Borderline

BORDERLINE PERSONALITY
DISORDER
A Clinical Guide
SECOND EDITION
This page intentionally left blank
BORDERLINE PERSONALITY
DISORDER
A Clinical Guide
SECOND EDITION
John G. Gunderson, M.D.

Professor of Psychiatry at Harvard Medical School
Director of Psychosocial and Personality Research and
Director of the Borderline Personality Disorder Center
McLean Hospital
Belmont, Massachusetts
With
Paul S. Links, M.D., F.R.C.P.C.
Professor of Psychiatry at the University of Toronto, Canada
Washington, DC
London, England
Note: The authors have worked to ensure that all information in this
book is accurate at the time of publication and consistent with general
psychiatric and medical standards, and that information concerning
drug dosages, schedules, and routes of administration is accurate at the
time of publication and consistent with standards set by the U.S. Food
and Drug Administration and the general medical community. As medi-
cal research and practice continue to advance, however, therapeutic stan-
dards may change. Moreover, specific situations may require a specific
therapeutic response not included in this book. For these reasons and
because human and mechanical errors sometimes occur, we recommend
that readers follow the advice of physicians directly involved in their care
or the care of a member of their family.
Books published by American Psychiatric Publishing, Inc., represent the
views and opinions of the individual authors and do not necessarily rep-
resent the policies and opinions of APPI or the American Psychiatric As-
sociation.
The authors of this book, John G. Gunderson, M.D., and Paul S. Links,
M.D., F.R.C.P.C., have no competing interests to disclose.
Copyright © 2008 American Psychiatric Publishing, Inc.
ALL RIGHTS RESERVED
Manufactured in the United States of America on acid-free paper
12 11 10 09 08 5 4 3 2 1
Second Edition
Typeset in Adobe’s New Baskerville and AvantGarde
American Psychiatric Publishing, Inc.
1000 Wilson Boulevard
Arlington, VA 22209-3901
www.appi.org
Library of Congress Cataloging-in-Publication Data
Gunderson, John G., 1942–
Borderline personality disorder : a clinical guide / John G. Gunderson
with Paul S. Links. — 2nd ed.
p. ; cm.
Includes bibliographical references and index.
ISBN 978-1-58562-335-8 (alk. paper)
1. Borderline personality disorder. I. Links, Paul S. II. Title.
[DNLM: 1. Borderline Personality Disorder. WM 190 G975ba 2008]
RC569.5.B67G863 2008
616.85'852—dc22
2008005061
British Library Cataloguing in Publication Data
A CIP record is available from the British Library.
CONTENTS
About the Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
1 The Borderline Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . 1

Origins of the Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Shifts in the Borderline Construct: From Organization to
Syndrome to Disorder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Epidemiology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
An Explication of the DSM-IV Criteria . . . . . . . . . . . . . . . . . . 9
A Clinical Synthesis: Intolerance of Aloneness . . . . . . . . . . 18
Misuses of the Borderline Diagnosis . . . . . . . . . . . . . . . . . . . 22
The Behavioral Specialty: Self-Injurious Behavior . . . . . . . . 24
Use of the Diagnosis in Adolescents . . . . . . . . . . . . . . . . . . . 27
How to Explain the Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . 27
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2 Differential Diagnosis: Overlaps, Subtleties, and

Treatment Implications . . . . . . . . . . . . . . . . . . . . . . . . 37
Overall Function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
The Changing Construct: From Schizophrenia to
Depression to Posttraumatic Stress Disorder to
Bipolar Disorder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Comorbidity and Differential Diagnosis . . . . . . . . . . . . . . . . 40
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3 Overview of Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Historical Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Generic Therapeutic Processes and the Functions
They Serve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Multiple Modalities and Step-Down Services:
An Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Sociotherapies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Establishing Goals: The Expectable
Sequence of Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Changes Within Four Spheres . . . . . . . . . . . . . . . . . . . . . . . . 77
The Initial Structuring of Treatment . . . . . . . . . . . . . . . . . . 80
Types and Sequence of Therapeutic Alliance . . . . . . . . . . . . 82
Countertransference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4 Case Management: The Primary Clinician . . . . . . . . 89

Qualifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Responsibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Liability Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Relationship Management . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Managing Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Responding to Recurrent Suicidality: The Principle of
False Submission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Implementing Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Boundaries, Violations, and Setting Limits . . . . . . . . . . . . . 106
Splits, Splitting, and the Virtues of Split Treatments . . . . . 107
Giving, Receiving, and Participating in Supervision . . . . . . 110
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5 Levels of Care: Indications, Structure, Staffing . . . . 113

Selecting or Changing a Level of Care . . . . . . . . . . . . . . . . 114
Level IV: Hospital Treatment—Makes Therapy Possible . . . 115
Level III: Residential/Partial Hospital Care/
Day Treatment—Basic Socialization . . . . . . . . . . . . . . . . 122
Level II: Intensive Outpatient Care—
Behavioral Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
Level I: Outpatient Care—Interpersonal Growth . . . . . . . . 133
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
6 Pharmacotherapy: Clinical Practices . . . . . . . . . . . 139
History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Overall Role of Medications . . . . . . . . . . . . . . . . . . . . . . . . . 140
Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
The Prescribing Psychiatrist’s Role . . . . . . . . . . . . . . . . . . . 143
Symptom Chasing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Attitudes, Meanings, and Attributions . . . . . . . . . . . . . . . . . 148
Transference-Countertransference Issues . . . . . . . . . . . . . . 149
Contraindications and Discontinuance . . . . . . . . . . . . . . . . 151
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
7 Pharmacotherapy: Selection of Medications . . . . . 157

Pharmacotherapy Models . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
The Symptom-Targeted Model . . . . . . . . . . . . . . . . . . . . . . 158
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
8 Family Interventions and Therapies . . . . . . . . . . . . . 177

History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
Therapists and Countertransferences . . . . . . . . . . . . . . . . . 180
Getting Started: Overcoming Resistance . . . . . . . . . . . . . . . 181
Phase 1: Initial Family Meetings . . . . . . . . . . . . . . . . . . . . . . 183
Phase 2: Establishing an Alliance . . . . . . . . . . . . . . . . . . . . . 187
Phase 3: Psychoeducational Family Therapy . . . . . . . . . . . . 189
Phase 4: Psychodynamic Family Therapy . . . . . . . . . . . . . . 196
Marital or Couples Therapy . . . . . . . . . . . . . . . . . . . . . . . . . 197
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
9 Group Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

Indications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
Engaging Patients and the Primary Clinician’s Role . . . . . 208
Skills Training Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Psychodynamic Group Therapies . . . . . . . . . . . . . . . . . . . . 214
Group Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
Common Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
10 Individual Psychotherapies: Getting Started . . . . . . 225
Introduction: Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . 225
Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Structuring the Therapeutic Frame
(External Boundaries) . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Therapists . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
Engagement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Generic Qualities of Effective Psychotherapies . . . . . . . . . 247
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
11 Cognitive-Behavioral Therapies: Dialectical

Behavior Therapy and Cognitive Therapies . . . . . . 253
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
Basic Operant Conditioning Applications for All
Treatment Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
Dialectical Behavior Therapy . . . . . . . . . . . . . . . . . . . . . . . . 256
Cognitive Therapies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
12 Psychodynamic Psychotherapies . . . . . . . . . . . . . . 279

Pre-Empirical Developments . . . . . . . . . . . . . . . . . . . . . . . . 280
Outcome Studies: Nonrandomized Trials . . . . . . . . . . . . . . 282
Transference-Focused Psychotherapy . . . . . . . . . . . . . . . . . 286
Overview of Change Processes . . . . . . . . . . . . . . . . . . . . . . . 289
Phase 2: A Relational Alliance . . . . . . . . . . . . . . . . . . . . . . . 291
Phase 3: Positive Dependency . . . . . . . . . . . . . . . . . . . . . . . 298
Phase 4: Secure Attachment, the Working Alliance,
and Consolidation of Self . . . . . . . . . . . . . . . . . . . . . . . . 304
Impasses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
13 Future Considerations . . . . . . . . . . . . . . . . . . . . . . . . 315

Treatment Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
Diagnostic Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
Public Awareness and Advocacy . . . . . . . . . . . . . . . . . . . . . . 321
Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
Appendix Psychoeducational Resources:
Printed Materials, Videos, Films, and Web Sites . . . 329
Printed Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329
Videos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
Films . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
Web Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
List of Sidebars
1–1: Where Were the Borderline Patients Before
the Diagnosis? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1–2: The Subjective Experience of Being Borderline . . . . . 13
1–3: Borderline Personality as an Iatrogenic Disorder . . . . 15
1–4: British Developmentalists: From Winnicott to
Bowlby to Fonagy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1–5: “Wisdom Is Never Calling a Patient Borderline” . . . . 23
1–6: Cutting: Social Contagion or
Psychopathology? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2–1: Was Vincent van Gogh Borderline? . . . . . . . . . . . . . 46
2–2: Is Martha Stewart Borderline? . . . . . . . . . . . . . . . . . . . 53
3–1: Should Consumers Receive Progress Reports? . . . . . 74
3–2: Myths About Alliance With Borderline Patients . . . . . 83
4–1: Guidelines to Avoid Liability . . . . . . . . . . . . . . . . . . . . 93
4–2: Is Contracting for Safety Safe? . . . . . . . . . . . . . . . . . . 99
5–1: Can Long-Term Hospitalization Be Desirable
for BPD? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5–2: How Psychotherapeutic Technique Relates to
Level of Care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
5–3: Vocational Counseling: Should a Borderline
Patient Return to School, Pursue a Career, or
Become a Caregiver? . . . . . . . . . . . . . . . . . . . . . . . . 123
5–4: Empirical Support for a Specialized
“Mentalization-Based” Day Hospital . . . . . . . . . . . . 128
6–1: Listening to Prozac: Can Selective
Serotonin Reuptake Inhibitors Cure BPD? . . . . . . . . 141
6–2: Liability Hazards of Split Treatment . . . . . . . . . . . . . . 147
8–1: “You Can’t Talk to My Parents” . . . . . . . . . . . . . . . . . 182
8–2: Families of Married Borderline Patients . . . . . . . . . . 184
8–3: Finessing the Guilt Issue . . . . . . . . . . . . . . . . . . . . . . . 186
8–4: “Good Cop/Bad Cop”: A Parental Problem . . . . . 191
8–5: Makes Sense, But Does it Work?
Preliminary Findings of the Psychoeducational
Multiple-Family Group . . . . . . . . . . . . . . . . . . . . . . . . 195
9–1: Research on Interpersonal Group Therapy
With BPD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
10–1: Quality Assurance: Should Therapists Be
Credentialed to Treat BPD? . . . . . . . . . . . . . . . . . . . 240
10–2: Listening to Kernberg or Linehan: Can Charisma
Cure BPD? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
11–1: Schema-Focused Therapy: Does It Work? . . . . . . . 269
12–1: Kernberg Versus Kohut/Adler: The Debate of
the 1970s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
12–2: Kernberg Versus Linehan: The Debate of
the 1990s . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
12–3: Transitional Objects: From Concept to
Phenomenon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
12–4: Is Regression Therapeutic? The Two Margarets . . . . 302
13–1: Were a Famous Borderline Person
to Go Public… . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
ABOUT THE AUTHORS
John G. Gunderson, M.D., is Professor of Psychiatry at Harvard Medical

School, Boston, Massachusetts. At McLean Hospital in Belmont, Massa-
chusetts, he serves as Director of Psychosocial and Personality Research
and of the Borderline Personality Disorder Center.
Dr. Gunderson has been integrally involved in the diagnosis and treat-
ment of borderline personality disorder since 1969. His studies helped to
define the diagnosis of borderline personality disorder and establish an
empirical standard for evaluating treatment of this disorder.
Paul S. Links, M.D., F.R.C.P.C., is Professor of Psychiatry at the Univer-

sity of Toronto, Canada. At St. Michael’s Hospital, he is serving his third
term as the Arthur Sommer Rotenberg Chair in Suicide Studies.
xi
INTRODUCTION
THIS BOOK IS A SEQUEL to Borderline Personality Disorder: A Clinical

Guide, published in 2001, which was a sequel to Borderline Personality Dis-
order, published in 1984. These books have summarized what was
known—or believed—about treating borderline personality disorder
(BPD) at the time of their publication. This revision of the 2001 book up-
dates the rapidly expanding treatment literature. Clinical perspectives,
concepts, and modalities continue to become more sophisticated, de-
tailed, and empirically buttressed. A great deal is currently known about
what to do and, just as important, what not to do to treat BPD effectively.
After only a 7-year interval, a significant revision was needed to accom-
modate the mushrooming of information, expertise, and specialization.
This book is meant to cover comprehensively all the recognized thera-
pies for BPD. It details long-term multimodal treatment, with an apprecia-
tion that no one modality is by itself sufficient. I attempt to emphasize
advances from empirical research and to synthesize them with what is feasi-
ble and with what derives from clinical experience. Above all, this book is
meant to be useful and practical, primarily for clinicians, but also potentially
for trainees, patients’ families, and health care administrators. Although no
treatment is excluded from consideration because of its cost, all treatments
are considered with issues of cost-effectiveness and feasibility in mind.
The first chapter covers the issue of the diagnosis itself: what the diag-
nosis means and the biases that affect its use. Special attention is given to
the borderline patients’ behavioral specialty (i.e., their self-destructive-
ness) and to the use of this diagnosis in adolescents. Perhaps the most im-
portant message for clinicians is that we do these patients (and their
families) a favor by identifying the diagnosis and educating them about it.
Patients and families deserve to know what is known, and, as often as not,
the success of treatments rests on their being included as responsible allies.
Chapter 2 describes BPD’s most common differential diagnostic is-
sues. These have shifted from schizophrenia to depression to posttrau-
matic stress disorder to the current controversy about the interface with
bipolar disorder. The last issue is now given more attention.
xiii
xiv ❘ BORDERLINE PERSONALITY DISORDER: A CLINICAL GUIDE
Three theories guide most clinicians—biological, cognitive-behav-

ioral, and psychodynamic. My psychodynamic background inevitably an-
chors much of this book, but the theory most central to this book’s goals
is a theory about therapies found in Chapter 3. Chapter 3 offers an em-
pirically and clinically anchored theory on the sequencing of goals (i.e.,
targets for intervention), on processes of change, and on the modalities
that are best suited for a patient’s changing needs. The chapter under-
scores the feasibility and value of establishing initial short-term goals. In
this chapter, I also describe the basic role of psychoeducation for patients
and families—both as a way to establish an alliance with longer-term goals
and as a therapeutic intervention in its own right. Psychoeducation su-
perimposes a logic on treatment planning, and it anticipates the se-
quence of the chapters that compose the rest of the book.
After Chapter 3, the book proceeds to chapters concerning the imple-
mentation of overall treatment plans (Chapters 4 and 5). These chapters
emphasize the need for someone to develop the plan, establish an appro-
priate level of care, and include rehabilitative services that address the
borderline patients’ typically severe impairments in social functioning.
Chapters 6 through 12 describe the specific modalities in a sequence con-
sistent with the severity of the borderline patient’s mental state and with
the length of time needed to meet the primary goals of each modality.
Chapter 4 outlines the primary clinician’s responsibilities. In an era
when managed care hovers in the background of treatment authoriza-
tion, and in which care of borderline patients moves across multiple set-
tings, it is easy to ignore the central requirement of having some one
clinician be identified to all, including the patient, as being in charge—
the primary clinician. This chapter introduces the thesis that rather than
being problematic, split treatments that emerge from the current multi-
modal, multitreater environment are treatment enhancing. This book re-
peatedly points out how a treatment having two or more components not
only adds breadth to the treatment goals but also offers a structure that
safeguards treatment against the borderline patients’ enactment of their
intrapsychic splits.
Chapter 5 concerns four levels of care. Here empirical evidence is in-
troduced about the potential value of the two most intensive levels: hospital
care (level IV) and residential or partial hospital (level III). The four levels
are not seen here as competitive but as having different goals, durations
(this was intended 7 years ago), structures, and staffing. Of most interest
may be the endorsement given to a newer level of care, intensive outpatient
programs. Intensive outpatient programs represent a level of care that, al-
though not widely available, may be more effective and certainly more cost-
beneficial than relying on hospital or partial hospital services.
Introduction ❘ xv
Reflecting the fact that medications have quietly become the single
most widely and uniformly used treatment for BPD, two chapters (6 and
7) are devoted to psychopharmacology. Chapter 6 offers an extensive ac-
count of the seemingly irrational in vivo complexities surrounding pre-
scribing medications and evaluating their effectiveness. In contrast,
Chapter 7 offers a rational algorithm to guide selection of medications
that should usefully inform prescribing physicians.
Chapter 8 encourages clinicians to involve families far more than has
been customary. I describe how clinicians can use consumer-friendly
psychoeducational interventions. Note that many interventions, albeit
brief and not called therapies, may be very valuable. Furthermore, use of
traditional dynamic family therapies is reserved for only selected cases
and then only in a late stage of treatment. For most families, the primary
treatments are parental coaching and assisted problem solving. Prelimi-
nary data that show the value of such coaching and problem solving are
offered.
Chapter 9 underscores the role that interpersonal groups should play
in the first year or so of most borderline patients’ treatment. This type of
treatment is readily exportable and nicely complements the functions
served by individual therapies or psychopharmacology by addressing the
interpersonal impairment that is central to most borderline patients’ dis-
order. The available empirical evidence underscores the need for more
use of and more research on interpersonal groups.
In Chapter 10, I argue that initiating individual psychotherapy should
be done selectively, taking into account the motivation, the aptitude, and
the social supports required of both patients and therapists. Otherwise
skilled cognitive-behavioral or dynamically oriented therapists still need
special training and experience, and perhaps special personality traits, to
do such therapies well. Chapter 10 also outlines some of the general over-
lapping characteristics of all effective psychotherapies.
Although cognitive-behavioral principles have always been needed
for adequate treatment of BPD, Chapter 11 recognizes that specific types
of cognitive-behavioral treatments have now become the cornerstone for
much modern theory and practice. Indeed, dialectical behavior therapy
(DBT) has rapidly become the most BPD-specific and empirically sub-
stantiated treatment for BPD. Unquestionably, DBT was the major ad-
vance in therapeutics of the 1990s. Chapter 11 tries both to acquaint the
uninitiated with DBT and to place it in some perspective. Other notable
developments cited in Chapter 11 include the recent addition of a prom-
ising second empirically validated cognitive-behavioral treatment,
schema-focused psychotherapy, and evidence for the potential for short-
term cognitive-behavioral therapies to be effective for discrete goals.
xvi ❘ BORDERLINE PERSONALITY DISORDER: A CLINICAL GUIDE
Chapter 12 is devoted to psychodynamic (i.e., psychoanalytic) psy-

chotherapy, the modality that for several decades was considered the
treatment of choice for BPD. Chapter 12 highlights the long-needed
emergence of empirical support for Kernberg’s transference-focused
form of therapy. It also delineates the phases of psychotherapeutic
progress and change over a period of 4 or more years. Both the corrective
power of the relationship and the growth made possible by learning (i.e.,
insights) are described. The case is made that progress should be ongo-
ing, and its absence should be cause for review and consultation. This op-
timistic message is set against the need for a protracted multiyear process.
Chapter 13 begins by noting that the borderline diagnosis has
achieved a place in the consciousness of the mental health world, but it is
only beginning to establish a place in the public consciousness. The de-
velopment and influence of the Borderline Personality Disorder Re-
search Foundation, the explicit recognition of a need for more research
on BPD by the National Institute of Mental Health, the rise of family ad-
vocacy groups, and the adoption of BPD as a brain disease by the National
Alliance on Mental Illness dramatically signal that this expansion is un-
der way. Chapter 13 also introduces how the rise in neurobiological re-
search is likely to greatly transform our understanding of borderline
patients.
The remarkable advancements in treatment for BPD described and
celebrated in this book can be expected to continue. The current diver-
sity of theory and research creates a healthy, vibrant vehicle for continued
growth.
Borderline patients require an array of clinical services, any of which
can be harmful or helpful. But to treat this disorder effectively requires
clinicians with specialized knowledge and training. When such condi-
tions are present, beneficial changes occur that greatly reduce patients’
dysphoric mental states and enhance social functioning. Effective treat-
ment results in a concurrent reduction in the burden on borderline pa-
tients’ significant others, an improved morale by treaters, and a decrease
in the otherwise enormous public health costs.
Chapter 1
THE BORDERLINE DIAGNOSIS
THE BORDERLINE PERSONALITY DISORDER (BPD) diagnosis en-

tered the American Psychiatric Association’s DSM-III in 1980 (American
Psychiatric Association 1980) and 12 years later, in 1992, was adapted for
the World Health Organization’s ICD-10 (World Health Organization
1992). The growth in the recognition and use of this diagnosis during the
period from 1975 to 1990 has been remarkable. It is easily the most widely
and commonly used diagnosis for personality disorders in modern clinical
practice (Loranger 1990; Loranger et a1. 1997). Individuals with BPD con-
stitute about 2%–3% of the general population (Swartz et al. 1990; Zim-
merman and Coryell 1989), about 25% of all inpatients, and about 15% of
all outpatients (Koenigsberg et al. 1985; Widiger and Weissman 1991).
Origins of the Diagnosis

The origins of the borderline diagnosis, illustrated in Figure 1–1, are usu-
ally traced to the clinical observations of Adolph Stern (1938), a psycho-
analyst in office practice, who recognized that a subgroup of his patients
disregarded the usual boundaries of psychotherapy and did not fit into the
existing classification system, a system concerned primarily with dividing
psychoses from neuroses. A scholarly review of the work preceding Stern’s
can be found in Mack (1975). The patient group became somewhat more
widely recognized in the early 1950s as a result of several influential papers
by Robert Knight (1953, 1954). He expanded the descriptor borderline
from relating to only the border with neurosis to being equally relevant to
the border with psychosis. Like Stern, he began by decrying the “wastebas-
ket” diagnostic status for such patients. However, he added that failure to
identify the unique needs of these patients was responsible for the trou-
bling disagreements between staff members on inpatient units; he further
stated that this failure led clinicians to ignore providing the structure such
1
2 ❘ BORDERLINE PERSONALITY DISORDER: A CLINICAL GUIDE
FIGURE 1–1. Development of the borderline construct, I.

BPD= borderline personality disorder; BPO= borderline personality organiza-
tion; BSz= borderline schizophrenia; STPD=schizotypal personality disorder.
patients needed to avoid regressing. After Knight, the term borderline to

denote troublesome patients who were neither psychotic nor neurotic
retained some currency but primarily within the community of psychoan-
alysts who worked in hospital settings (Sidebar 1–1).
Sidebar 1–1: Where Were the Borderline Patients

Before the Diagnosis?
A review of medical records from Danish and British psychi-
atric institutions before the diagnosis was used showed that
borderline patients existed (Gunderson et al. 1983; Kroll et
al. 1982). Although Freud himself used the term borderline
only to differentiate delinquent acting-out adolescents
from those with neuroses (Aichhorn 1925/1945, Introduc-
tion), years later Wolberg (1973) rediagnosed one of
Freud’s most famous patients, the “Wolf Man,” as being
borderline. Certainly, before the diagnosis, clinicians (Aich-
horn 1925/1945; Alexander 1930; W. Reich 1949) described
impulse-driven disorders presaging what was to become
the BPD diagnosis. Therefore, there is every reason to be-
lieve that borderline patients were present in clinical set-
tings long before the diagnosis.
Still, it is possible that what was formerly rare is now far
more common. Grinker et al. (1968) suggested that bor-
The Borderline Diagnosis ❘ 3
derline psychopathology is a by-product of social changes

during the twentieth century. The earlier burdens of man-
ual labor and the earlier restrictions of travel, communica-
tion, and leisure time may have offered the structure,
survival activities, and monitors that silently kept such psy-
chopathology in check. Millon (1987) developed a thesis
(subsequently elaborated by Paris [1992]) about sociocul-
tural causes for BPD that, if taken to its extreme, is consis-
tent with the possibility that BPD would have been far less
common in other eras. At present, this thesis can be tested
only by epidemiological work showing whether the inci-
dence and prevalence of BPD vary between cultures and
their levels of modernity.
Use of the term borderline for atypical, clinically troubling cases stag-
gered along in the periphery of psychiatric thinking without notable
progress until developments in the late 1960s. At this point, the conflu-
ence of three independent investigations forced the questions about a
borderline consciousness.
The first of these investigations came from Otto Kernberg (1967).
Even as a relatively young man, Kernberg authoritatively added to the psy-
choanalytic perspective of the borderline construct. He defined borderline
personality organization as one of three forms of personality organization, to
be differentiated from sicker patients, who had psychotic personality organi-
zation, and healthier patients, who had neurotic personality organization (Fig-
ure 1–2). Borderline personality organization was characterized by failed
or weak identity formation, primitive defenses (namely, splitting and pro-
jective identification), and reality testing that transiently lapsed under
stress. Kernberg’s scheme was a conceptual advance within the psycho-
analytic community by virtue of integrating object relations with ego
psychology and the instincts and by virtue of giving a rationale and orga-
nization to a basic classification system. However, the effect of his scheme
within the larger mental health community derived more from the opti-
mistic therapeutic mandates that he gained from his way of understand-
ing these patients than from the concept itself (see Kernberg 1968, 1975).
The second seminal contribution was provided by Roy Grinker et al.
(1968), a senior and respected statesman within American psychiatry. Ar-
mored with a brief personal analysis by Freud himself, Grinker had be-
come chairman of psychiatry at The University of Chicago and editor of
the Archives of General Psychiatry. As one of the early champions of the
need for empirical research, and having already made major contribu-
tions to studies of depression and posttraumatic stress disorder (PTSD),
Grinker undertook the first empirical study of borderline patients. With
FIGURE 1–2. Concepts of borderline disorders.

BPD= borderline personality disorder; PD= personality disorder.
Source. Reprinted from Gunderson JG: Borderline Personality Disorder. Washing-
ton, DC, American Psychiatric Press, 1984, p. 12. Copyright 1984, John G. Gun-
derson. Used with permission.
the publication of The Borderline Syndrome in 1968, Grinker and colleagues

established the accessibility of this patient group to clinical research
methods and offered the first empirically based criterion set. The criteria
were 1) failures of self-identity, 2) anaclitic relationships, 3) depression
based on loneliness, and 4) the predominance of expressed anger.
The third major investigation to impel borderline patients into the
consciousness of the mental health community was not intended to ad-
dress these patients at all. In the Danish adoption studies that proved to be
the cornerstone for establishing a biogenetic basis for schizophrenia, Kety
and colleagues (1968) were forced to develop criteria by which to identify
whether nonpsychotic relatives had schizophrenia spectrum (i.e., “bor-
derline schizophrenic”) disorders. (Despite the power of adoptive designs
to test heritability, the basic prevalence of schizophrenia is so small—
about 1%—that it would have taken an infeasible number of adoptees to
see statistically enhanced rates of schizophrenia per se in relatives.) Hence
the genetic transmission of schizophrenia was established by document-
ing the higher-than-expectable rates of relatives with “borderline” (mean-
ing atypical) schizophrenia. Once genetic transmission was established, it
became critically important to develop replicable ways to identify who
these “borderline schizophrenic” patients were. Although these individu-
als were subsequently shown to have personality disorder (Gunderson et
al. 1983), the effect of this work was to stimulate further research interest
in the borderline diagnosis and to move the theorizing about such pa-
tients into the realms of genetic transmission and biological therapies.
In the historical context of these three independent investigations—
analytic, descriptive, and genetic—my own contribution began. At Massa-
chusetts Mental Health Center in 1969, I conducted a small study char-
acterizing the diagnostically “wastebasket” patients who most distressed
my group of beginning residents. My interest subsequently intensified
while at the National Institute of Mental Health (NIMH), where I became
aware of the three investigations—and the three primary investigators—
cited earlier. This interest prompted a collaboration with Carpenter and
Strauss to disentangle borderline patients from those with a diagnosis of
schizophrenia (Gunderson et al. 1975); more important, it prompted the
review and synthesis of all the relevant literature in collaboration with
Singer. That review, “Defining Borderline Patients: An Overview” (Gun-
derson and Singer 1975), received such surprising acclaim when it was
published in 1975 that my involvement greatly intensified. What followed
was the development of a structured interview (Diagnostic Interview for
Borderline Patients [DIB]; Gunderson et al. 1981), with which the diag-
nosis could be made reliably and with which we were able to identify a set
of discriminating characteristics (Gunderson and Kolb 1978). Spitzer, as
overseer of the development of DSM-III, used these characteristics in a
survey of clinical practices, and, with the addition of the criterion about
identity diffusion that derived from Kernberg, the characteristics were all
validated as being the most discriminating in clinical practice (Spitzer et
al. 1979). The disorder defined by these criteria narrowed the syndrome
from the definitions offered by Kernberg and Grinker (see Figure 1–2).
In 1980, the BPD diagnosis, amid considerable controversy, entered the
official classification system, DSM-III.
It was official, but what was it?
Shifts in the Borderline Construct:

From Organization to Syndrome to Disorder
The borderline construct has undergone several major shifts since the
1960s (Gunderson 1994). It was first a personality organization and then
a syndrome; it is now a disorder (see Figure 1–2). These three versions of
the construct reflect more general epochs within the field of psychiatry,
as psychiatry itself has shifted from a psychoanalytic paradigm, to a re-
medicalization with empirical and pharmacological bases, to the ongo-
ing search for psychiatric diagnoses to convey meaning in terms of
specific etiology and specific treatment.
The identification of the borderline personality became very wide-

spread in the United States and elsewhere during the 1970s. As is de-
scribed in Chapter 3, this increase in usage was largely a result of the
optimistic endorsements of ambitious long-term psychoanalytic treat-
ments by Kernberg (1968) and Masterson (1971). What was meant by
calling a patient borderline was intellectually tied to borderline person-
ality organization, an intrapsychic organization that aided psychody-
namic clinicians in understanding these patients. In practice, borderline
usually meant an angry, manipulative patient who would be a problem
but who might receive long-term, psychoanalytically based hospitaliza-
tions or psychotherapies.
In the aftermath of DSM-III and the adoption of standardized criteria,
the term borderline became used by a much wider segment of the mental
health community. A study of first-admission diagnoses in Denmark
showed a dramatic increase in use of this diagnosis between 1970 and
1985: the increase in Copenhagen was fivefold (Mors 1988). Within aca-
demic circles, the borderline diagnosis shifted from intrapsychic organi-
zation to descriptive syndrome—that is, a cluster of phenomena that co-
occurred with greater-than-chance frequency and that could discrimi-
nate borderline patients from other types of patients. The value of the
syndromal concept was that it incited researchers to establish the syn-
drome’s meaning through studies of the syndrome’s course, genetics, co-
morbidity, development, and treatment response and, of course, by
documenting its discrimination from other disorders in all these areas. It
also opened the door to a new array of therapeutic modalities. Figure 1–
3 identifies the sequence by which the borderline construct has been re-
fined to include the domains of cognition, affect, impulse, and trauma.
The figure also highlights the progression by which a group of enterpris-
ing empiricists have added validating evidence with respect to phenom-
enology, family history, course, treatment response, and development.
Figures 1–4 and 1–5 illustrate the remarkable explosion of publica-
tions and research that occurred between 1968 and the 1990s. Of partic-
ular note is the parallel rate of growth of published articles, both clinical
and empirical, although the number of clinical reports remained nearly
10 times as great as the number of empirical studies. The number of
books on BPD showed a similar logarithmic rise in number up to 1994
(Figure 1–5). Notably, the percentage of those books reflecting a psycho-
analytic perspective has steadily declined, from 80% in 1974 to only 23%
of those written between 1995 and 1999. Also notable is that since the
mid-1990s, there has been a dramatic decline in the number of new
books. I think this is because psychoanalytic and other clinical observa-
tions have already said most of what they can say. The field awaits a syn-
FIGURE 1–3. Development of the borderline construct, II.

BPD= borderline personality disorder.
thesis and implementation of what has been learned, and, as described

throughout this book, it awaits advances that are more empirically based.
Since DSM-III, the descriptive characteristics of BPD have been the
subject of an enormous amount of research. These studies have found
that 1) BPD has a high level of comorbidity with Axis I disorders and per-
sonality disorders (see Chapter 2); 2) 29 longitudinal studies have shown
considerable heterogeneity in course (Grilo et al. 1998); and 3) sub-
groups can be identified on the basis of defenses, medication response,
neurobiological impairment, trauma, and factor analyses. Still, despite
these evidences of heterogeneity, the overall results have tended to vali-
date the integrity and clinical utility of the diagnosis. The clinical utility of
the diagnosis is discussed in Chapter 2, but from a scientific point of view,
the validating evidence derives from the following developments:
• Establishing a course that is distinctive from that of either psychotic or

depressive disorders
• Showing that few borderline patients resolve into a psychotic or a
mood disorder
• Establishing a pathogenesis marked by significant heritability (Tor-
gersen et al. 2001), family environments with high conflict and unpre-
dictability (Gunderson and Zanarini 1989), and high frequencies of
FIGURE 1–4. Number of research studies and clinical reports on

borderline personality disorder, 1970–1990.
FIGURE 1–5. Number of books on borderline personality disorder,

1968–1999.
Source. Search of Library of Congress database, 1999.
sexual abuse and other trauma (Gunderson and Sabo 1993; Zanarini
et al. 1997)
• Confirming that both modalities and techniques specific to this diag-
nosis have preferential benefits
Therefore, although boundaries remain blurred into a variety of

other diagnostic types (see Chapter 2), the diagnosis has established itself
as a viable disorder awaiting the more conclusive evidence of measurable
neuropathology (see Chapter 13).
Epidemiology
The epidemiological data about BPD remains methodologically weak.
Thus, all the figures reported in Table 1–1 should be considered best
estimates. Of particular interest, but particularly speculative, are the au-
thors’ estimates about age at onset, which are based on retrospective ac-
counts. Almost certainly, prodromal signs of this disorder (e.g., cutting)
are identifiable for most of the patients who develop it, but we know little
about this. Similarly, for prevalence, the critical studies in the general
population have not been done. Still, it seems clear that it will probably
be about 1%.
Although the epidemiological data are not strong, borderline pa-
tients constitute a high proportion (approximately 20%) of psychiatric
inpatients and outpatients. They are also high consumers of emergency
department services, crisis lines, and psychiatric consultative liaisons to
other medical services (Ellison et al. 1989; J. Reich et al. 1989).
Patients with BPD represent 9%–33% of all suicides (Kullgren et al.
1986; Runeson and Beskow 1991). Among patients 15 years or older pre-
senting to the hospital with suicide attempts, 41% were given the diagno-
sis of BPD, and 56% of the female attempters had BPD (Persson et al.
1999). At least 50% of chronically suicidal patients with four or more vis-
its in a year to a psychiatric emergency department are patients with BPD
(Bongar et al. 1990). Such patients accounted for more than 12% of all
psychiatric emergency department visits during the year studied. De-
pending on the study, the lifetime risk of suicide among patients with
BPD is between 3% and 10% (Paris and Zweig-Frank 2001).
An Explication of the DSM-IV Criteria

The DSM-IV (American Psychiatric Association 1994) criteria set for BPD
changed only modestly from the original in DSM-III. Changes were
based on an extensive series of descriptive studies (see Gunderson et al.
1996).
TABLE 1–1. Demographics of borderline personality disorder

AGE AT ONSETa
Age group Percentage
Adolescence (ages 13–17) 15
Early adulthood (ages 18–25) 50
Young adulthood (ages 26–30) 25
Adulthood (ages 31–48) 10
PREVALENCE
Population Percentage
General population 0.4–3b
Inpatient populations 15–18c
Outpatient populations 15–25d
Socioeconomic status (SES) Possible increase in low SES,
otherwise evenly distributed
Gender 75% female
Race No known variations
aThe percentages estimated to have onset in each age group are based on the au-
thor’s experience; no epidemiological data are available.
bCoryell and Zimmerman 1989; J. Reich et al. 1989; Swartz et al. 1990; Torgersen
et al. 2001.
cDahl 1986; Loranger 1990; Widiger and Weissman 1991.
dKoenigsberg et al. 1985; Widiger and Weissman 1991.
Table 1–2 shows the nine criteria in DSM-IV for diagnosing BPD, as
well as changes in the criteria from DSM-III-R (American Psychiatric As-
sociation 1987) to DSM-IV. The criteria in the table are organized accord-
ing to their association as factors and in the approximate order of their
diagnostic value (differing from their order in DSM-IV), and only signif-
icant changes are shown. The text that follows is an amplification of each
criterion to emphasize its clinical meaning.
Disturbed Relationships
1. Unstable, intense relationships. This criterion describes the interper-

sonal manifestations of intrapsychic splitting. A hallmark of border-
line psychopathology is the inability to see significant others (i.e.,
potential sources of care or protection) as other than idealized, if
TABLE 1–2. DSM-III-R and DSM-IV diagnostic criteria for

borderline personality disorder (adapted)
1. A pattern of unstable and intense interpersonal relationships
characterized by alternating between extremes of [over]idealization
and devaluation
2. Frantic efforts to avoid real or imagined abandonment (do not
include suicidal or self-mutilating behavior covered in Criterion 7)
3. Chronic feelings of emptiness [or boredom]
4. Affective instability [: marked shifts from baseline mood to
depression, irritability, or anxiety] due to a marked reactivity of mood
(e.g., intense episodic dysphoria, irritability, or anxiety usually lasting a
few hours and only rarely more than a few days)
5. Inappropriate, intense anger or lack of control of anger (e.g.,
frequent displays of temper, constant anger, recurrent physical
fights)
6. Impulsivity in at least two areas that are potentially self-damaging,
e.g., spending, sex, substance abuse, reckless driving, binge eating
(do not include suicide or self-mutilating behavior covered in
Criterion 7)
7. Recurrent suicidal behavior, gestures, or threats, or self-mutilating
behavior
8. Identity disturbance [uncertainty about at least two of the following:
self-image, sexual orientation, goals or career choice, type of friends,
values]; markedly and persistently unstable self-image and/or sense
of self a
9. Transient, stress-related paranoid ideation or severe dissociative symptomsb
Note. Text in italics is significant text that was not in DSM-III-R (American Psy-
chiatric Association 1987) but was introduced in DSM-IV (American Psychiatric
Association 1994). Text in brackets is significant text that appears in DSM-III-R
but does not appear in DSM-IV.
aE.g., he or she may feel that he or she does not exist or embodies evil [author’s
note].
bOr depersonalization, derealization, or hypnagogic illusions [author’s note].
Source. Adapted from American Psychiatric Association: Diagnostic and Statistical

Manual of Mental Disorders, 3rd Edition, Revised. Washington, DC, American Psy-
chiatric Association, 1987; American Psychiatric Association: Diagnostic and Statis-
tical Manual of Mental Disorders, 4th Edition. Washington, DC, American
Psychiatric Association, 1994. Used with permission. Copyright 1987, 1994 Amer-
ican Psychiatric Association.
gratifying, or devalued, if ungratifying. Kernberg (1967) is responsi-

ble for identifying the importance of the Kleinian construct of split-
ting for BPD. His theory traces splitting to unmitigated anger, initially
intended toward still-needed caregivers. As such, this criterion is de-
velopmentally closely tied to the abandonment criterion.
2. Abandonment fears. This criterion reflects Masterson’s seminal contri-
bution to the borderline construct (Masterson 1972; Masterson and
Rinsley 1975). The criterion needs to be differentiated from the more
common, less pathological separation anxieties. Although borderline
patients are quite aware of abandonment fears, some patients are so
accustomed to acting out in response to such fears that they do not rec-
ognize the fears. This criterion was reframed as “intolerance of alone-
ness” by Gunderson and Singer (1975) and Adler and Buie (1979).
Although Masterson attributed this trait to failure in the rapproche-
ment subphase of development (ages 16–24 months), Mahler and
Kaplan’s (1977) empirical investigation indicated that children can
develop these anxieties without noticeable problems in the rapproche-
ment subphase or can fail to develop these fears despite very notice-
able rapprochement failures. This criterion is now recognized as a
symptom of early insecure attachment (Fonagy 1991; Gunderson
1996).
3. Emptiness. Early analysts (Abraham 1927; Freud 1908/1959) con-
ceived of an oral phase of development that, if unsuccessfully com-
pleted, created a disposition toward adult depression and dependent,
object-hungry relatedness. This conceptualization was modified by
object relations theorists (e.g., M. Klein 1932, 1946) who suggested
that insufficiencies of early caretaking resulted in a failure to introject
a soothing other (i.e., an internalized sense of oneself as being cared
for), with a resultant inability to self-soothe or to conjure up represen-
tations of soothing others. This internal absence leaves the child vul-
nerable and is theorized to be evident in the subjective experience of
emptiness.
The early literature and the widespread use of the DIB brought
this criterion to attention in partnership with boredom (Gunderson
and Kolb 1978). Boredom was eliminated from DSM-IV because it ac-
tually proved more characteristic of narcissistic personality disorder
(Gunderson et al. 1996). Its link to emptiness in earlier accounts re-
flects the unclear clinical and theoretical discrimination between
these two types of personality disorder (Singer 1977). Emptiness is a
visceral feeling, usually in the abdomen or chest, not to be confused
with fears of not existing or with existential anguish. Emptiness is an
exemplary criterion—discriminating BPD from other types of de-
pression (Westen et al. 1992) and linking the borderline individual’s

subjective experience to presumed developmental failures. Balint
(1992) identified the feeling of “something missing” as the basic
fault. Although emptiness is a valuable criterion, recent work has sug-
gested that other aspects of the subjective experience of borderline
patients may be equally or even more discriminating (Sidebar 1–2).
Sidebar 1–2: The Subjective Experience

of Being Borderline
Zanarini and Frankenburg (1994) identified the borderline
patient’s typical “hyperbolic” style, referring to his or her in-
tense, insistent, and dramatic style of communicating feel-
ings and wishes to others. As a sequel, Zanarini et al. (1998)
conducted a study documenting “the pain of being bor-
derline.” Systematic inquiries of 50 dysphoric feelings and
thoughts dramatically underscored their thesis: patients
with BPD scored higher than patients without BPD on all 50!
Patients with BPD spent far higher percentages of their time
feeling overwhelmed (61.7%), worthless (59.5%), very angry
(52.6%), lonely (63.5%), or misunderstood (51.8%). More re-
vealing was that some borderline patients reported suffer-
ing for high percentages of the time for reasons and in
ways that nonborderline patients rarely do: feeling aban-
doned (44.6%), betrayed (35.9%), evil (23.5%), out of con-
trol (33.5%), like a small child (39.1%), and like hurting or
killing themselves (44%). The intensity of pain and the
amount of time suffering pain reported by borderline pa-
tients allow clinicians to easily discriminate patients with
BPD from those without. Granted, the borderline sample
was newly admitted inpatients, who can be counted on to
maximize their reports of pain to garner support, but the re-
sults underscore that this disorder involves a terrible way to
experience life.
Affective Instability
4. Affective instability. This criterion developed out of the work of early
clinical observers (e.g., Grinker et al. 1968; Zetzel 1971) who were im-
pressed by the intensity, volatility, and range of the borderline pa-
tient’s affects. As described earlier, such observations prompted D.
Klein (1975, 1977), Stone (1979, 1980), and Akiskal (1981, 1985) to
propose that the basic psychopathology of borderline individuals in-
volved the same problems of affective regularity found in people with
mood disorders—originally depression, now bipolar II disorder.
Linehan and other cognitive-behavioral clinicians have adopted the

concept of affective dysregulation as the borderline individual’s core
psychopathology, suggesting that intense emotions propel the behav-
ior problems (see Chapter 11, section “DBT Theory”). Such theories
have encouraged the testing and widespread use of mood-regulating
medications (Chapter 6). Revisions of this criterion since DSM-III
have tried to distinguish the affect shifts of borderline patients as be-
ing more reactive (read “less autonomous”) and less enduring than
those in mood disorders.
5. Anger. As noted, Kernberg (1967) first suggested that the source of
borderline psychopathology involved excessive aggression, due ei-
ther to a temperamental excess or to the infant’s response to exces-
sive frustration. The result, whether genetic or environmental, was
too much anger, which caused further problems, such as splitting
(Criterion 1) and self-destructive behaviors (Criterion 7).
The patient’s anger can be discovered by the clinician’s taking a
history or actively inquiring about anger (many borderline patients
are aware of feeling angry much of the time, even though they rarely
express it); sometimes the anger becomes more apparent after cessa-
tion of a patient’s acting-out behaviors (Criteria 5 and 7), which have
defended against this feeling.
Impulsivity
6. Impulsivity. This criterion evolved out of the early literature describ-
ing the problems within psychotherapies of acting out as a resistance
to, or flight from, feelings and conflicts. Empirical studies then found
that the impulsivity of borderline individuals is to some extent differ-
ent from that found in manic/hypomanic or antisocial patients by
virtue of its being self-damaging. Thus, the person with BPD who is a
substance abuser would be likely to relapse if angry at his or her Al-
coholics Anonymous sponsor or because of that sponsor’s absence or
unavailability. This one criterion, impulsivity, provides a way of incor-
porating as symptoms what are otherwise considered distinct disor-
ders (e.g., bulimia and substance abuse). It is not uncommon for
borderline patients to substitute one impulse pattern for another—
for example, exchanging cutting for purging for abusing drugs. As
noted elsewhere, the impulsivity of borderline patients has been con-
sidered a basic temperamental disposition and has linked BPD to an-
tisocial personality disorder (see Chapter 2).
7. Suicidal or self-mutilating behaviors. Recurrent suicidal attempts, ges-
tures, or threats or self-mutilating behaviors are the borderline pa-
tient’s behavioral specialty. This criterion is so prototypical of persons

with BPD that the diagnosis rightly comes to mind whenever recur-
rent self-destructive behaviors are encountered. The presence of this
pattern helps signal concurrent BPD in patients whose presenting
symptoms are depression or anxiety (R. C. Friedman et al. 1983; S.
Friedman et al. 1992). Clinicians must differentiate threats or acts
that communicate a cry for help from those that have other motiva-
tions. Responding to cries for help as if they were suicidal gestures
probably contributed to the idea that this was a postprognosis disor-
der (Sidebar 1–3). The clinical importance of this criterion is de-
scribed in more detail later in this chapter (see section “How to
Explain the Diagnosis”).
Sidebar 1–3: Borderline Personality as

an Iatrogenic Disorder
Joe Triebwasser had only recently completed his psychiatry
residency training at McLean when he wrote a paper bear-
ing the title of this sidebar. The paper described instances in
which well-meaning therapists and staffs of inpatient ser-
vices were unwittingly encouraging patients to get atten-
tion and intensive levels of care by responding with
disproportionate and undue alarm to the red flag of self-
harm. As a psychiatrist responsible for administration for
these patients, Triebwasser became impatient with the sub-
sequent problems of trying to undo the effects of these na-
ive interventions. He proposed that troubled adolescents or
young adults were transformed into patients with diagnos-
able borderline personality by iatrogenic processes.
Drawing on the evidence that the natural course of BPD
is far more benign than expected and that well-designed
treatments are beneficial for most borderline patients, sev-
eral recent reviews have independently proposed that
much of the earlier literature concerning the failures or over-
whelming difficulties in treating BPD were a result of treat-
ments that worsened the symptoms. Fonagy and Bateman
(2006) argued that psychosocial treatments have often “im-
peded the borderline patient’s capacity to recover... and
prevented them from harnessing advantageous changes in
their social circumstances” (p. 2). They suggested that, in
part, the better prognosis observed with modern borderline
cohorts is a result of “harmful treatments being less fre-
quently offered.”
Gunderson noted that the numerous books on psycho-
analytic therapy that were written in the 1980s largely dwell
on the intense—sometimes psychotic—transferences, the
problems with noncompliance and maintaining bound-

aries, the potential for excessive aggression, and the seem-
ingly ubiquitous extreme countertransference reactions
(and the danger of their enactment) (Gunderson and Links
2007; Gunderson et al. 2007). Very little documentation of
benefits is found. In retrospect, the problems these books
detail may be iatrogenic—caused by therapies that were
uninformed or infused by unrecognized countertransfer-
ence reactions.
It is reassuring that the frequency with which mental
health services now reinforce self-harm behaviors has
greatly diminished as a result of better awareness about
these dangers. What remains disturbing, however, is that
borderline psychopathology exists without our help and for
reasons that often remain difficult to reverse.
8. Identity disturbance. This criterion is derived from Kernberg’s descrip-

tion of borderline personality organization (see earlier in this chapter,
section “Shifts in the Borderline Construct”). Since DSM-III, this cri-
terion has undergone modifications intended to differentiate it from
the generic identity issues that are normal parts of development, most
notably adolescence. The identity disturbance criterion is meant to
encompass the body image distortions seen in persons with anorexia
or body dysmorphic disorder; more important, it is meant to recog-
nize the pathological disorders of self that are more specific to border-
line patients—that is, adults whose values, habits, and attitudes are
dominated by whomever they are with to the extent that they feel they
have no identity. Here, too, the interpersonal context for these iden-
tity problems associates this criterion with early attachment failures.
9. Lapses in reality testing. The criterion reflecting the issue that was
added in DSM-IV is a sharply demarcated derivative of the earlier
clinical literature that spoke of psychotic transferences and the
potential for psychotic regressing within unstructured treatment set-
tings (e.g., see Hoch and Polatin 1949; Knight 1953). Indeed, specu-
lation about the possible relation of borderline psychopathology to
schizophrenia was fueled by concerns about psychotic regression in
unstructured settings as disparate as Rorschach testing and psycho-
analysis.
John Frosch (1964) refined the description by distinguishing
lapses in a sense of reality (not knowing whether one’s experience is
real) from a generally intact ability to test reality (being able to correct
distortions of reality with feedback). The lapses in sense of reality that
typify borderline patients involve depersonalization, derealization,
and hallucinogenic phenomena (e.g., “I thought I heard my

mother’s voice, so I turned on the lights”). The ability to correct real-
ity distortions would be exemplified by the fact that the patient can
turn the lights on and be reassured by the fact that her mother is not
there. The addition of this criterion also ties the phenomenology of
BPD to its possible pathogenesis insofar as such lapses in reality test-
ing can be understood as sequelae of childhood neglect and abuse
(Shearer 1994a; Silk et al. 1995).
Because the diagnostic system has traditionally set apart psychoses
as major mental illnesses, there were reservations about introducing
criteria suggesting transient lapses. If such lapses were present, they
seemed when DSM-III was written to fit better with preconceptions
that a reality-testing criterion was more consistent with the schizo-
typal personality disorder construct than with BPD. The accumulated
data from many studies (Table 1–3) were needed to document the
presence of such phenomena in samples of patients with BPD.
TABLE 1–3. Prevalence of cognitive/perceptual symptoms in

borderline personality disorder samples
COGNITIVE/PERCEPTUAL PROBLEM STUDYa RANGE (%)
Depersonalization 1, 3, 5, 7, 8, 9 30–85
Derealization 1, 3, 4, 7, 8, 9 30–92
Paranoid experiences 3, 5, 6, 7, 8, 9 32–100
Visual illusions 3, 7, 8 77–88
Muddled thinking 4 52
Magical thinking 5, 6, 9 34–68
Ideas of reference 5, 6, 9 49–74
Odd speech 5, 6 30–59
Disturbed thoughts 2, 9 39–68
aNumbers
in this column refer to the following studies: 1= Frances et al. 1984;
2= Pope et al. 1985; 3= Chopra and Beatson 1986; 4 =George and Soloff 1986;
5= Jacobsberg et al. 1986; 6 = Widiger et al. 1987; 7 =Links et al. 1988; 8 = Silk et al.
1989; and 9 =Zanarini et al. 1990.
Source. Reprinted from Gunderson JG, Zanarini MC, Kisiel CL: “Borderline Per-
sonality Disorder,” in DSM-IV Sourcebook, Vol 2. Edited by Widiger TA, Frances AJ,
Pincus HA, et al. Washington, DC, American Psychiatric Association, 1996, p.
725. Used with permission. Copyright 1996 American Psychiatric Association.
A Clinical Synthesis: Intolerance of Aloneness

Aloneness is experienced as a terrifying loss of self (Criterion 3) that the
person with BPD may defend against by action (Criterion 2) or by distort-
ing reality (Criterion 9). Aloneness also can be diminished either by the use
of transitional objects (discussed in Chapter 12) or by another person’s pro-
viding reassuring evidence that he or she cares for the person with BPD.
Identification of intolerance of being alone as one of the defining cri-
teria for the diagnosis of BPD can be traced to the clinical and theoretical
contributions of Modell (1963), Winnicott (1965), and Masterson (1972,
1976; Masterson and Rinsley 1975). Modell posited that borderline pa-
tients’ basic developmental failure involved an inability to cope with the
separateness of their caregivers—what Winnicott (1953) had defined as
transitional relatedness. Masterson emphasized the abandonment fears of
borderline patients and the origins of these fears in traumatic childhood
separation experiences. The DIB operationalized this trait and estab-
lished it as one of the more discriminating features of the disorder (Gun-
derson and Kolb 1978). The inability to conjure up representations of
absent others (object inconstancy) was subsequently emphasized by
Adler and Buie (1979). This intolerance of aloneness and this object in-
constancy have been empirically confirmed (Richman and Sokolove
1992). The reason that the BPD diagnosis, so prevalent in clinical settings
and now so much a part of the mental health world’s consciousness,
failed to be identified earlier is, I believe, because the presenting phe-
nomenology is extremely dependent on the interpersonal context. This
formulation has been given empirical support in work by Perry and Coo-
per (1986) and more extensively by Benjamin (1993).
These clinical and conceptual characterizations of adult borderline
patients led to the development of the diagnostic criteria, but various
British child analysts have continued to explicate the childhood experi-
ences that illuminate the pathogenesis of BPD (Sidebar 1–4).
Sidebar 1–4: British Developmentalists:

From Winnicott to Bowlby to Fonagy
Three British child analysts have lent their clinical and theo-
retical contributions to the understanding of the develop-
ment of borderline psychopathology.
D.W. Winnicott, originally a pediatrician, distinguished
himself in the 1950s by his keen clinical observations, by his
personal charisma, and by his creative conceptualization
of key concepts such as the holding environment and tran-
sitional objects. The concept of the analyst’s role as hold-
ing—that is, serving as a container for—the borderline

patient’s aggression has been widely used to understand
the functions served by hospitals as well as to understand
the need for the clinician’s accepting the patient’s hostili-
ties without withdrawing. The concept of transitional ob-
jects has been picked up within object relations theory and
developmental psychology. Modell (1963) used this con-
cept to describe the function that borderline patients need
their therapists to serve—that is, as if the therapists were ex-
tensions of their patients who lacked separate identities or
feelings. This contribution later generated a round of em-
pirical investigations (see Sidebar 12–3).
Bowlby’s (1969, 1973, 1980, 1988) primary contributions
started to gain influence in the 1970s. He was a child psy-
chiatrist, analytically trained, who systematized his obser-
vations of children and thereby established an empirically
based and observer-friendly scheme for child develop-
ment organized around the acquisition of secure attach-
ments. He proposed that all infants possess a basic instinct
toward attachment to caregivers. Darwinian adaptations
for survival impel infants to evolve interpersonal behaviors
that function to maintain their proximity (availability and re-
sponsiveness) to a caregiver. Caregiver proximity is re-
quired for the development of internal feelings (i.e., by
introjection or internalization) of security and lovability.
Children whose early attachments were insecure become
adults whose interpersonal behavioral adaptations are de-
veloped in response to inconsistent, absent, or frustrating
caregivers. Ainsworth et al. (1978) later developed the
Strange Situation experiment to assess toddlers’ response
to separation from their caregivers. They operationalized
subtypes of insecure attachment. One pattern called anx-
ious/ambivalent includes the need to check for caregiver
proximity, signaling the need to establish contact by
pleading or other calls for attention or help, and clinging
behaviors. This pattern can alternate unpredictably with a
different subtype of attachment pattern called disorga-
nized/disoriented (Lyons-Ruth and Jacobvitz 1999; Main
and Solomon 1990), consisting of the denial of dependent
needs, the apparent absence of separation anxiety, and a
reluctance or fearfulness about becoming attached. Such
behaviors, intermittently present in many patients with BPD,
develop in response to primary caregivers who are de-
pressed, disturbed, or abusive (Crittendon 1988; Main and
Hesse 1990)—qualities that unfortunately are common in
the childhood caregivers of many borderline patients
(Gunderson and Zanarini 1989; Links 1990). Many clinician-
scientists believe that this alternating attachment pattern is
the core psychopathology for borderline patients (Adler
and Buie 1979; Benjamin 1993; Fonagy 1991; Gunderson

1984, 1996; Perry and Cooper 1986).
Fonagy’s influence is only beginning to be felt, and,
although it is premature to place him alongside Winnicott
and Bowlby, his work extends their contributions in creative
and significant ways (Fonagy 1991; Fonagy et al. 1995a,
1995b). Fonagy has added specificity and detail to the
parent-child interactions that beget those anxious/ambiv-
alent or disorganized/disoriented forms of insecure attach-
ment typical of borderline patients. Specifically, Fonagy et
al. (1991) used the Strange Situation experiment to identify
parental misunderstandings (i.e., misattributions) of the
child’s internal states. Thus, a caregiver who misperceives a
frightened child and labels the feelings as angry, or who
misconceives a child’s normal attention seeking as being
demanding, will respond in ways that impair the child’s
capability to develop stable and realistic concepts of
self—that is, to capably “mentalize” his or her own and oth-
ers’ intentions, desires, or feelings. Fonagy (1991; Fonagy et
al. 1995a, 2000) proposed that BPD develops in children
who acquire only a limited capacity to depict feelings and
thoughts in themselves and in others. His work essentially
begins to chart the interactive processes of early child-
hood that affect attachments and may have a neurobio-
logical base. Fonagy’s approach builds bridges between
cognitive and dynamic theories and has provided a base
for an empirically validated form of therapy for borderline
patients (see Chapter 5, section “Level III: Residential/Par-
tial Hospital Care/Day Treatment—Basic Socialization”;
and Chapter 12 on psychodynamic psychotherapies).
Table 1–4 shows how borderline patients’ feelings of being securely

held, threatened by separation, or alone in relation to their primary ob-
ject (i.e., their needed other) account for changes in their clinical phe-
nomenology.
When the person with BPD feels cared for, “held,” he or she appears
like a depressed waif—easy to sympathize with, grateful for signs of care,
and, like a healthier neurotic patient, receptive to therapeutic interpre-
tations. Symptoms such as depression, eating disorders, substance abuse,
or PTSD often become the focus of therapy.
When the person with BPD is confronted with the potential loss of the
caring, holding other, a different set of clinical phenomena becomes ev-
ident—phenomena that link the theme of intolerance of aloneness to
the DSM-IV criteria for BPD. Now, prompted by fears of abandonment,
the angry devaluation or the self-injurious behaviors become apparent,
often with unexpected suddenness and intensity. The self-injurious be-
TABLE 1–4. How borderline personality disorder (BPD) patients’

perceived attachment relates to BPD
phenomenology
INTERPERSONAL OTHERS’
CONTEXT PHENOMENOLOGY RESPONSES CLINICAL IMPLICATION
Held/ Empty, Sympathetic Collaborative;
idealizing dysfunctional, interpretations;
symptomatic patient needs ex-
pressive, involv-
ing therapies
Threatened/ Angry, self- Scared, Confrontations;
devaluing destructive guilty, patient needs
pleas for help angry social supports,
behavior change
Alone Terrified, Rescue, Words unimportant;
dissociated, avoid patient needs
paranoid, containment,
substance- medications
abusing,
promiscuous
Source. Adapted from Gunderson 1984.
haviors may be potentially lethal, indicating a readiness to die unless the

person whose absence is threatened, or someone new, establishes that he
or she wants the person with BPD to live. Often, this involves not leaving
the person with BPD or otherwise rescuing him or her. This sets in mo-
tion a characteristic dilemma in the other: the other feels guilty about any
impending separation from the person with BPD but finds the prospect
of staying very distasteful.
When the person with BPD feels that he or she is without a caring
other, without a holding environment, a third set of clinically significant
phenomena becomes evident. The experience of aloneness leads to a loss
of a sense of reality (dissociative or hallucinogenic symptoms) or to para-
noid ideation (which conjures up a malevolent other, a situation prefer-
able to being alone). Alternatively, the experience of aloneness is
obviated by desperate object-seeking behaviors (e.g., promiscuity), often
made possible by the disinhibiting influence of alcohol or other drugs.
This formulation of intolerance of aloneness as the central or core
psychopathology of borderline patients contrasts with dynamic formula-
tions that give equal emphasis to borderline patients’ fears of too much
closeness—that is, fear of “fusion” (Lewin and Schulz 1992). In my view,
when fear of fusion is equal to or greater than fears of aloneness, the pa-
tient is more likely to have predominantly schizoid or narcissistic psycho-
pathology. The formulation offered here is consistent with Fairbairn’s
(1963) thesis (and subsequently Bowlby’s) that humans have an innate
drive for attachment; they are biologically object seeking.
It is only by longitudinal and interpersonally focused observations
that these changing phenomena become evidence of a single underlying
pathological process. Descriptive psychiatry has been too cross-sectional
and too distant to see the interpersonal patterns. Psychoanalytic psychia-
try has been too single-case-based and interpersonally intimate to identify
the phenomenological pattern.
Misuses of the Borderline Diagnosis

Controversies persist within the mental health community about the bor-
derline diagnosis. It is easy and not uncommon to misuse this diagnosis,
and this possibility remains in large measure a result of the emotional re-
sponses such patients engender.
There are reasons for the overuse of the diagnosis, starting with the
breadth of Kernberg’s construct of borderline personality organization and
the value that his conceptualization retains for psychodynamic therapists.
Notwithstanding the merits of his contribution, a deep skepticism exists
within the psychoanalytic community about defining diagnoses by exter-
nal, observable (read “superficial”) phenomena. Mental health profession-
als, whether analysts or not, whose primary identity lies in doing dynamic
therapy may still use the borderline diagnosis for all “primitive characters”
who show immature defenses such as projection and acting-out.
Occasionally, overuse can even come from the managed care environ-
ment. Little time is afforded for extended evaluations, and clinicians must
identify diagnoses early on to justify their costs. From this perspective, it is
convenient—as well as usually correct—to identify anyone who has carried
out repeated self-destructive acts or who is an inappropriately angry young
woman as “301.83” (the DSM-IV diagnostic code number for BPD).
One reason for underuse of the diagnosis parallels the first source of
overuse noted earlier. Some psychiatrists believe that the foundations of
our diagnostic system should be more biologically based than is the bor-
derline diagnosis. They believe that dynamic considerations are superfi-
cial and that the major therapeutic importance of diagnosis is to guide
pharmacotherapies. This perspective can be inferred from the prolifer-
ating studies of bipolar II disorder, in which “comorbidity” (or overlap
with) BPD has not, to my knowledge, yet been assessed. Offsetting this
tendency, the borderline diagnosis remains useful for most biological
psychiatrists because of its role in explaining mood disorders that prove

resistant to medications.
A second source of underuse is more subtle. Distinguished mental
health professionals have suggested that the borderline diagnosis is pejo-
rative. Vaillant (1992) argued that the diagnosis is primarily used by cli-
nicians to label patients they do not like (Sidebar 1–5). Clinical leaders
from the Stone Center argued that the label misleadingly conveys that
the patients are angry and manipulative and that it therefore interferes
with a clinician’s empathic availability for patients who are often better
conceptualized as trauma victims (Jordan et al. 1991; Stiver 1991). They
are joined by Heller (1991), who argued that because the borderline la-
bel “implies a character problem,” it causes doctors and therapists to
shun patients with BPD rather than provide proper—meaning in Heller’s
view pharmacological—therapies. Both the Stone Center and Heller
agree that the BPD label assigns too much accountability for socially un-
desirable behaviors.
Sidebar 1–5: “Wisdom Is Never Calling

a Patient Borderline”
Beginning in 1974, George Vaillant entertained audiences
with a talk by this name that eventually found its way to
publication (Vaillant 1992). His thesis that clinicians use the
label borderline for patients they do not like captures an
unhappy truth—most clinicians do not like borderline pa-
tients. More to the point, most clinicians do not like patients
who are angry, critical, rejecting, mocking, or even con-
temptuous toward them. Vaillant is right. Such attitudes do
not warrant being diagnosed as borderline. Disliking a pa-
tient—that is, a hostile countertransference—is not a rea-
son to make the borderline diagnosis; it is a reason to
understand one’s reaction.
There is an alarming tendency for clinicians who are working within

institutional settings—for example, hospitals and outpatient clinics—to
underuse the borderline diagnosis. Zimmerman and Mattia (1999)
showed that clinicians in a private practice group at the Rhode Island
Hospital outpatient psychiatry clinic diagnosed BPD in only 0.4% of the
patients, whereas the frequency rose to 14.4%—36 times as great—when
a similar sample of patients was given structured interviews. The authors
argue that clinicians typically give priority to Axis I diagnoses and treat-
ment, leaving insufficient time for Axis II assessments. Consistent with
this conclusion is the much lower use of the BPD diagnosis in state men-
tal health facilities than in nonstate facilities (Oldham and Skodol 1991).
I believe this conclusion is true but that the underuse may have even
more to do with a strong bias toward diagnosing and offering treatments
for only what managed care payers and biological psychiatrists deem
treatable.
The Behavioral Specialty: Self-Injurious Behavior

When John Mack (1975) called for a “behavioral specialty” to establish the
borderline diagnosis, self-injurious behaviors offered a vivid and distinctive
exemplar. Such behaviors are found in about 75% of borderline patients
(Clarkin et al. 1983; Gardner and Cowdry 1985; Gunderson 1984; Zisook
et al. 1994). The frequency with which self-destructive behaviors occur (e.g.,
unprotected sexual intercourse with strangers, drinking while taking An-
tabuse) would increase this rate into the 90% range. Self-injurious behav-
iors, most often cutting (80%) but also frequently bruising (24%), burning
(20%), head banging (15%), and biting (7%) (Shearer 1994b), are the
most common symptoms by which people with BPD come to clinical atten-
tion. For most people, and certainly for school counselors, clergy, friends,
and family, the evidence of willful self-harm is an alarming indication of
suicidal intentions.
Repeated self-destructive acts by any patient should alert clinicians to
the fact that these acts may not be suicidally intended. Many self-destruc-
tive acts are done for self-punitive reasons (Shearer 1988) and are some-
times associated with an experience of relief from painful or intolerable
affective states (Soloff et al. 1994). Moreover, these acts are also per-
formed with progressively greater awareness of the controlling effects
that such acts have on significant others. Indeed, the power that self-
mutilative behaviors have in drawing attention and concern from others
is probably the reason that they can become contagious in adolescents.
However, the relation of self-injurious behavior to suicidality is complex.
These behaviors may be done with real suicidal intentions. According to
Stone et al. (1987), BPD patients with self-injurious behaviors are at in-
creased risk for suicide attempts—especially when they are associated
with higher levels of depression, hopelessness, and impulsivity. Stanley et
al. (2001) reported that a history of self-injurious behavior, regardless of
the intention, actually doubles the likelihood of suicide. Borderline pa-
tients with self-injurious behavior can misperceive and underestimate the
lethality of their acts. Often, the patient will relate his or her self-injurious
behaviors to multiple intentions, including a wish to die (Santa Mina et
al. 2006). The common intentions of self-injurious behavior are listed in
Table 1–5.
TABLE 1–5. Functions of self-injurious behavior

FUNCTION % OF PATIENTS
To feel physical pain—to overcome psychic pain 59
To punish self for being “bad” 49
To control feelings 39
To exert control 22
To express anger 22
To feel—to overcome numbness 20
Source. Adapted from Shearer SL: “Phenomenology of Self-Injury Among Inpa-
tient Women With Borderline Personality Disorder.” Journal of Nervous and Mental
Diseases 182(9):524–526, 1994. Used with permission.
A patient struggling with her impulse to cut wrote the following:
I want to cut. I want to see pain, for it is the most physical thing to show.
You can not show pain inside. I want to cut, cut, show, show. Get it out.
What out? Just pain.
It is clearly here an expression of pain intended to be seen and re-

sponded to. In recent years, this wish to communicate one’s pain by self-
mutilative behaviors has occasionally rendered this symptom socially con-
tagious (Sidebar 1–6). Yet it is unsafe ever to assume that self-mutilative
behaviors are merely attention-getting. Borderline patients do commit
suicide, often under circumstances that may have begun as a gesture but
in which they have miscalculated the response of those from whom a sav-
ing response was expected. Borderline patients will recount after serious
overdoses that they were fully aware that they could die and that they
were knowingly placing their fate in the hands of chance. Moreover, hav-
ing once, or ever, made an actual suicide attempt greatly increases the
likelihood of later suicide. However, clinicians must remember that de-
spite the high frequency with which borderline patients perform multi-
ple self-destructive acts, the comparative frequency of acts that result in
actual suicide is low (Soloff et al. 1994; Stone 1990).
Chapter 4 discusses the “acute-on-chronic model” to assist the clini-
cian in assessing suicide risk and resumes a discussion of the clinical man-
agement issues surrounding self-injurious and suicidal behavior. Chapter
11 describes several cognitive-behavioral therapies that are specifically tar-
geted at diminishing such behaviors. Chapter 8 offers suggestions to fam-
ilies and other nonprofessionals about how they can respond helpfully.
Sidebar 1–6: Cutting: Social Contagion or

Psychopathology?
Favazza (1996) has pioneered a scholarly examination of
the historical and cultural context for self-mutilation. He intro-
duces his book by speculating that the remarkable lack of
attention given to this widespread and long-standing as-
pect of human behavior reflects the horror and shameful
fascination such behaviors hold. He notes that cutting was
identified as the work of evil spirits by Jesus, who performed
an exorcism, and that ritualistic cutting has been docu-
mented in various religions or other subcultural practices
since the thirteenth century. He estimated that about 2 mil-
lion Americans self-mutilate (about 0.75%). Favazza devel-
oped a thesis that whether self-mutilative acts are culturally
sanctioned or are the products of personal anguish, they of-
ten help relieve pain and that their adaptive, life-enhancing
qualities deserve recognition.
More recently, cutting has received more of its over-
due attention. In the July 27, 1997, New York Times Maga-
zine, Jennifer Egan (1997) suggested that cutting is an
extreme expression of the same impulses that are making
tattoos and piercing contagion among modern adoles-
cents. Indeed, she noted that several communities have
witnessed an epidemic of cutting among adolescent girls.
Still more recently, a series of epidemiological studies have
found that cutting occurs in 2%–14% of the general popu-
lation and that this is part of a wider escalation of deliber-
ate self-injurious behavior (Favaro et al. 2007; Gratz et al.
2002; Muehlenkamp and Gutierrez 2004).
The danger of this sociological perspective is that it
could minimize the personal and clinical significance at-
tached to cutting by those who do it repeatedly and when it
is usually a private and highly shameful act. For adolescents
or young adults who develop BPD, cutting usually begins as
a private act of desperation, reflecting either an inability to
communicate in words or a call for help. Not all those who
self-mutilate have BPD, but many do, and all need to be
taken seriously. Studies have shown that about one-third of
the women with BPD began cutting before age 13 (Zanarini
et al. 2006). Because the diagnosis of BPD underscores a se-
rious, long-standing mental health problem, the diagnosis
should not be offered reflexively to anyone who cuts or oth-
erwise self-mutilates. However, the diagnosis should never be
excluded because “we didn’t want to believe it was seri-
ous.” The latter response is likely to evoke further alienation
and more serious acts of self-destruction in the subgroup of
cutters who are vulnerable to BPD.
Use of the Diagnosis in Adolescents

The diagnosis of BPD is not recognized for children or adolescents in
DSM-IV-TR (American Psychiatric Association 2000). Yet the diagnosis is
increasingly used in adolescents, and adaptations of an adult therapy, di-
alectical behavior therapy, appear to be helpful (Rathus and Miller
2002). Adolescent psychiatrists believe that the diagnosis can be made
with considerable confidence. Impulsive patterns are sufficiently com-
mon in adolescents that they have less weight for the borderline diagno-
sis than in adults. However, as noted in Sidebar 1–6, repeated self-
injurious behavior, especially if private, is a strong risk marker. The crite-
ria of affective instability (Becker et al. 2002) and the pattern of seeking
or needing exclusive friendships or romantic relationships are also risk
markers. The bottom line is that adolescents seeking or coming for help
can and should be given the BPD diagnosis if their symptoms meet the
diagnostic criteria.
How to Explain the Diagnosis

Two common issues are whether someone has the diagnosis of BPD and
what causes it to arise. Someone might ask whether he or she (or his or
her relative) has BPD. When asked these questions, it is useful to be able
to answer in ways that are relatively jargon free, allowing patients, their
families, or other laypersons to raise questions and reach their own con-
clusions.
The following exemplifies an answer to the question about whether
someone has the diagnosis:
People with BPD are born with a genetic disposition to be emotional,

have low frustration tolerance, and be very sensitive to signs of rejection.
They have grown up feeling that they were unfairly treated and that they
did not get the attention or care they needed. They are angry about that,
and as young adults, they set out in search of someone who can make up
to them for what they feel is missing. When they think they have found
such a person, they set in motion intense, exclusive relationships, which
predictably will fail because they place unrealistic expectations on the
other person. Upon failing, they feel rejected or abandoned, and either
their rage about being treated unfairly gets reawakened or they feel they
are bad and deserved the rejection, in which case they become suicidal or
self-destructive. Sometimes, their anger about being mistreated causes
others to feel guilty, and sometimes their self-destructiveness evokes pro-
tective feelings in others. Such guilty or rescuing responses from others
validate the borderline person’s often unrealistically negative perceptions
of mistreatment and encourages their unrealistically high expectations of
having their needs met. Thus, the cycle is apt to repeat itself.
With respect to the second question about what caused the person to
have BPD, the causes are still poorly understood, so the response needs to
convey this complexity:
The cause of BPD is not fully understood, but we know it involves multiple
factors. Like all other major psychiatric disorders, BPD arises when an in-
dividual with a genetic predisposition is exposed to environmental stress-
ors. Although the genetic predisposition is still being researched, we
believe that this involves three personality dimensions, each of which has
multiple genes. These dimensions, called phenotypes, involve affective
(emotional) instability, impulsivity, and interpersonal (rejection) hyper-
sensitivity. The environmental stressors that lead to the diagnosis are
highly variable from one individual to another; however, for many indi-
viduals, histories of neglect or trauma during childhood are highly rele-
vant. Research is beginning to tie the three personality dimensions or
phenotypes to neurobiological pathways. Neuroimaging studies suggest
that the emotional gateway within the brain, the amygdala, is overly ac-
tive, whereas the normal inhibitory system within the brain, the prefron-
tal cortex, is hypoactive. Although much more research needs to be done,
no one cause is adequate to explain the diagnosis of BPD.
Summary
That patients fulfill criteria for the borderline syndrome is well estab-
lished, and the use of the diagnosis has become more uniform and uni-
versal. The meaning of the diagnosis is still undergoing revision as greater
specificity is added to our understanding of the etiology and pathogenesis
of this disorder. A basic thesis of this book is that the diagnosis already car-
ries great specificity in terms of treatment but that a great deal of exper-
tise is required to provide such treatment well, whereas uninformed
treatment is very easy to do harmfully. With the emergence of this diagno-
sis as a valid and widely recognized entity, it is important that clinicians be-
gin using the diagnosis openly with patients and families. A way to do so
has been presented here. This chapter’s larger message is that it is highly
useful to be explicit and unapologetic in making this diagnosis and that
to do otherwise is often a product of our countertransference feelings
about such patients.
References
Abraham K: Selected Papers. London, Hogarth Press, 1927
Adler G: Hospital treatment of borderline patients. Am J Psychiatry
130:32–35, 1973
Adler G, Buie D: The psychotherapeutic approach to aloneness in bor-
derline patients, in Advances in Psychotherapy of the Borderline Pa-
tient. Edited by LeBoit J. New York, Jason Aronson, 1979, pp 433–448
Aichhorn A: Wayward Youth (1925). New York, Viking, 1945

Ainsworth MDS, Blehar MC, Waters E, et al: Patterns of Attachment: A Psy-
chological Study of the Strange Situation. Hillsdale, NJ, Erlbaum, 1978
Akiskal HS: Subaffective disorders: dysthymic, cyclothymic and bipolar II
disorders in the “borderline” realm. Psychiatr Clin North Am 4:25–
46, 1981
Akiskal HS: The nosologic status of borderline personality: clinical and
polysomnographic study. Am J Psychiatry 142:192–198, 1985
Alexander F: The neurotic character. Int J Psychoanal 11:292–311, 1930
American Psychiatric Association: Diagnostic and Statistical Manual of
Mental Disorders, 3rd Edition. Washington, DC, American Psychiat-
ric Association, 1980
Mental Disorders, 3rd Edition, Revised. Washington, DC, American
Psychiatric Association, 1987
Mental Disorders, 4th Edition. Washington, DC, American Psychiat-
Mental Disorders, 4th Edition, Text Revision. Washington, DC,
American Psychiatric Association, 2000
Balint M: The Basic Fault: Therapeutic Aspects of Regression. Evanston,
IL, Northwestern University Press, 1992
Becker DF, Grilio CM, Edell WS, et al: Diagnostic efficiency of borderline
personality disorder criteria in hospitalized adolescents: comparison
with hospitalized adults. Am J Psychiatry 159:2042–2047, 2002
Benjamin LS: Interpersonal Diagnosis and Treatment of Personality Dis-
orders. New York, Guilford, 1993
Bongar B, Peterson LG, Golann S, et al: Self-mutilation and the chroni-
cally “suicidal” emergency room patient. Ann Clin Psychiatry 2:217–
222, 1990
Bowlby J: Attachment and Loss, Vol 1: Attachment. New York, Basic
Books, 1969
Bowlby J: Attachment and Loss, Vol 2: Separation: Anxiety and Anger.
New York, Basic Books, 1973
Bowlby J: Attachment and Loss, Vol 3: Loss. New York, Basic Books, 1980
Bowlby J: A Secure Base: Parent-Child Attachment and Healthy Human
Development. New York, Basic Books, 1988
Chopra HD, Beatson JA: Psychotic symptoms in borderline personality
disorder. Am J Psychiatry 143:1605–1607, 1986
Clarkin JF, Widiger TA, Frances A, et al: Prototypic typology and the bor-
derline personality disorder. J Abnorm Psychol 93:263–275, 1983
Coryell WH, Zimmerman M: Personality disorder in the families of de-
pressed, schizophrenic, and never-ill probands. Am J Psychiatry
146:496–502, 1989
Crittendon PM: Relationships at risk, in Clinical Implications of Attach-

ment. Edited by Belsky J, Nezworski T. Hillsdale, NJ, Erlbaum, 1988,
pp 136–174
Dahl AA: Some aspects of the DSM-III personality disorders illustrated by
a consecutive sample of hospitalized patients. Acta Psychiatr Scand
328(suppl):61–67, 1986
Egan J: The thin red line. New York Times Magazine, July 27, 1997, pp
21–25, 39–44, 48
Ellison J, Barsky A, Blum NR: Frequent repeaters to an emergency ser-
vice. Hosp Community Psychiatry 40:958–960, 1989
Fairbairn WRD: Synopsis of an object-relations theory of the personality.
Int J Psychoanal 44:224–225, 1963
Favaro A, Ferrara S, Santonastaso P: Self-injurious behavior in a commu-
nity sample of young women: relationship with childhood abuse and
other types of self-damaging behaviors. J Clin Psychiatry 68:122–131,
2007
Favazza M: Bodies Under Siege, 2nd Edition. Baltimore, MD, Johns Hop-
kins University Press, 1996
Fonagy P: Thinking about thinking: some clinical and theoretical consid-
erations in the treatment of a borderline patient. Int J Psychoanal
72(Pt 4):639–656, 1991
Fonagy P, Bateman A: Progress in the treatment of borderline personality
disorder. Br J Psychiatry 188:1–3, 2006
Fonagy P, Steele H, Moran G, et al: The capacity for understanding men-
tal states: the reflective self in parent and child and its significance
for security of attachment. Infant Ment Health J 13:200–217, 1991
Fonagy P, Leigh T, Kennedy R, et al: Attachment, borderline states and
the representation of emotions and cognitions in self and other, in
Emotion, Cognition, and Representation (Rochester Symposium on
Developmental Psychopathology, Vol 6). Edited by Cicchetti D, Toth
SL. Rochester, NY, University of Rochester Press, 1995a, pp 371–414
Fonagy P, Steele M, Steele H, et al: Attachment, the reflective self, and
borderline states: the predictive specificity of the Adult Attachment
Interview and pathological emotional development, in Attachment
Theory: Social, Developmental, and Clinical Perspective. Edited by
Goldberg S, Muir R, Kerr J. Hillsdale, NJ, Analytic Press, 1995b, pp
223–278
Fonagy P, Target M, Gergely G: Attachment and borderline personality
disorder: a theory and some evidence. Psychiatr Clin North Am
23:vii–viii, 103–122, 2000
Frances A, Clarkin J, Gilmore M, et al: Reliability of criteria for border-
line personality disorder: a comparison of DSM-III and the Diagnos-
tic Interview for Borderline Personality Disorder. Am J Psychiatry
141:1080–1084, 1984
Freud S: Character and erotism (1908), in Standard Edition of the Com-

plete Psychological Works of Sigmund Freud, Vol 9. Translated and
edited by Strachey J. London, Hogarth Press, 1959, pp 169–175
Friedman RC, Aronoff MS, Clarkin JF, et al: History of suicidal behavior
in depressed borderline inpatients. Am J Psychiatry 140:1023–1026,
1983
Friedman S, Jones JC, Chernen L, et al: Suicidal ideation and suicide at-
tempts among patients with panic disorder: a survey of two outpa-
tient clinics. Am J Psychiatry 149:680–685, 1992
Frosch J: The psychotic character: clinical psychiatric considerations. Psy-
chiatr Q 38:1–16, 1964
Gardner DL, Cowdry RW: Suicidal and parasuicidal behavior in border-
line personality disorder. Psychiatr Clin North Am 8:389–403, 1985
George A, Soloff P: Schizotypal symptoms in patients with borderline
personality disorders. Am J Psychiatry 143:212–215, 1986
Gratz KL, Conrad SD, Roemer L: Risk factors for deliberate self-harm
among college students. Am J Orthopsychiatry 72:128–140, 2002
Grilo CM, McGlashan TH, Oldham JM: Course and stability of personal-
ity disorders. Journal of Practical Psychiatry and Behavioral Health
1(4):61–75, 1998
Grinker R, Werble B, Drye R: The Borderline Syndrome: A Behavioral
Study of Ego Functions. New York, Basic Books, 1968
Gunderson JG: Borderline Personality Disorder. Washington, DC, Amer-
ican Psychiatric Press, 1984
Gunderson JG: Building structure for the borderline construct. Acta Psy-
chiatr Scand Suppl 379:12–18, 1994
Gunderson JG: The borderline patient’s intolerance of aloneness: inse-
cure attachments and therapist availability. Am J Psychiatry 153:752–
758, 1996
Gunderson JG, Kolb JE: Discriminating features of borderline patients.
Am J Psychiatry 135:792–796, 1978
Gunderson JG, Links P: Borderline personality disorder, in Gabbard’s
Treatments of Psychiatric Disorders, 4th Edition. Gabbard GO, Edi-
tor-in-Chief. Washington, DC, American Psychiatric Publishing,
2007, pp 805–821
Gunderson JG, Sabo AN: The phenomenological and conceptual inter-
face between borderline personality disorder and PTSD. Am J Psychi-
atry 150:19–27, 1993
Gunderson JG, Singer M: Defining borderline patients: an overview. Am
J Psychiatry 132:1–10, 1975
Gunderson JG, Zanarini MC: Pathogenesis of borderline personality dis-
order, in American Psychiatric Press Review of Psychiatry, Vol 8. Ed-
ited by Tasman A, Hales RE, Frances AJ. Washington, DC, American
Psychiatric Press, 1989, pp 25–48
Gunderson JG, Carpenter WT, Strauss JS: Borderline and schizophrenic

patients: a comparative study. Am J Psychiatry 132:1257–1264, 1975
Gunderson JG, Kolb JE, Austin V: The Diagnostic Interview for Border-
line Patients. Am J Psychiatry 138:896–903, 1981
Gunderson JG, Siever LJ, Spaulding E: The search for a schizotype: cross-
ing the border again. Arch Gen Psychiatry 41:15–22, 1983
Gunderson JG, Zanarini MC, Kisiel CL: Borderline personality disorder,
in DSM-IV Sourcebook, Vol 2. Edited by Widiger TA, Frances AJ, Pin-
cus HA, et al. Washington, DC, American Psychiatric Association,
1996, pp 717–731
Gunderson JG, Bateman A, Kernberg O: Alternative perspectives on psy-
chodynamic psychotherapy of borderline personality disorder: the
case of “Ellen.” Am J Psychiatry 164:1333–1339, 2007
Heller LM: Life at the Border. Okeechobee, FL, Dyslimbia Press, 1991
Hoch P, Polatin P: Pseudoneurotic forms of schizophrenia. Psychiatr Q
23:248–276, 1949
Jacobsberg I, Hymowitz P, Barasch A, et al: Symptoms of schizotypal per-
sonality disorder. Am J Psychiatry 143:1222–1227, 1986
Jordan J, Kaplan A, Miller J, et al (eds): Women’s Growth in Connection:
Writings From the Stone Center. New York, Guilford, 1991
Kernberg O: Borderline personality organization. J Am Psychoanal Assoc
15:641–685, 1967
Kernberg O: The treatment of patients with borderline personality orga-
nization. Int J Psychoanal 49:600–619, 1968
Kernberg O: Borderline Conditions and Pathological Narcissism. New
York, Jason Aronson, 1975
Kety S, Rosenthal D, Wender P, et al: The types and prevalence of mental
illness in the biological and adoptive families of adopted schizo-
phrenics, in The Transmission of Schizophrenia. Edited by Rosen-
thal D, Kety S. New York, Pergamon, 1968, pp 345–362
Klein D: Psychopharmacology and the borderline patients, in Borderline
States in Psychiatry. Edited by Mack I. New York, Grune & Stratton,
1975, pp 75–92
Klein D: Psychopharmacological treatment and delineation of border-
line disorders, in Borderline Personality Disorders: The Concept,
the Syndrome, the Patient. Edited by Hartocollis P. New York, Inter-
national Universities Press, 1977, pp 365–384
Klein M: The Psychoanalysis of Children. London, Hogarth Press, 1932
Klein M: Notes on some schizoid mechanisms. Int J Psychoanal 27:99–
110, 1946
Knight RP: Borderline states. Bull Menninger Clin 17:1–12, 1953
Knight RP: Management and psychotherapy of the borderline schizo-
phrenic patient, in Psychoanalytic Psychiatry and Psychology. Edited
by Knight RP, Freidman CR. New York, International Universities
Press, 1954, pp 110–122
Koenigsberg HW, Kaplan RD, Gilmore MM, et al: The relationship be-
tween syndrome and personality disorder in DSM-III: experience
with 2,464 patients. Am J Psychiatry 142:207–212, 1985
Kroll J, Carey K, Sines L, et al: Are there borderlines in Britain? A cross-
validation of US findings. Arch Gen Psychiatry 39:60–63, 1982
Kullgren G, Renberg E, Jacobsson L: An empirical study of borderline
personality disorder and psychiatric suicides. J Nerv Ment Dis
174:328–331, 1986
Lewin RA, Schulz CG: Losing and Fusing: Borderline Transitional Object
and Self Relations. New York, Jason Aronson, 1992
Links PS (ed): Family Environment and Borderline Personality Disorder.
Washington, DC, American Psychiatric Press, 1990
Links PS, Steiner M, Offord D, et al: Characteristics of borderline person-
ality disorder: a Canadian study. Can J Psychiatry 33:336–340, 1988
Loranger AW: The impact of DSM-III on diagnostic practice in a univer-
sity hospital. Arch Gen Psychiatry 47:672–675, 1990
Loranger AW, Janca A, Sartorius N (eds): Assessment and Diagnosis of Per-
sonality Disorders: The ICD-10 International Personality Disorder Ex-
amination. Cambridge, England, Cambridge University Press, 1997
Lyons-Ruth K, Jacobvitz D: Attachment disorganization: unresolved loss,
relational violence and lapses in behavioral and attentional strate-
gies, in Handbook of Attachment Theory and Research. Edited by
Cassidy J, Shaver PR. New York, Guilford, 1999, pp 520–574
Mack J (ed): Borderline States in Psychiatry. New York, Grune & Stratton,
1975
Mahler M, Kaplan L: Developmental aspects in the assessment of narcis-
sistic and so-called borderline personalities, in Borderline Personality
Disorders: The Concept, the Syndrome, the Patient. Edited by Harto-
collis P. New York, International Universities Press, 1977, pp 71–86
Main M, Hesse E: Parents’ unresolved traumatic experiences are related
to infant disorganization status: is frightened and/or frightening pa-
rental behavior the linking mechanism? in Attachment in the Pre-
school Years. Edited by Greenberg MT, Cicchetti D, Cummings EM.
Chicago, IL, University of Chicago Press, 1990, pp 161–184
Main M, Solomon J: Procedures for identifying infants as disorganized/
disoriented during the Ainsworth Strange Situation, in Attachment in
the Preschool Years. Edited by Greenberg MT, Cicchetti D, Cum-
mings EM. Chicago, IL, University of Chicago Press, 1990, pp 121–160
Masterson J: Treatment of the adolescent with borderline syndrome (a
problem in separation-individuation). Bull Menninger Clin 35:5–18,
1971
Masterson J: Treatment of the Borderline Adolescent: A Developmental
Approach. New York, Wiley, 1972
Masterson J: Psychotherapy of the Borderline Adult. New York, Brun-
ner/Mazel, 1976
Masterson JF, Rinsley DB: The borderline syndrome: the role of the
mother in the genesis and psychic structure of the borderline per-
sonality. Int J Psychoanal 56:163–177, 1975
Millon T: On the genesis and prevalence of the borderline personality
disorder: a social learning thesis. J Personal Disord 1:354–372, 1987
Modell A: Primitive object relationships and the predisposition to schizo-
phrenia. Int J Psychoanal 44:282–291, 1963
Mors O: Increasing incidence of borderline states in Denmark from
1970–1985. Acta Psychiatr Scand 77:575–583, 1988
Muehlenkamp JJ, Gutierrez PM: An investigation of differences between
self-injurious behavior and suicide attempts in a sample of adoles-
cents. Suicide Life Threat Behav 34:12–23, 2004
Oldham JM, Skodol AE: Personality disorders in the public sector. Hosp
Community Psychiatry 42:481–487, 1991
Paris J: Social risk factors for borderline personality disorder: a review
and hypothesis. Can J Psychiatry 37:510–515, 1992
Paris J, Zweig-Frank H: A 27-year follow-up of patients with borderline
personality disorder. Compr Psychiatry 42:482–487, 2001
Perry JC, Cooper SH: A preliminary report on defenses and conflicts as-
sociated with borderline personality disorder. J Am Psychoanal Assoc
34:863–893, 1986
Perry JC, Klerman GL: The borderline patient: a comparative analysis of
four sets of diagnostic criteria. Arch Gen Psychiatry 35:141–150, 1978
Persson MII, Runeson BS, Wasserman D: Diagnoses, psychosocial stres-
sors and adaptive functioning in attempted suicide. Ann Clin Psychi-
atry 11:119–128, 1999
Pope HG Jr, Jonas JM, Hudson JI, et al: An empirical study of psychosis
in borderline personality disorder. Am J Psychiatry 142:1285–1290,
1985
Rado S: Theory and therapy: the theory of schizotypal organization and
its application to the treatment of decompensated schizotypal behav-
ior, in Psychoanalysis of Behavior, Vol 2. Edited by Rado S. New York,
Grune & Stratton, 1962, pp 127–140
Rathus JH, Miller AL: Dialectical behavior therapy adapted for suicidal
adolescents. Suicide Life Threat Behav 32:146–157, 2002
Reich J, Bostler H, Yates W, et al: Utilization of medical resources in per-
sons with DSM-III personality disorders in a community sample. Int
J Psychiatry Med 19:1–9, 1989
Reich W: Character Analysis, 3rd Edition. New York, Orgone Institute
Press, 1949
Richman NE, Sokolove RL: The experience of aloneness, object repre-
sentation, and evocative memory in borderline and neurotic pa-
tients. Psychoanalytic Psychology 9:77–91, 1992
Rorschach H: Psychodiagnostics, 5th Edition. Bern, Switzerland, Hans
Huber, 1942
Runeson B, Beskow J: Borderline personality disorder in young Swedish

suicides. J Nerv Ment Dis 179:153–156, 1991
Santa Mina EE, Gallop R, Links P, et al: The Self-Injury Questionnaire:
evaluation of the psychometric properties in a clinical population.
J Psychiatr Ment Health Nurs 13:221–227, 2006
Shearer SL: Dissociative phenomena in women with borderline person-
ality disorder. Am J Psychiatry 151:1324–1328, 1994a
Shearer SL: Phenomenology of self-injury among inpatient women with
borderline personality disorder. J Nerv Ment Dis 182(9):524–526,
1994b
Shearer SL, Peters CP, Quaytman MS, et al: Intent and lethality of suicide
attempts among female borderline inpatients. Am J Psychiatry
145:1424–1427, 1988
Silk K, Lohr N, Westen D, et al: Psychosis in borderline patients with de-
pression. J Personal Disord 3:92–100, 1989
Silk KR, Lee S, Hill EM, et al: Borderline personality disorder symptoms
and severity of sexual abuse. Am J Psychiatry 152:1059–1064, 1995
Singer M: The experience of emptiness in narcissistic and borderline
states, II: the struggle for a sense of self and the potential for suicide.
Int Rev Psychoanal 4:471–479, 1977
Soloff PH, Lis JA, Kelly T, et al: Self-mutilation and suicidal behavior in
borderline personality disorder. J Personal Disord 8:257–267, 1994
Spitzer RL, Endicott J, Gibbon M: Crossing the border into borderline
personality and borderline schizophrenia: the development of crite-
ria. Arch Gen Psychiatry 36:17–24, 1979
Stanley B, Gameroff MJ, Michalsen V, et al: Are suicide attempters who
self-mutilate a unique population? Am J Psychiatry 158:427–432,
2001
Stern A: Psychoanalytic investigation and therapy in the borderline
group of neuroses. Psychoanal Q 7:467–489, 1938
Stiver IP: The meaning of care: reframing treatment models, in Women’s
Growth in Connection: Writings From the Stone Center. Edited by
Jordan J, Kaplan A, Miller J, et al. New York, Guilford, 1991, pp 250–
267
Stone M: Contemporary shift of the borderline concept from a sub-
schizophrenic disorder to a subaffective disorder. Psychiatr Clin
North Am 2:577–594, 1979
Stone M: The Borderline Syndromes. New York, McGraw-Hill, 1980
Stone MH: The Fate of Borderline Patients. New York, Guilford, 1990
Stone MH, Hurt SW, Stone DK: The P.I. 500: long term follow up of bor-
derline inpatients meeting DSM-III criteria, I: global outcome. J Per-
sonal Disord 1:291–298, 1987
Swartz M, Blazer D, George L, et al: Estimating the prevalence of border-
line personality disorder in the community. J Personal Disord 4:257–
272, 1990
Torgersen S, Kringlen E, Cramer V: The prevalence of personality disor-

ders in a community sample. Arch Gen Psychiatry 58:521–616, 2001
Vaillant GE: The beginning of wisdom in never calling a patient border-
line. J Psychother Pract Res 1:117–134, 1992
Westen D, Moses MJ, Silk KR, et al: Quality of depressive experience in
borderline personality disorder and major depression: when depres-
sion is not just depression. J Personal Disord 6:382–393, 1992
Widiger TA, Weissman MM: Epidemiology of borderline personality dis-
order. Hosp Community Psychiatry 42:1015–1021, 1991
Widiger TA, Trull T, Hurt S, et al: A multidimensional scaling of the DSM-
III personality disorders. Arch Gen Psychiatry 44:557–563, 1987
Winnicott DW: Transitional objects and transitional phenomena. Int J
Psychoanal 34:89–97, 1953
Winnicott DW: The Maturational Process and the Facilitating Environ-
ment. London, Hogarth Press, 1965
Wolberg AR: The Borderline Patient. New York, Intercontinental Medi-
cal Book Corporation, 1973
World Health Organization: The ICD-10 Classification of Mental and Be-
havioural Disorders: Clinical Descriptions and Diagnostic Guide-
lines. Geneva, Switzerland, World Health Organization, 1992
Zanarini MC, Frankenburg FR: Emotional hypochondriasis, hyperbole,
and the borderline patient. J Psychother Pract Res 3:25–36, 1994
Zanarini MC, Gunderson JG, Frankenburg FR: Cognitive features of bor-
derline personality disorder. Am J Psychiatry 147:57–63, 1990
Zanarini MC, Williams AA, Lewis RE, et al: Reported pathological child-
hood experiences associated with the development of borderline
personality disorder. Am J Psychiatry 54:1101–1106, 1997
Zanarini MC, Frankenburg FR, DeLuca CJ, et al: The pain of being bor-
derline: dysphoric states specific to borderline personality disorder.
Harv Rev Psychiatry 6:201–207, 1998
Zanarini MC, Frankenburg FR, Ridolfi ME, et al: Reported childhood on-
set of self-mutilation among borderline patients. J Personal Disord
20:9–15, 2006
Zetzel ER: A developmental approach to the borderline patient. Am J
Psychiatry 127:867–871, 1971
Zimmerman M, Coryell W: DSM-III personality disorder diagnoses in a
nonpatient sample: demographic correlates and comorbidity. Arch
Gen Psychiatry 46:262–269, 1989
Zimmerman M, Mattia JI: Differences between clinical and research
practices in diagnosing borderline personality disorder. Am J Psychi-
atry 156:1570–1574, 1999
Zisook S, Goff A, Sledge P, et al: Reported suicidal behavior and current
suicidal ideation in a psychiatric outpatient clinic. Ann Clin Psychia-
try 6:27–31, 1994
Chapter 2
DIFFERENTIAL DIAGNOSIS
Overlaps, Subtleties, and Treatment
Implications
Overall Function
All major psychiatric diagnoses represent interactions between baseline
genetic diathesis and adverse environmental stress. Borderline personal-
ity disorder (BPD) exemplifies this diagnosis; and, given its fuzzy borders
with many other psychiatric diagnoses, the importance of making this
particular diagnosis can easily be underestimated. However, failure to
recognize the diagnosis will always create clinical problems.
Identifying BPD is important for several specific reasons.
First, the diagnosis anchors the patient’s and the clinician’s expec-
tations about course. Even when priority may be given to symptoms, be-
haviors, or situational crises, the perspective of a long-term seriously
handicapped person sets realistic boundaries to what can be expected.
BPD patients almost always present with depression, eating disorders, or
substance abuse, but it is only when the BPD diagnosis is identified that
realistic prognostications can occur.
Second, the borderline diagnosis establishes a basis for developing a
treatment alliance by offering patients a developmental and therapeutic
context that they will experience as meaningful and appropriate. As de-
scribed in this chapter, this alliance often develops from the initial reas-
surance that borderline patients feel when they learn that their problems
are shared by others and that their clinicians have a body of relevant
knowledge to draw from.
Third, the diagnosis prepares clinicians for what lies ahead—including
the option of referring the patient to those who may be better able to pro-
37
vide what is needed. Most specifically, the diagnosis helps therapists focus
on the characteristic defensive adaptations that these patients have made
(e.g., regressing, idealizing, blaming) lest therapists unwittingly enact the
roles that these patients commonly project (i.e., caregiver, controller, or
abuser). Indeed, it is because of such countertransference enactments that
astute clinicians began to appreciate that a particular type of personality
psychopathology that lay behind the fluctuating phenomenology could
help explain why clinicians had these problems. Fear of aloneness, for ex-
ample, is a stable underlying trait that gives coherence to the descriptive
characteristics of BPD (see Chapter 1) and conveys added meaning in
terms of both etiology and treatment. Such a characteristic helps clinicians
discriminate BPD from posttraumatic stress disorder (PTSD; Gunderson
and Sabo 1993), narcissistic personality disorder (Plakun 1987; Ronning-
stam and Gunderson 1991), and depressive disorders (Westen et al. 1992).
The Changing Construct:

From Schizophrenia to Depression to
Posttraumatic Stress Disorder to Bipolar Disorder
Figure 2–1 begins to divide the population of individuals with personality
disorders into subtypes created by temperaments, level of impairment,
biogenetics, and phenomenological features. A hierarchy is present, such
that—as in clinical practice (Herkov and Blashfield 1995; Westen
1997)—the presence of a more severe personality disorder makes fulfill-
ing criteria for lesser types superfluous. In this respect, the figure differs
from DSM-IV-TR (American Psychiatric Association 2000), in which the
lesser types are considered comorbid. In this scheme—similar to one rec-
ognized by Kernberg (1986) and buttressed in large measure by empiri-
cal support from Livesley et al. (1992, 1998) and Skodol et al. (A. E.
Skodol, T. H. McGlashan, C. M. Grilo, R. L. Stout, and J.G. Gunderson,
unpublished manuscript, July 2000)—BPD remains one of the major
forms of disorders of the self, alongside schizoid and antisocial personal-
ity disorders. Each of these disorders of self constitutes so severe an im-
pairment that extended social rehabilitative treatments are required.
The presence of these disorders must assume priority in treatment plan-
ning: their presence will greatly complicate, or override, the usual treat-
ments of concurrent Axis I disorders or even medical problems. Thus,
they deserve categorical status or equal weight with Axis I disorders. To
mental health clinicians, BPD is by far the most important type of person-
ality disorder. Socially isolated people with schizoid personality disorder
or socially exploitative people with antisocial personality disorder are not
self-destructive care seekers like those with BPD.
Differential Diagnosis
FIGURE 2–1. How borderline personality disorder (BPD) fits in with neighboring diagnoses.
ASPD= antisocial personality disorder; AVPD= avoidant personality disorder; Bip-II= bipolar II disorder; HPD=histrionic personality disorder;
MDD=major depressive disorder; NPD=narcissistic personality disorder; PTSD= posttraumatic stress disorder; SPD= schizoid personality dis-
❘
order; STPD= schizotypal personality disorder.
39
Comorbidity and Differential Diagnosis

Table 2–1 offers a summation of the literature studying the co-occurrence
between BPD and other diagnoses that can make differential diagnosis
difficult. Although these estimates are from the sources cited in the table,
readers should understand that the estimates are not based on epidemi-
ologically generalizable samples. It is very clear from this extensive, albeit
seriously flawed, literature that rates of overlap increase with higher lev-
els of care (e.g., hospitalized samples will have much higher rates of over-
lap than will outpatient samples). It is notable that, with the possible
exception of the eating disorders, the co-ocurrence rate for the other di-
agnoses is higher in the BPD samples than is the rate of the presence of
BPD in samples with the other diagnosis (e.g., the rate of depression in
BPD samples is about three times as high as the rate of BPD in depression
samples). This phenomenon in co-occurrence rates has fueled the per-
sistent idea that BPD represents an atypical form of Axis I disorder.
BPD and Depression

As shown in Table 2–1, most BPD patients meet criteria for depressive dis-
orders—at least half with major depressive disorder (MDD), dysthymia,
or both. Yet the descriptive characteristics of patients with these condi-
tions seem so disparate—for example, gloomy, anergic depressed pa-
tients versus angry, impulsive borderline patients—that it is not obvious
why the differentiation of them should pose problems. A series of studies
has established that the quality of the depressive experience of border-
line patients is unique and quite distinct from that of depressed or other
patient types (Kurtz and Morey 1998; Rogers et al. 1995; Westen et al.
1992) (see Figure 2–2 for distinctions between BPD and MDD). These
studies have highlighted the emptiness, the primitive guilt, and the neg-
ative, devaluative attitudes of BPD depressions.
Still, and despite the fact that BPD and MDD often co-occur, deciding
which diagnosis should assume treatment priority has often proved diffi-
cult (Gunderson and Phillips 1991; Rogers et al. 1995; Westen et al. 1992).
This difficulty occurs when a patient meets criteria for MDD in the context
of a troubled relationship, with a threat of separation, and with the onset
of suicidal impulses or actions. The clinician must then make a judgment
about whether the patient’s suicidality is a communication motivated by
the wish to gain a sympathetic and binding response (a borderline dy-
namic) or motivated by despair and hopelessness (a depressive mental
state). Did the suicide “attempt” fail because of design or ineptitude?
New evidence shows that when the disorders co-occur, MDD is resis-
tant to antidepressants (see Chapter 7) but that when BPD improves, it is
Differential Diagnosis ❘ 41
TABLE 2–1. Estimated co-occurrence of borderline personality

disorder (BPD) and other diagnoses
BPD PATIENTS WITH PATIENTS WITH
OTHER DIAGNOSIS OTHER DIAGNOSIS
DIAGNOSIS (%) WITH BPD (%)
Depression 50 15
Dysthymia 70 10
Bipolar II disorder 11 16
Bipolar I disorder 9 11
Eating disorder 25 No estimate
Bulimia 20 20
Anorexia 5 20
Obesity 5 10
Posttraumatic stress disorder 30 8
Substance abuse 35 10
Alcohol abuse only 25 5
Somatization 5 10
Narcissistic personality disorder 25 ~15
Antisocial personality disorder 25 ~25
Source. Estimates based on the following review articles: Dolan et al. 2001; Fyer
et al. 1988; Gunderson and Sabo 1993; Gunderson et al. 1991, 1999; Herzog et al.
1992; Hudziak et al. 1996; McGlashan et al. 2000; Paris et al. 2007; Stern et al.
1993; Tyrer et al. 1997; Zanarini et al. 1998a, 1998b.
usually followed by remission of depression (Gunderson et al. 2004). In

the 20%–30% of borderline patients whose depressions do respond to
medication, this is not likely to be followed by significant improvement of
BPD (Gunderson et al. 2003, 2004). These are reasons to give priority to
treating BPD when it co-occurs with MDD. A complicating issue is that pa-
tients may appear to qualify for either diagnosis when they present, but
later the clinician determines that the less phenomenologically evident
diagnosis is primary. This occurs most dramatically for depressed patients
who meet all criteria for MDD and have very modest responses to antide-
pressants but then have a very rapid remission when admitted to a hospi-
FIGURE 2–2. Distinctions and overlapping characteristics be-

tween borderline personality disorder (BPD) and major depressive
disorder (MDD).
tal. The clinician may only then learn, for example, that the “depressed”
patient had held her husband hostage by her dysfunction for several
months—ever since he had begun amicable discussions with her es-
tranged mother. The borderline diagnosis is primary in this instance.
Having said this, clinicians must not assume that any depressed patient
who self-harms is borderline.
Vignette
A patient in whom BPD was diagnosed on the basis of self-mutilative be-
haviors was referred for psychotherapy, where it became clear that she was
chronically isolated and had a developmental history marked by gloomy,
introverted parents and adherence to rigid religious values. She reported,
“I did not know I’d been depressed much of my life. I thought it was nor-
mal, just the way life is.” Her acts of cutting were the outgrowth of long-
standing moral preoccupations and offered her temporary relief from
them. In this patient, a depressive diagnosis was primary.
The clinical significance of this differentiation involves the degree of

optimism that clinicians communicate about what to expect from antide-
pressant therapies. When the disorders co-occur, the pragmatic, empiri-
cal approach described in Chapter 6 is necessary. To convey undue hope
that the patient should rely on antidepressants will be misleading and

evoke unnecessary despair. Although many MDD patients benefit from
psychotherapies, they are often unnecessary, whereas BPD almost always
involves sustained use of several somewhat specialized psychosocial mo-
dalities (see Chapter 3).
BPD and the Bipolar Spectrum

BPD is considered one of the bipolar disorders’ most indistinct boundaries
(Blacker and Tsuang 1992). The similarity of the bipolar disorder (BP)
and BPD acronyms is deserved. The overlap in phenomenology—mood la-
bility and impulsivity—probably relates to a common underlying temper-
ament—that is, a genetically derived disposition (Akiskal 1981; Gunderson
et al. 1999; Silverman et al. 1991). Still, bipolar I is phenomenologically dis-
tinct by virtue of manic episodes, which, when present, should mandate
treatment with mood stabilizers. When the mania remits, clinicians will be
able to observe whether the borderline phenomena persist or were epi-
phenomena. Surprisingly, the co-occurrence of bipolar disorder does not
appear to have much effect on BPD’s course (Gunderson et al. 2006).
The more complex differential diagnostic issue involves bipolar II.
BPD and bipolar II disorder are in fact so similar phenomenologically
(Figure 2–3) and otherwise (i.e., predominantly diagnosed in females
with unstable relationships and heightened risk of suicide) that it is un-
clear whether these are two independent disorders (Gunderson et al.
1999; Kopacz and Janicak 1996; Paris et al. 2007). Remarkably, given the
high frequency with which borderline patients receive the diagnosis of
bipolar II, the actual overlap when criteria are assessed is only about 11%
(see Table 2–1) (Paris et al. 2007). Similarly, reports on patients with
bipolar II have shown that only 11%–23% of them meet criteria for BPD
(Paris et al. 2007).
Table 2–2 offers some ways to distinguish BPD from bipolar II disor-
der. Especially telling can be the differential responsiveness of BPD and
bipolar II patients to confrontation or interpretation (Bolton and Gun-
derson 1996). Borderline patients will react, sometimes constructively
and sometimes not. Bipolar II patients are not fazed: they are likely to go
on as if the intervention had not occurred—either by not responding at
all, by changing the topic, or by glibly rationalizing. Both bipolar II and
BPD patients may respond to external controls by rage or flight, but the
borderline patients’ responses will always and clearly be emotional. They
will believe that much is at stake, either about their self-esteem or about
the clinician’s trustworthiness. A related finding by Benjamin and Won-
derlich (1994) was that hospitalized borderline patients perceived more
hostility and autonomy in others than did bipolar depressed inpatients.

tween borderline personality disorder (BPD) and bipolar II disorder.
Vignette
A 34-year-old man who had undergone a female-to-male sex change op-
eration was flirtatious and had many affairs with members of both sexes.
After being hospitalized for suicidal impulses, he quickly became the “life
of the unit” and wondered aloud why his therapist would have thought he
was suicidal. When the patient was confronted with the facts that his re-
cent vocational and relational failures were doubtless related to these im-
pulses, he angrily stood up and declared, “How dare you talk to me like
that. You have no right to call me a ‘loser’! Do you want me to kill myself?”
He then promptly filed a formal complaint about his treatment.
The subtleties that differentiated this patient’s diagnosis were the in-
discriminate thrill-seeking or attention-seeking aspects of his behavior,
his confidence that authorities would help him punish the transgressors
(the confrontational staff), and the patient’s interest in keeping all rela-
tionships transient (as opposed to exclusive and binding). Particularly
important, in my experience, was the glibness of his feelings—it was dif-
ficult to take them seriously or to empathize with them. These character-
istics tilted the diagnostic balance toward bipolar II.
Vignette
A 28-year-old female litigation lawyer was referred for psychiatric consul-
tation after assaulting her boyfriend for having resumed drinking and
having lied to her about it. She had recently had an abortion with his en-
TABLE 2–2. Comparison of borderline personality disorder (BPD)

and bipolar II disorder
TRAIT BPD BIPOLAR II
Mood lability/ Due to interpersonal Autonomous and persis-

impulsivity sensitivity tent (person acts out)
Affects Deep, intense; evoke Include elation; lack
strong empathic depth, pain; hard to
response empathize with
Prototypical Care seeking: seeks Energetic self-initiated
behavior pattern exclusivity, is sensitive activities that are left
to rejection incomplete; requiring
others to clean up, attend
to details
Defense Splitting: polarizes reali- Denial: ignores undesir-
ties and, if challenged, able realities and, if con-
becomes angry at chal- fronted with a reality,
lenger or changes to denies its emotional sig-
opposite view nificance
couragement. Their 5 years of cohabitation had been marked by her

workaholic habits and growing success and by their both having had brief
affairs during periods of drinking.
On evaluation, she initially seemed pressured—to convey her story as
quickly as possible and to impress the consultant with her intelligence
and innocence. This gave way to depressing themes of being unlovable,
being distrusting, and her long-standing experience of presenting a false
self. She sobbed with remorse about having had the abortion.
When asked to consider the suitability of both the bipolar and the
borderline diagnoses, she was resistant. She wanted to be told. When she
read the borderline diagnosis, she first said, “That fits me perfectly!” She
then said that she did not want to lose her identity. When she saw the bi-
polar diagnosis, she said, “So now are you telling me it’s genetic? Does
that mean you don’t want to treat me?”
This patient presented with characteristics of both disorders. Her bipo-

lar disorder was evident in pressure of speech, abrupt switches to depres-
sion, and a history of hypomanic levels of activity. Her borderline disorder
was evident in her splitting responses to diagnostic queries, in her interper-
sonal sensitivity and instability, and in her history of identity problems.
She is a reminder that these disorders do co-occur. Such co-occurrence
presents significant treatment problems. A clinician must both attend to
the significant—often lifelong—role that medications will have in treat-
ment planning and enlist the patients’ collaboration in changing them-

selves. It can be difficult to appreciate the relative contribution of each
disorder to impulsive acts (e.g., substance abuse or unprotected sexual in-
tercourse). The history of acting without thinking may have become so ha-
bitual that a clinician’s expectations of introspection or attachment will be
more foreign and more feared in patients with this comorbidity than in pa-
tients with BPD alone. The existing evidence suggests that comorbid bipo-
lar disorder does not much affect BPD’s course (Gunderson et al. 2006).
Whether BPD affects bipolar I or II’s course is unknown.
Mood stabilizers are often used for BPD (Gunderson et al. 2006). BPD
responds to lithium atypically—more in the domain of impulsivity than
mood lability (Links et al. 1990). Bipolar II disorder also responds less
consistently or clearly to mood stabilizers than does bipolar I disorder. In-
deed, the overall profile of medication responsiveness in bipolar II disor-
der is closer to that of BPD than to bipolar I (Paris et al. 2007). Because
the borderline disorder has proved amenable to various psychosocial in-
terventions and the bipolar II diagnosis is typically treated only with med-
ications—which are not particularly effective (Suppes and Dennehy
2002)—it is best to focus treatment on BPD (Gunderson et al. 2006). The
advantage of documenting the bipolar disorder is primarily pragmatic—
it may enhance the patient’s insurance coverage.
The complexity of the interface between these disorders now offers a
Rorschach-test-like opportunity for clinicians to project their biases
(Sidebar 2–1). In years past, schizophrenia, depression, and PTSD have
each offered a preferred diagnostic option for clinicians who were resis-
tant to making the BPD diagnosis. These diagnoses reduced the level of
accountability and hostility they assigned to those patients. A similar di-
agnostic process takes place in the case of bipolar II—but by a different
mechanism. Many clinicians, perhaps especially biological psychiatrists,
prefer Axis I diagnoses because they offer a rationale for a pharmacolog-
ical approach that will keep managed care overseers at bay and will limit
the level of involvement with such patients (sometimes actually the clini-
cian’s preferred level). The diagnosis of BPD means that emotional in-
volvement will be an essential aspect of meaningful treatment. Clinicians
whose interest and skills are primarily psychotherapeutic will be prone to
see the same patients as having BPD and may err by overlooking bipolar
or cyclothymic phenomena that could respond to mood stabilizers.
Sidebar 2–1: Was Vincent van Gogh Borderline?

The most celebrated self-mutilative act in history is Vincent
van Gogh’s cutting off the lobe of his left ear. Psychiatric
opinions about van Gogh have been numerous and varied
(Blumer 2002). Most have noted his seizures, suicide, inter-

mittent dependency on absinthe, and extreme emotional-
ity. The last symptom has suggested that he had a form of
bipolar disorder (Blumer 2002; Morrant 1993).
A Dutch psychiatrist, Erwin van Meekeren (2003), has
revisited van Gogh’s history. He highlighted the artist’s ex-
treme interpersonal sensitivity. On Christmas Eve shortly
after Gauguin had announced he would leave, van Gogh
had thrown absinthe in Gauguin’s face. Gauguin then took
van Gogh home and put him to bed. But van Gogh fol-
lowed Gauguin out into the street, threatening with his
razor. Again, Gauguin returned van Gogh to his bed, and
sometime after this, van Gogh cut his earlobe.
Without question, van Gogh’s life was marked by un-
stable moods, behaviors, and relationships. His longings for
love, his sudden mood changes (and most particularly his
seemingly unpredictable and unwarranted rages), and his
pattern of impulsive acts, including substance abuse, are
all recognizable components of the borderline syndrome.
Perhaps most telling, he ended his life in an unpremedi-
tated impulsive act when confronted by the potential loss
of his needed brother’s support.
It was argued by van Meekeren (2003) that BPD offers
a way to understand the issues that compelled van Gogh’s
self-mutilation and provides a more comprehensive way to
understand other vicissitudes of van Gogh’s stormy life. By
focusing on van Gogh’s enduring personality traits, his ep-
isodes of dysfunction, depression, and psychosis became
more understandable and predictable extensions of sta-
ble underlying vulnerabilities. Insofar as psychiatric hospi-
tals provided him with a “holding” environment, it helps
explain why his stays were among the two most productive
and self-satisfied periods of his life. Whether van Gogh was
borderline or not, it is a useful prism through which to view
his troubled life.
BPD and Posttraumatic Stress Disorder

The issue of differential diagnosis between BPD and PTSD involved the
same basic question in 1990 as did depression in 1980 and as does bipolar
II at present: are these really separable disorders? The question of the
border with PTSD is usually raised when a depressed, self-mutilative, and
impulsive patient has a developmental history marked by significant
childhood trauma. The clinician then considers whether that reaction to
trauma was sufficient to account for the presenting adult’s emotional and
behavior problems (i.e., PTSD) or whether the trauma was itself emblem-
atic of sustained developmental problems that formed the patient’s dis-
turbed personality (i.e., BPD). This vignette illustrates such a problem:
Vignette
A 44-year-old woman presented with flashbacks that disrupted her sleep
and concentration. Her childhood included eight hospitalizations be-
tween ages 13 and 18 for treatment of a congenital disease. Twenty-six
years later, she could still access the feeling of being “helpless and alone.”
In response, she would become agitated, with bursts of accusatory, offen-
sive anger toward her husband and children, which she would then
deeply regret as unfair. This remorse then prompted self-destructive or
suicidal impulses.
The diagnosis of complex PTSD (Herman 1992) is warranted when

such patients have flashbacks or sustained dissociative experiences and an
interpersonal style marked by wariness and fears of attachment—such that
in adulthood, social isolation is usual and only intermittently interrupted
by brief, often alcohol-related, social forays. If the patient is very hungry
for attention and protection and is expressive of angry feelings when hurt,
the effect of trauma is less likely to have been dominant, and the patient is
better conceptualized as having BPD. Westen (1990) pointed out that
when patients are organized around their abuse experiences (i.e., when
they have complex PTSD), they are more likely to respond with paranoid
accusations of malevolence within the context of an ongoing relationship,
whereas borderline patients are more likely to become accusatory when
threatened by the loss of their other (see Chapter 1).
The interface between the disorders is complex (Gunderson and
Sabo 1993; Herman et al. 1989). Abusive experiences predispose children
to a variety of serious psychiatric illnesses, including BPD. For the approx-
imately 70% of BPD patients who have childhood histories of physical or
sexual traumas, the sexual abuse most distinguishes them from traumas
associated with antisocial personality disorder. Adult BPD patients are vul-
nerable to developing PTSD by virtue of their recklessness and emotional
hyperreactivity. Indeed, PTSD co-occurs in about 30% (lifetime 40%) of
BPD patients (Swartz et al. 1990; Zanarini et al. 1998a). The social condi-
tions needed for BPD to develop require emotional estrangement from
parents. This estrangement gives abusive experiences during childhood
an effect that is far more traumatic in warping character development
than is the effect of similar events on children who have the opportunity
to find support, talk about the events, and react with their families.
The presence of childhood trauma has clinical significance—as much
for the attitudes of treaters as for the nature of the treatments required
(Gunderson and Chu 1994). Many clinicians, influenced by their instinc-
tive sympathy for victims of violence, may prefer the PTSD diagnosis for
anyone with childhood traumas because it encourages a deep involve-
ment while sidestepping these patients’ hostility.
Vignette
A 34-year-old unmarried woman sought psychotherapy because she
“needs support.” She related this to a series of recent events.
She loved her job but, after becoming convinced that she was under-
paid, demanded more pay from her employer. She consequently lost her
job. She also had a fight with her landlord, insisting on her rights. This too
resulted in her being kicked out. In both instances, she perceived injustices
in the situations correctly, but she experienced the injustice too personally,
and her anger was disproportionate. Depressed about the consequences of
her fights and about the prospects of having no husband and no children,
she moved back to live with her mother and with her 40-year-old brother.
This brother had sexually abused her when she was between ages 6 and 10.
Her mother knew but had coped by alternating between helplessness and
denial.
The patient presented as very sensitive, wary, and vigilant to rejection
and criticism, with a defensive response to interpretations. She acknowl-
edged fears of intimacy and attachments. Her defensiveness made explor-
atory therapy unlikely. Even when a supportive therapist attempted to
work with her, she resisted getting attached.
This patient might have been given a BPD diagnosis by virtue of her an-
ger and need for support, but in my opinion, she would better be identified
as having complex PTSD, as proposed by Herman (1992). The bleakness of
her interpersonal life and her resistance to any attachment set the effects of
trauma apart from what is seen in BPD. Although the PTSD diagnosis is
sometimes overused by clinicians sympathetic to victims, its clinical signif-
icance often means that an intensive, exploratory, or close therapeutic re-
lationship will not be as possible as it is with patients having BPD.
BPD and Eating Disorders

Eating disorders constitute one of the three most common presenting
complaints of patients with BPD. Bulimia is the most common type of co-
occurring eating disorder (Table 2–1). Individuals with bulimia are more
impulsive than are those with anorexia; the latter are more perfectionis-
tic and conscientious as people and are more purposeful and persistent
in their personal deprivations than are those with bulimia. Even more sig-
nificant is the way in which bingeing and purging offer outward expres-
sion of internal splits. Starving oneself, the ascetic ethic of denying one’s
appetites and needs, is “dutiful and good.” Eating is “bad,” associated
with defying control and with being too aggressive.
Sustained deprivation accompanied by persistent body image distor-
tions or illusions of purity and perfection characterize the prototypical
individual with anorexia nervosa. In such individuals, BPD is unlikely or,
if present, is secondary.
Those with bulimia, however, alternate this state of “good” depriva-

tion with angry and entitled feeling states based on feeling that they have
suffered more than their share. Under these circumstances—rebelling
against their self-imposed restrictions—they binge. No sooner does this
occur than they conclude that they have been self-indulgent, feel deeply
ashamed, and need to be punished. The punishment often takes the
form of renewed anorexia or the recurrence of self-destructive behaviors.
Thus, their purging, their impulsivity, and the volatility of their self-image
are indicators of underlying BPD.
Other complications involve appearance. Both the underweight and
the overweight borderline individuals are making both a private state-
ment of goodness or badness and a public statement of neediness. Both
also may want to deflect attention or interest in their sexuality. Weight
problems may indicate an inability to take care of oneself—without hav-
ing to acknowledge unrecognized or humiliating dependency needs.
Finally, eating disorders can be a particularly effective way to torture
mothers who see their provision of food as extensions of their love or who
see their child’s appearance as a narcissistic extension of themselves. In
this way, overeating or undereating is like most other self-punishing be-
haviors by borderline patients: it contains both intrapsychic and inter-
personal meanings (see Chapter 4). Eating disorder behaviors leave
responsibility for the patients’ safety and welfare in the hands of others.
Indeed, it is central to having BPD that eating (i.e., living) can be justified
only if there is reassuring and concrete evidence that others want you to
eat (i.e., want you to live) and that they will take responsibility for keeping
you alive.
Borderline patients who are obese are likely to have a history of sexual
abuse (Sansone et al. 1995) and often have conscious desires to deflect
sexual interests. Complex PTSD (see previous section, “BPD and Posttrau-
matic Stress Disorder”) needs to be considered. In addition, in an era of
multiple medications, secondary obesity in borderline patients is dramat-
ically increasing (see Chapter 7). This can be a particularly unfortunate
side effect in young women who already feel alienated and unwanted. Ed-
ucation and revised treatment plans should follow from this.
The distinction between patients with eating disorders and BPD and
those with eating disorders but without BPD usually can be found in their
developmental history. Borderline patients will typically have impulsivity
and dysfunction related to markedly unstable family situations. Many
patients with eating disorders but without BPD have histories in which
family problems are not easily recognizable. These patients have devel-
opmental histories that typically reflect the narcissistic issues of counter-
dependence and expectations of high achievement. Such patients have
more sensitivity to inferiority and fewer concerns about rejection than

are experienced by borderline patients.
BPD and Substance Abuse

When a patient has a history of heavy substance use and, along with it, a
history of desperate, impulsive, self-endangering relationships, the clini-
cian must determine whether these typically “borderline-like” relation-
ships are really evidence of BPD or are simply behavioral by-products of
the drugs—that is, outgrowths of a primary substance abuse disorder.
Whether the BPD symptoms are secondary or not, clinicians should ap-
preciate that these disorders are related by virtue of an underlying dispo-
sition or phenotype for impulsivity (White et al. 2003).
If the substance abuse is primary, the behaviors and relationships are
caused by drug-seeking or drug-related disinhibition. The patient is the
victim of his or her drug needs and feels regretful about the relationships.
Alternately, the drug use can be a behavioral by-product of the cravings of
the person with BPD to soothe himself or herself (i.e., to self-medicate dys-
phoric feeling states). This drug use, described in Khantzian’s “self-medi-
cation hypothesis” (Khantzian 1985), also has a desirable disinhibiting
effect that permits borderline patients who would otherwise find it unac-
ceptable to make active efforts to seek relationships. From this perspective,
the substance abuse would be a secondary symptom of the primary BPD.
As with the general population, the most common type of substance
abused by borderline patients is alcohol, but what is most specific to these
patients is that the type of substance is not very important—that is, they are
polysubstance abusers (Nace 1989). Their abuse tends to be—but is not al-
ways—episodic and impulsive, and they use whatever drug is available. Even
if hard-core addicted persons have BPD, the treatment of their substance
abuse needs to take priority. For such people, the structures, supports, and
ideology of substance abuse programs are ideal and are essential for gain-
ing control over their drug habit. Even for borderline patients whose sub-
stance abuse is episodic and clearly secondary, the treatment options used
for substance abusers still have value—especially for patients who can ac-
cept the diagnosis of substance abuse but resist the idea of having more sus-
tained problems with relationships and impulsivity. The ubiquitous and
daily access to Alcoholics Anonymous or Narcotics Anonymous self-help
services meets borderline patients’ needs for support, crisis management,
and networking in ways that the mental health system can rarely replicate.
Vignette
A 24-year-old woman with diagnoses of BPD and polysubstance abuse was
transferred to a hospital after being kicked out of her third substance
abuse program. In each case, she had violated every restriction by resum-
ing her pain-drug habit. While failing several placements in residential
programs, she became attached to a therapist, and gradually a new pre-
cipitant for her substance abuse relapse became apparent: relapse oc-
curred when her mother (although geographically distant from the
patient) traveled to see the patient’s brother or was visited by him. At this
point, the psychodynamic BPD issues took precedence over the substance
abuse issues in the therapist’s mind. Concretely, this idea surfaced when,
in response to the patient’s expressed wish to relocate home, her sub-
stance abuse counselor said, “When you have remained sober for 3
months,” and the therapist reframed the criterion as “When you can man-
age visits between your mother and brother without relapsing.”
Although this vignette illustrates how, with the understanding that

comes from psychotherapy, priority can shift from a substance abuse di-
agnosis to a BPD diagnosis, such cases often do not work out so well. The
presence of comorbid substance abuse appreciably increases the likeli-
hood of suicide and diminishes the overall prognosis for the BPD patient
(Stone 1990a). On the other hand, borderline patients who recover from
substance abuse habits may have very good outcomes (Gunderson et al.
2003; Stone 1990b).
Still, few borderline patients can change or grow without therapies
that are directed at their problems with close relationships and at the
management of their angry and anguished feelings. It is therefore impor-
tant for substance abuse programs to be sensitive to these ongoing issues
for these patients—whose success in living, despite their appearing com-
mitted to sobriety, depends on other, more borderline-specific therapies.
BPD and Somatoform and Somatization Disorders

Borderline patients are care seekers. Patients with somatoform disorders
also seek care, but these people go to medical doctors, surgeons, and
emergency departments with physical complaints as a way to get care. In-
deed, Zanarini and Frankenburg (1994) termed the borderline patient’s
use of emotional displays to elicit care emotional hypochondriasis. This form
of care seeking can be carried out with any significant other, but within
medicine it usually involves the mental health disciplines. Some border-
line patients are adamant about seeing psychiatrists (rather than psychol-
ogists or social workers) because they insist that their mental problems are
somatic diseases. Patients who have multiple somatic complaints without
obvious or well-documented physical pathology (i.e., somatizers) raise
questions of malingering or hypochondriasis. The difference is that ma-
lingerers sometimes consciously but guiltlessly manufacture symptoms
(sometimes called Munchausen syndrome), whereas the patient with hy-
pochondriasis, having no more of a physical basis for his or her symptoms,
consciously believes that medical care can relieve the symptoms. Border-
line patients may use physical complaints in either way (Nadelson 1985).
Although patients with Cluster C personality disorders are even more
likely to have somatoform disorders than are patients with BPD (Fink
1995; Stern et al. 1993), the risks associated with having BPD make it par-
ticularly important for medical services staff to be aware of this disorder.
Some borderline patients may wish to be injured (i.e., they may con-
sciously or unconsciously seek mistreatment), and then liability issues
can haunt unsuspecting doctors. Indeed, somatoform patients exemplify
why making the borderline diagnosis can be of critical importance.
Knowing about BPD in a somatizing patient increases doctors’ awareness
about potential misuse of and placebo effects from medications. It helps
redirect the search for care into more explicit and less dangerous com-
munications. It also encourages clinicians to emphasize basic health care
messages about diet, sleep, and exercise that many borderline patients
otherwise neglect. These messages are standard aspects of how cognitive-
behavioral therapists assist chronic pain patients. Finally, it will be a help
to such patients for other physicians to refer them for psychiatric care.
BPD and Narcissistic Personality Disorder

The differential diagnosis issue between BPD and narcissistic personality
disorder is usually triggered by the concurrence of inappropriate anger,
feelings of entitlement, and suicidality (Figure 2–4). People with either of
these personality disorders are likely to use defenses such as devaluation,
projection, and counterdependence (Perry and Perry 2004). This diag-
nostic question is more common in males because they are more typically
unaware of dependency needs and thus more apt to seem narcissistic (i.e.,
they are more likely to present as distant and haughty). People with either
type of personality disorder can become either angry or self-destructive in
response to rejections or even criticisms. These apparent areas of overlap
often yield discrimination on closer scrutiny (Sidebar 2–2).
Sidebar 2–2: Is Martha Stewart Borderline?

On September 2, 1997, the National Enquirer headlined the
story “Martha Stewart: Mentally Ill” and went on to report ex-
perts who judged her to have BPD. Their conclusions were
based on reading the unauthorized biography of her written
by Jerry Oppenheimer, titled Martha Stewart: Just Desserts.
The supposed experts cited the book’s documentation of
abandonment fears, demandingness, rages, self-destructive
acts, threats, shifts from idealizing to devalued views of oth-
ers, and impulsive acts. These experts argued that she would
appear to fulfill the current DSM diagnostic criteria for BPD.

tween borderline personality disorder (BPD) and narcissistic per-
sonality disorder (NPD).
Yet in the biography, other impressions that cast doubt

on this diagnosis become evident. The behaviors, fears, and
rages are not sustained patterns. They take place against a
life patterned on high achievement in challenging and
competitive fields of endeavor. The behaviors are reactive
to events in which the issues do not involve feeling deprived
and alone, but rather injured pride. Martha Stewart’s self-
esteem is tied to her public image of beauty, productivity,
and perfectionism. That she is demanding of others is not in
the service of nurturance; it is in the service of completing
tasks according to her lofty standards. Although she is surely
capable of rages, they ensue when she feels frustrated, crit-
icized, or defeated, not when she feels neglected. Her self-
esteem is precariously perched on standards of superiority,
not precariously sustained by evidence of lovability.
The primary reason for consideration of the borderline
diagnosis involves Martha Stewart’s alleged response to
her husband’s leaving her. She became enraged with feel-
ings like “How could he do this to me?” and “He has no
right to do this.” It is the functions he served for her that she
felt abandoned by, not the man per se. His departure
seemed an insulting and humiliating betrayal of her. There
is little evidence that she felt she needed his love or sup-
port. For her, those needs would be admissions of weak-
ness. More at stake seemed to be her belief in her ability to
control him. These are narcissistic, not borderline, issues.
Reading her unauthorized biography could justify

speculation about narcissistic personality disorder, but the
dangers of such speculation are twofold: the potential for
mistakes from such a data source and the potential for be-
ing the target of a response like Martha Stewart’s response
to the National Enquirer. Her response to the Enquirer was
similar to her reaction to her husband: vindictive rage. She
brought suit against the newspaper for libel: defamation of
character. Defamation is what hurt.
Vignette
Matthew, an 18-year-old man who used what seemed to be his girlfriend’s
idealization of him to sustain his fantasies of becoming a great poet, be-
came very agitated and had suicidal ideas when he learned of his girl-
friend’s plans to relocate to another school—despite her assurances of
ongoing love. In therapy, he talked about being enraged by the disparity
between what she meant to him and what he meant to her—“otherwise,
she would never leave.” He hated himself for “being so stupid” as to let
her mean so much.
There is no question that Matthew is narcissistically injured, but the

differential diagnosis of narcissistic personality disorder or BPD cannot
be clarified without knowing 1) whether the injury revolves around the
prototypically borderline fear of abandonment or around the prototypi-
cally narcissistic reaction to perceived threats to a grandiose self-image,
2) whether his relationship was sustained by his girlfriend’s idealization
(narcissistic personality disorder) or by her caregiving (BPD), and 3)
whether he has had a pattern of intense relationships that ended because
of his becoming too needy (BPD).
Beyond the similarities noted earlier, however, these disorders di-
verge (Ronningstam and Gunderson 1991). Persons with BPD expe-
rience rejections as abandonments that trigger their fears of stark
aloneness. For them, criticisms may be intolerable because what is in-
tended as discrete becomes generalized into an indictment of their whole
person (i.e., their overall badness). Persons with narcissistic personality
disorder experience either rejection or criticisms as shameful humilia-
tions that trigger feelings of defeat or inferiority but, as Rinsley (1984)
pointed out, do not involve issues of survival of self or others. Akhtar and
Thomson (1982) believe that the grandiosity or feelings of omnipotence
that characterize patients with narcissistic personality disorder mask (i.e.,
compensate for) covert convictions of inferiority. The self-destructive re-
sponse of a borderline patient is likely to be impulsive, consistent with
other impulsive behaviors; or, if it is a calculated response, it will be de-
signed to regain caring attention. The self-destructive response of a per-
son with narcissistic personality disorder is less likely to be part of a

pattern and more likely to have lethal intentions.
People with either narcissistic personality disorder or BPD feel enti-
tled to special privilege, attention, or care. People with narcissistic per-
sonality disorder believe they deserve it because they are unique and
exceptional and have “earned” it. Persons with BPD feel entitled to spe-
cial privilege, attention, or care because they have suffered and because
they need more. Persons with narcissistic personality disorder would be re-
luctant to recognize or acknowledge being needy; for them, it would be
humiliating.
This differentiation has both theoretical and psychotherapeutic im-
plications (Ronningstam and Gunderson 1991). The theoretical implica-
tions (discussed further in Chapter 12) involve the question of whether
aggression—along with its symptoms, rage, hostility, and anger—is 1) a
primary drive whose misdirection and dyscontrol create the psychopa-
thology (i.e., is part of BPD) or is 2) reactive to environmental insults, so
that its misdirection or dyscontrol is a symptom of an overly fragile self
(i.e., is part of narcissistic personality disorder). This question has thera-
peutic significance. The treatment of choice for narcissistic personality
disorder involves a long-term corrective attachment whose effectiveness
depends on the therapist’s empathy and sensitivity to not bruising the pa-
tient’s self-esteem so much as to precipitate flight. A role for medications
has not been established. Although long-term corrective attachment ex-
periences are also important in the treatment of BPD, this disorder, in
contrast to narcissistic personality disorder, typically requires pharmaco-
logical (Chapters 6 and 7) and social rehabilitative (Chapters 3, 5, 8, 9)
modalities. As described throughout this book, initial treatments should
be directed at behavioral and affect controls. Within the individual psy-
chotherapies, borderline patients require that more attention be paid to
contracting, boundaries, regressions, and negative transference issues
(Chapters 3, 10–12).
BPD and Antisocial Personality Disorder

About 75% of BPD patients are female, about 75% of antisocial person-
ality disorder patients are male, and about 25% of patients with either di-
agnosis will meet criteria for the other (Zanarini and Gunderson 1997).
Hatzitaskos et al. (1997) found that persons with BPD had more intro-
verted hostility and that those with antisocial personality disorder had
more extroverted hostility. Antisocial personality is usually marked by ac-
tion-oriented defenses and, as emphasized by Livesley et al. (1989), a
cold, interpersonally exploitative way of relating to others (Figure 2–5).
The diagnostic dilemma often occurs when a female patient, otherwise

tween borderline personality disorder (BPD) and antisocial per-
sonality disorder (ASPD).
prototypically borderline, has a pattern of calculated (conscious) deceit-

fulness (including malingering) and/or episodic violence. The diag-
nostic dilemma also occurs when a male patient with a clear pattern of
violence or interpersonal irresponsibility has recurrent suicidality and
deep-seated feelings of badness about himself that he struggles to keep
out of his awareness.
The issue is often made no easier after a developmental history is ob-
tained because patients with either of these diagnoses often report ne-
glect, abuse, and alienation (Robins 1966; Widom 1997; Zanarini et al.
1989). Their family experience often includes marital discord, abandon-
ment, violence, and substance abuse. Kernberg’s (1967, 1986) placement
of these disorders together within the overarching construct of a border-
line personality organization marked by broken identities, primitive de-
fenses, and transient failures in reality testing seems incontrovertible. So
disturbed is the development of many BPD or antisocial personality dis-
order individuals that it seems unnecessary to invoke genetic causality.
Yet it seems likely that the aggressivity and impulsivity shared by these pa-
tients have their roots in a temperamental predisposition (Coccaro and
Kavoussi 1991; Paris 1997; Siever and Davis 1991; Zanarini and Gunder-
son 1997). Relatives of borderline patients have a high frequency of an-
tisocial personality disorder (White et al. 2003). Both disorders score
similarly on the personality dimensions of high neuroticism and low

agreeableness and conscientiousness (Widiger et al. 1994).
These similarities in phenomenology and development suggest that
these two diagnoses may be highly related forms of psychopathology. In-
deed, I think they are and that the distinctions are probably gender re-
lated, from both a genetic and an environmental perspective. (As a
parallel observance, even infant boys can be differentiated from infant
girls: the boys are more instrumental, and the girls are more affiliative
[Gilligan 1982; Jordan et al. 1991]. This presumably genetic difference
may be supported by environmental experiences that support girls’ so-
cialization and boys’ task competence.) Relevant to this is the finding that
adult borderline patients report a high frequency of use of transitional
objects in their childhoods (Morris et al. 1986). Transitional objects, as
observed by Winnicott (1953), are a normal way in which children at-
tempt to diminish anxieties normally associated with separation from a
caregiver. With borderline patients, transitional-object use often extends
past its usual spontaneous ending by age 4 or 5 (Arkema 1981; Lobel
1981; Morris et al. 1986). The nursing staff on inpatient units confidently
predict that when a newly admitted patient insists on his or her stuffed an-
imal, blanket, or other inanimate source of comfort, the person is bor-
derline (Cardasis et al. 1997; Labbate and Benedek 1996). In contrast,
Horton et al. (1974) reported, with conflicting replications (Cooper et al.
1985; Morris et al. 1986), that adult patients with antisocial personality
disorder report no use of transitional objects. This finding is consistent
with both genetic and environmental pressure for boys to manage their
needs, or drives, in ways that are less interpersonal than the way girls do.
Vignette
Mr. A, a 23-year-old man with divorced parents, developed an intense, ide-
alized relationship with his very supportive but inexperienced substance
abuse counselor. Because of Mr. A’s continuing to steal from his family
and from stores and to drive too fast despite repeated encounters with the
law, his mother sought consultation. When a change to a more confron-
tational and intensive therapy was recommended, Mr. A became very abu-
sive and threatened his mother and stepfather with a knife. When his
counselor, frightened by Mr. A’s desperate calls and by his threats to kill
himself, joined the mother in support of a change in treatment, Mr. A ran
away. The next contact from him was a telephone call apologizing for his
flight and requesting that his mother send him money to pay a debt and
transport him home.
Had Mr. A been 16 years old, a borderline diagnosis would have

guided his treatment and thus would have probably required sustained
care on level III (partial hospital). As a 23-year-old, his borderline issues
(i.e., need for caring attention) were still dominant, but it was unclear to
what extent his past drug use, his potential violence, and his dishonesty
(i.e., antisocial personality disorder issues) made such treatment unlikely
to succeed. His telephone call for help does nothing to resolve the ques-
tion of whether his motives were exploitative or were guided by a real
wish for rapprochement. By now, his call would probably be better re-
sponded to as an exploitative and manipulative act (i.e., as if he primarily
had antisocial personality disorder).
The differentiation of these disorders has major significance to clini-
cians. To mistakenly diagnose a borderline patient as having antisocial per-
sonality disorder often consigns a potentially treatable patient to minimal
treatment. To mistakenly diagnose a patient with antisocial personality
disorder as having BPD is to initiate the ineffective use of valuable clinical
resources and to expose other patients, even the treaters, to potential ex-
ploitation and, at worse, physical harm. Having said this, the case can be
made that, when the diagnosis is in doubt, it is best to honor evidence of
the patient’s interest in treatment and to make a serious effort (Zanarini
and Gunderson 1997). Tipping the balance toward treatment are 1) evi-
dence of a hunger to be attached, 2) a capacity to bear negative feelings
(e.g., shame, envy) or self-critical attitudes, 3) any history of sustained role
functioning, 4) availability of significant supports for the treatment from
people the patient needs or respects, and 5) adequate monitoring of the
patient’s use of a therapy. Keeping an eye on these guidelines will allow cli-
nicians to stop a therapy before harmful consequences occur. Unfor-
tunately, for borderline patients who also fulfill criteria for antisocial
personality disorder, their responsiveness to treatment usually will be re-
duced (Clarkin et al. 1994).
Summary
This discussion of the most common and difficult differential diagnostic
issues has established that the boundaries separating BPD from neigh-
boring disorders are often inherently unclear. The decision about prior-
itizing the diagnosis of BPD versus that of its overlapping neighbor
should be guided by whether the treatment implications will benefit the
patient. In most instances, making treatment plans that overlook the bor-
derline diagnosis when it is present sets the stage for therapeutic im-
passes or worse (splits, regressions, countertransference enactments);
common examples where this can occur are with depression, bipolar II,
and bulimia. Still, exceptions do exist; substance abuse, anorexia, and bi-
polar I disorder require attention and stabilization before BPD can be
treated. Clearly woven into these diagnostic considerations are counter-
transference issues. There is an inevitable inclination to diagnose BPD by

those who believe they can treat it, to ignore it when clinicians believe
that they cannot treat it, and to invoke it as a retrospective explanation
for patients who prove noncompliant or unresponsive. In this chapter, I
encourage clinicians to consider thoughtfully whether a borderline diag-
nosis is apt and then to recognize how its presence can usefully inform
clinical decisions.
References
Akhtar S, Thomson JA Jr: Overview: narcissistic personality disorder. Am
J Psychiatry 139:12–20, 1982
46, 1981
Arkema PH: The borderline personality and transitional relatedness. Am
J Psychiatry 138:172–177, 1981
Benjamin LS, Wonderlich SA: Social perceptions and borderline person-
ality disorder: the relation to mood disorders. J Abnorm Psychol
103:610–624, 1994
Blacker D, Tsuang MT: Contested boundaries of bipolar disorder and the
limits of categorical diagnosis in psychiatry. Am J Psychiatry 149:
1473–1483, 1992
Blumer D: The illness of Vincent van Gogh. Am J Psychiatry 159:519–526,
2002
Bolton S, Gunderson JG: Distinguishing borderline personality disorder
from bipolar disorder: differential diagnosis and implications. Am J
Psychiatry 153:1202–1207, 1996
Cardasis W, Hochman JA, Silk KR: Transitional objects and borderline
Clarkin JF, Hull JW, Yeomans F, et al: Antisocial traits as modifiers of treat-
ment response in borderline inpatients. J Psychother Pract Res
3:307–312, 1994
Coccaro EF, Kavoussi RJ: Biological and pharmacological aspects of
borderline personality disorder. Hosp Community Psychiatry
42(10):1029–1033, 1991
Cooper SH, Perry JC, Hoke L, et al: Transitional relatedness and border-
line personality disorder. Psychoanalytic Psychology 2(2):115–128,
1985
Dolan RT, Krueger RF, Shea MT: Co-occurrence with syndrome disor-
ders, in Handbook of Personality Disorders. Edited by Livesley WJ.
New York, Guilford, 2001, pp 84–106
Fink P: Psychiatric illness in patients with persistent somatization. Br J

Psychiatry 166:93–99, 1995
Fyer MR, Frances AJ, Sullivan T, et al: Comorbidity of borderline person-
ality disorder. Arch Gen Psychiatry 45:348–352, 1988
Gilligan C: In a Different Voice: Psychological Theory and Women’s De-
velopment. Cambridge, MA, Harvard University Press, 1982
Gunderson JG, Chu JA: Treatment implications of past trauma in border-
line personality disorder. Harv Rev Psychiatry 1:75–81, 1994
Gunderson JG, Phillips KA: A current view of the interface between bor-
derline personality disorder and depression. Am J Psychiatry 148:
967–975, 1991
atry 150:19–27, 1993
Gunderson J, Zanarini M, Kisiel C: Borderline personality disorder: a re-
view of data on DSM-III-R descriptions. J Personal Disord 5:340–352,
1991
Gunderson JG, Triebwasser J, Phillips KA, et al: Personality and vulnera-
bility to affective disorders, in Personality and Psychopathology. Ed-
ited by Cloninger CR. Washington, DC, American Psychiatric Press,
1999, pp 3–32
Gunderson JG, Bender D, Sanislow C, et al: Plausibility and possible de-
terminants of sudden “remissions” in borderline patients. Psychiatry
66:111–119, 2003
Gunderson JG, Morey LC, Stout RL, et al: Major depressive disorder and
borderline personality disorder revisited: longitudinal interactions.
J Clin Psychiatry 65:1049–1056, 2004
Gunderson JG, Weinberg I, Kueppenbender KD, et al: Descriptive and
longitudinal observations on the relationship of borderline person-
ality disorder (BPD) and bipolar disorders. Am J Psychiatry 163:
1173–1178, 2006
Hatzitaskos PK, Soldatos CR, Sakkas PN, et al: Discriminating borderline
from antisocial personality disorder in male patients based on psy-
chopathology patterns and type of hostility. J Nerv Ment Dis 185:442–
446, 1997
Herkov MJ, Blashfield RK: Clinician diagnoses of personality disorders:
evidence of a hierarchical structure. J Pers Assess 65:313–321, 1995
Herman J: Trauma and Recovery. New York, Basic Books, 1992
Herman JL, Perry JC, van der Kolk BA: Childhood trauma in borderline
Herzog DB, Keller MB, Lavori P, et al: The prevalence of personality dis-
orders in 210 women with eating disorders. J Clin Psychiatry 53:147–
153, 1992
Horton PC, Louy J, Coppolillo HP: Personality disorder and transitional
relatedness. Arch Gen Psychiatry 30:618–622, 1974
Hudziak JJ, Boffeli TJ, Kreisman JJ, et al: Clinical study of the relation of
borderline personality disorder to Briquet’s syndrome (hysteria), so-
matization disorder, antisocial personality disorder, and substance
abuse disorders. Am J Psychiatry 153:1598–1606, 1996
15:641–685, 1967
Kernberg O: Severe Personality Disorders: Psychotherapeutic Strategies.
New Haven, CT, Yale University Press, 1986
Khantzian EJ: The self-medication hypothesis of addictive disorders: fo-
cus on heroin and cocaine dependence. Am J Psychiatry 142:1259–
1264, 1985
Kopacz DR, Janicak PG: The relationship between bipolar disorder and
personality. Psychiatr Ann 26:644–650, 1996
Kurtz JE, Morey LC: Negativism in evaluative judgments of words among
depressed outpatients with borderline personality disorder. J Per-
sonal Disord 12:351–361, 1998
Labbate LA, Benedek DM: Bedside stuffed animals and borderline per-
sonality. Psychol Rep 79:624–626, 1996
Links PS, Steiner M, Boiago I, et al: Lithium therapy for borderline pa-
tients: preliminary findings. J Personal Disord 4:173–181, 1990
Livesley WJ, Jackson DN, Schroeder ML: A study of the factorial structure
of personality pathology. J Personal Disord 3:292–306, 1989
Livesley WJ, Jackson DN, Schroeder ML: Factorial structure of traits de-
lineating PDs in clinical and general population samples. J Abnorm
Psychol 101:432–440, 1992
Livesley WJ, Jang KL, Vernon PA: Phenotypic and genetic structure of traits
delineating personality disorder. Arch Gen Psychiatry 55:941–948, 1998
Lobel L: A study of transitional objects in the early histories of borderline
adolescents. Adolesc Psychiatry 9:199–213, 1981
McGlashan TH, Grilo CM, Skodol AE, et al: The collaborative longitudi-
nal personality disorders study: baseline Axis I/II and II/II diagnos-
tic co-occurrence. Acta Psychiatr Scand 102:256–264, 2000
Morrant JCA: The wing of madness: the illness of Vincent van Gogh. Can
J Psychiatry 38:480–484, 1993
Morris H, Gunderson JG, Zanarini MC: Transitional object use and bor-
derline psychopathology. Am J Psychiatry 143:1534–1538, 1986
Nace EP: Substance use disorders and personality disorders: comorbidity.
Psychiatr Hosp 20:65–69, 1989
Nadelson T: False patients/real patients: a spectrum of disease presenta-
tion. Psychother Psychosom 44:175–184, 1985
Paris J: Antisocial and borderline personality disorders: two separate di-
agnoses or two aspects of the same psychopathology? Compr Psychi-
atry 38:237–242, 1997
Paris J, Gunderson JG, Weinberg I: The interface between borderline

personality disorder and bipolar spectrum disorders. Compr Psychi-
atry 48:145–154, 2007
Perry JD, Perry JC: Conflicts, defenses, and the stability of narcissistic per-
sonality features. Psychiatry 67:310–330, 2004
Plakun EM: Distinguishing narcissistic and borderline personality disor-
ders using DSM-III criteria. Compr Psychiatry 28:437–443, 1987
Rinsley DB: A comparison of borderline and narcissistic personality dis-
orders. Bull Menninger Clin 48:1–9, 1984
Robins LN: Deviant Children Grow Up. Baltimore, MD, Williams &
Wilkins, 1966
Rogers JH, Widiger TA, Krupp A: Aspects of depression associated with
borderline personality disorder. Am J Psychiatry 152:268–270, 1995
Ronningstam E, Gunderson J: Differentiating borderline personality dis-
order from narcissistic personality disorder. J Personal Disord 5:225–
232, 1991
Sansone RA, Sansone LA, Fine MA: The relationship of obesity to border-
line personality symptomatology, self-harm behaviors, and sexual
abuse in female subjects in a primary-care medical setting. J Personal
Disord 9:254–265, 1995
Siever LJ, Davis KL: A psychobiologic perspective on the personality dis-
orders. Am J Psychiatry 148:1647–1658, 1991
Silverman JM, Pinkham L, Horvath TB, et al: Affective and impulsive per-
sonality disorder traits in the relatives of patients with borderline per-
sonality disorder. Am J Psychiatry 148:546–554, 1991
Stern J, Murphy M, Bass C: Personality disorders in patients with somati-
zation disorder: a controlled study. Br J Psychiatry 163:785–789, 1993
Stone MH: The Fate of Borderline Patients. New York, Guilford, 1990a
Stone MH: Treatment of borderline patients: a pragmatic approach. Psy-
chiatr Clin North Am 13:265–285, 1990b
Suppes T, Dennehy DB: Evidence-based treatment of bipolar II disorder.
Swartz M, Blazer D, George L, et al: Estimating the prevalence of border-
line personality disorder in the community. J Personal Disord 4:257–
272, 1990
Tyrer P, Gunderson J, Lyons M, et al: Special feature: extent of comorbid-
ity between mental state and personality disorders. J Personal Disord
11:242–259, 1997
van Meekeren E: Starry Starry Night: Life and Psychiatric History of Vin-
cent van Gogh. Amsterdam, The Netherlands, Benecke NI, 2003
Westen D: Towards a revised theory of borderline object relations: con-
tributions of empirical research. Int J Psychoanal 71:661–693, 1990
Westen D: Divergences between clinical and research methods for assess-
ing personality disorders: implications for research and the evolution
of Axis II. Am J Psychiatry 154:895–903, 1997
Westen D, Moses MJ, Silk KR, et al: Quality of depressive experience in

borderline personality disorder and major depression: when depres-
sion is not just depression. J Personal Disord 6:382–393, 1992
White CN, Gunderson JG, Zanarini MC, et al: Family studies of border-
line personality disorder: a review. Harv Rev Psychiatry 11:8–19, 2003
Widiger TA, Trull TJ, Clarkin JF, et al: A description of the DSM-III-R and
DSM-IV personality disorders with the five-factor model of personal-
ity, in Personality Disorders and the Five-Factor Model of Personality.
Edited by Costa PT, Widiger TA. Washington, DC, American Psycho-
logical Association, 1994, pp 41–58
Widom CS: Child abuse, neglect and witnessing violence, in Handbook
of Antisocial Behavior. Edited by Stoff DM, Breiling J, Maser JD, et al.
New York, Wiley, 1997, pp 159–170
Zanarini MC, Gunderson JG: Differential diagnosis of antisocial and bor-
derline personality disorders, in Handbook of Antisocial Behavior.
Edited by Stoff DM, Breiling J, Maser JD, et al. New York, Wiley, 1997,
pp 83–91
Zanarini MC, Gunderson JG, Marino MF, et al: Childhood experiences of
borderline patients. Compr Psychiatry 30:18–25, 1989
Zanarini MC, Frankenburg FR, Dubo ED, et al: Axis I comorbidity of bor-
derline personality disorder. Am J Psychiatry 155:1733–1739, 1998a
Zanarini MC, Frankenburg FR, Dubo ED, et al: Axis II comorbidity of
borderline personality disorder. Compr Psychiatry 39:296–302,
1998b
Chapter 3
OVERVIEW OF TREATMENT
Historical Overview
Changes in perspectives on the treatment of borderline personality disor-
der (BPD) since the 1970s parallel the larger shifts in psychiatry, in health
care services, and (as noted in Chapter 1) in the diagnostic construct it-
self. Psychiatry has become more medicalized, health care services have
become more diagnosis specific and cost conscious, and the borderline
diagnosis has been validated. Psychoanalysts made the initial observations
about borderline patients largely on the basis of the uniquely vexing clin-
ical problems that these patients created in testing boundaries and in re-
gressing when in unstructured settings. When Kernberg (1968) and
especially Masterson (1971, 1972) wrote optimistic reports about the
treatability of this disorder, they inspired a tide of ambitious long-term,
psychoanalytically informed treatments in both inpatient and outpatient
settings. As was shown in Figure 1–5, since 1968, 56 books about psycho-
analytic psychotherapies have been written, cresting with 19 between
1990 and 1994 (as found in a Library of Congress database search). In in-
stitutional settings, most notably in prestigious private hospitals, long-
term units devoted to treating BPD had developed by the 1980s. Both the
psychoanalytic outpatient psychotherapies and these long-term inpatient
treatments were based on ambitious hopes for curative changes.
Even as the swell of intensive long-term psychoanalytic treatments was
peaking, the excesses, limitations, and narrowness of the approach were
being recorded. Many clinicians, including notable analysts (Adler 1981,
1986; H. J. Friedman 1969; Zetzel 1971), thought that long-term institu-
tional care was regressive and that short-term stays had advantages. Oth-
ers who had worked in long-term settings noted that this care often led to
intractable control struggles (Gunderson 1984), such as Kaysen (1993)
65
described in her best-selling book Girl, Interrupted. In outpatient settings,

similar observations occurred. Even psychoanalytic therapies by experts
often ended abruptly with the borderline patient’s flight (Waldinger and
Gunderson 1984), and these therapies were, in any event, economically
or logistically unfeasible for all but a few.
Important shifts in psychiatry from a psychoanalytic to a biological
paradigm and from a clinical to an empirical base of knowledge changed
treatment standards for borderline patients. The foundations on which
ambitious long-term psychoanalytic therapies were built during the 1970s
(i.e., compelling theories and expert opinions) were permanently under-
mined during the 1980s. Those theories and expert opinions were re-
framed as hypotheses, whose continued implementation would require
empirical nourishment. Since the 1980s, empiricism and reliance on
standardized diagnostic and outcome measures have established new
foundations for testing treatment efficacy. These foundations showed
that different treatments may effect changes in different and discrete sec-
tors of psychopathology (e.g., mood, cognition, and behavior). More-
over, the continued growth of biological psychiatry underscored the
potential for medications to change symptoms that often had been resis-
tant to psychological therapies. These changes were reflected in a series
of controlled medication trials conducted during the 1980s with reliably
diagnosed borderline patients.
In the 1990s, clinical and research attention slowly turned to socio-
therapeutic modalities—that is, the role of partial hospital, group, family,
and cognitive-behavioral therapies. This change was long overdue: that
the social functioning of borderline patients was as handicapped as that
of schizophrenia patients and far worse than that of depressed patients
had already been documented in 1975 (Gunderson et al. 1975). I believe
that this delayed attention to treating the severe social dysfunction of bor-
derline patients stemmed from the resistance to such attention found in
most borderline patients themselves. They often actively avoid or react
with disdain to talk of maladaptive functioning, or of the need for voca-
tional or social skills, as if these factors are unimportant or as if attention
to these issues means the therapist does not really care about the patient.
No doubt advances in overcoming such resistance have been pushed by
more awareness of deficits, as well as by deinstitutionalization: the social
rehabilitative needs of these patients had been masked within hospitals
and by the psychoanalytic focus on intrapsychic issues.
An informed approach to treating BPD now involves the thoughtful
deployment of multiple modalities. It is in the context of these historical
developments in treating BPD that modern professionals are confronted
with the need for much more complicated treatment planning and the
Overview of Treatment ❘ 67
need to integrate the component modalities. In this chapter, I offer an

overview of the selection and conceptualization of treatment services.
Generic Therapeutic Processes and

the Functions They Serve
A conceptual framework about therapeutic processes helps clinicians and
patients understand what functions can be served by different therapeutic
programs. In this section, five therapeutic functions are described: contain-
ment, support, structure, involvement, and actualization. This accord is a
revision of an earlier thesis (Gunderson 1978). The first two functions, con-
tainment and support, are often performed unilaterally by staff to or at least
for patients; therefore, they are treatments. The latter three are usually per-
formed in collaboration with patients, requiring their consent and desire.
Table 3–1 summarizes how these functions differ in goals, implementation,
and applicability to borderline patients (see also Figure 3–1; levels of care
referenced in Table 3–1 and Figure 3–1 are fully discussed in Chapter 5).
Containment
Containment functions to preserve or enhance the physical well-being of
people. For borderline patients, containment usually involves securing
their safety by provision of asylum from stressful situations, sometimes
even with locked doors and supervision, but usually only with monitored
food and medications. Containment refers to external imposition of con-
trol and is the most concrete form of what Winnicott (1965) referred to as
a “holding environment.” It alleviates the responsibility for self-control
and offers borderline patients a basic form of caregiving. For borderline
patients who feel angry about their responsibilities for caring for them-
selves, too much containment may become habit forming, thereby cre-
ating a regressive option that is antitherapeutic. For most borderline
patients, the initial relief at containment is followed by fears of being con-
trolled. (Medications often dramatize such a shift; see Chapter 6). Dur-
ing the course of successful treatment, borderline patients internalize
controls so that by the time they are nearly well, the holding environment
can be created and sustained by talking, and by the time they are well,
they can sustain the “holding” function intrapsychically.
Support
Support functions to make patients feel better and to enhance their self-
esteem. Support can be given by accommodating patients’ limitations
(e.g., tutors for those with learning disabilities or clarification for those
with poor reality testing). Support is most direct when it consists of as-
TABLE 3–1. Therapeutic functions

FUNCTION IMPLEMENTATION INDICATIONS
Containment Level IV Lack of self-control,

Medications dangerousness
Monitoring
Support All levels (IV–I) Self-care deficits, low
Caregiving self-esteem
Attention and validation (all
modalities)
Structure Levels IV–II Skill deficits,
Direction: cognitive-behavioral maladaptive
therapies and behaviors
psychopharmacology
Education: patient and family
Contingencies
Involvement Levels III–I Maladaptive or
One-to-one interaction insecure
Milieu, group, and family interpersonal
therapies relationships, social
isolation
Actualization Level I Identity or self-image
Individual psychotherapy: problems, intimacy
interpretation, empathy or risk aversion
Note. Levels of care: level IV =hospital; level III= residential/partial hospital/day
treatment; level II= intensive outpatient; level I= outpatient.
sistance with travel or feeding or when it involves verbal activities such as

direction or education. It is most obvious when it consists of praise or re-
assurance. A very basic supportive technique is validation—affirming the
reality of patients’ perceptions or the justification for their feelings. Val-
idation usually begins with an empathic recognition of patients’ pain and
past misfortunes. (This does not mean affirming that patients are victims
of malevolence [see Chapters 1 and 10], only that they suffered from
what were “unfair” burdens or stresses.) Validation allows borderline pa-
tients to develop closer relationships and in time diminishes distrust. Val-
idation has emerged as a central component of all of the empirically
validated individual therapies (Chapters 10–12).
For borderline patients, attention is always a critically important sup-
portive process. Even negative attention is better than inattention. Often
Overview of Treatment
❘
FIGURE 3–1. Multiple modalities and step-down services.
69
borderline patients appeal for direct support through somatic complaints

or through their accounts of prior mistreatment. Seeing the borderline
patient as handicapped, disabled, or mentally ill is a way to enhance sym-
pathetic responses by staff or by families.
Although the absence of significant amounts of support often will
evoke attributions of rejection or cruelty in borderline patients, too much
support can evoke unrealistic expectations and encourage “pseudother-
apies” (Bateman and Fonagy 2004) and distract from the “dialectical”
need for self-change (Linehan 1993). As with containment, support can
become less behavioral and more verbal as borderline patients improve.
Support is, however, a process that is a necessary part of all therapeutic ac-
tivities throughout treatment. The necessity of this function has some-
times been overlooked in psychoanalytic psychotherapies, but its role
with borderline patients has become well documented.
Structure
Structure functions to make the environment predictable—as simple and
repetitious as possible. It involves organizing the patient’s time, place,
and person. Structure is an impersonal holding, neither invasive nor ne-
glectful. This therapeutic function is served by schedules, clarity of roles
and goals, privilege systems, controls, contracts, and clear consequences
for behaviors. It is most important in addressing the borderline individ-
ual’s socially maladaptive behaviors, such as rages or impulsivity. Struc-
ture has particular importance for BPD: its absence invites regression and
projection. It is a more central component of cognitive-behavioral ther-
apies (see Chapter 11) than dynamic therapies, and it is usually appealing
and relieving to borderline patients.
Examples of structure would be a contract regarding wrist slashing
that would mandate a visit to an emergency department and one missed
therapy session as consequences. More generally, it would mean starting
and ending sessions on time, always at the same site, and in the same
seats. Within sessions, more structure is desirable early in treatment—for
example, consistently recounting reactions to the last visit and to any in-
tervening contacts or consistently reviewing work or health issues.
Involvement
Involvement evolves from the structured interpersonal interactions with
treaters and other patients. Involvement strengthens tolerance for inter-
personal relationships and identifies and modifies maladaptive interper-
sonal traits (e.g., devaluing or idealizing). Examples are development of
shared goals and collaboration on treatment planning. All group activi-
ties make involvement a central process, especially those during which a
discussion of group members’ ways of relating to one another is involved.

Therapeutic communities as described by Main (1946) and Jones (1952,
1956) made involvement the therapy.
For borderline patients, involvement is facilitated by interpreting
symptomatic behaviors as being interpersonally motivated: “You threat-
ened the nurse after your mother failed to call”; “You retreated to bed
right after your therapist announced her vacation”; “You avert your eyes
whenever I appear disapproving.” To return to the wrist-slashing exam-
ple, the process of involvement would be invoked by interpreting the ac-
tion as having sadistic or controlling intentions.
Borderline patients generally fear aloneness and are hungry for in-
volvement, but the presence of involvement usually prompts other anxi-
eties. Lewin and Schulz (1992) titled a book Losing and Fusing to underscore
this dilemma and to give equal attention to the borderline individual’s anx-
ieties about too much involvement—his or her intolerance of togetherness.
As noted in Chapter 1, I believe that the fear of losing (i.e., being rejected or
abandoned) is the more central and specific aspect of borderline patients’
problems.
Actualization
Actualization affirms and consolidates patients’ uniqueness—their individ-
uality—and helps them use or fulfill their potential. These goals often in-
volve customizing treatment by one-to-one talks, by attention to patients’
history, and by new learning—encouraging patients to extend themselves
into areas of uncertain competence or consequence.
For borderline patients, the process of actualization often begins by a
therapist’s developing a vision of the patient’s potential for health, com-
petence, and happiness. The process also underscores the individual’s
uniqueness and the significance of his or her life history in creating a life
narrative in which his or her own agency is recognized. It makes patients
feel understandable, less toxic, and fortunate to be alive.
Multiple Modalities and Step-Down Services:

An Overview
Whereas virtually all borderline patients require different levels of treat-
ment (see Chapter 5) and multiple modalities (Chapters 6–12), few will
require them all. There are several ways to think about the interactions,
complementarity, and sequencing of different modalities of treatment.
The usual sequencing of modalities moves from biological therapies to
sociotherapies and finally to psychological (or intrapsychic) approaches.
As shown in Figure 3–1, the five therapeutic functions discussed earlier
are also offered in progression through the various levels and modalities
of care. The sequencing is anchored by considerations of the modalities’
usual duration, their expectable costs and benefits, their relative levels of
empirical support, and their replicability. In practice, the sequence often
involves a shift from the highest level of care (hospital settings) to the low-
est (outpatient care) via a series of intermediary step-down services (e.g.,
residential care, day care). These services and patients’ relative length of
stay are schematized in Table 5–1 and further described in Chapter 5.
Sociotherapies
The most common pair of outpatient modalities (level I in Figure 5–1)—
psychotherapy and psychopharmacology—overlook the significant con-
tribution that sociotherapies can make to treating BPD. Sociotherapies re-
fer to that middle range of therapies that more directly addresses the
observable social impairment and social adjustment issues—the issues
that Links (1993) says require a psychiatric rehabilitation model (Table 3–
2). Although the social rehabilitative needs of borderline patients are
clearly central to the structured community or milieu therapy aspects of
residential (level III) services (as discussed later), these rehabilitative
needs are usually not addressed by the time patients begin outpatient
care. At present, some manual-guided outpatient sociotherapies for BPD
have been established—namely, for some forms of family therapy (Chap-
ter 8) and group therapy (Chapter 9). Very little has been written about
and few have recognized the role of vocational rehabilitation. These mo-
dalities or others that improve social skills and adaptation need to become
more central to treatment plans and more available in outpatient clinics.
Establishing Goals:
The Expectable Sequence of Change
The growth of cost-benefit considerations, the contingence of a new stan-
dard for empirical validation, and the expansion of the cognitive-behav-
ioral paradigm have each contributed to the still-growing awareness that
treatments should have goals and that setting those goals constitutes an
essential first step in planning treatments. Within the context of treating
BPD, we have moved from the era in which goals were long-term objec-
tives stated in abstract language (improve object relations, decrease re-
liance on splitting) to less abstract but still broad goals (develop more
independence, diminish impulsivity) to the current era in which short-
term and more specific goals can be identified (learn to control temper,
ask for help). This progression may reflect stages that are inherent in the
maturation of therapeutics for any disorder, but without question these
TABLE 3–2. Goals of sociotherapies

Enhance social skills
Manners
Listening
Comfort (e.g., chitchat)
Enhance self-awareness in interpersonal situations
Tendencies to misattribute
Typical reactions to praise, criticism, or competition
Tendencies to disclose (feelings or attitudes) vs. isolate/withdraw vs.
deceive/mislead
Goals specific to groups
Confront undesirable styles
Diminish aggression and defensiveness
Enhance awareness of effects on others
Enhance disclosure, expression of feelings, and assertiveness
Improve tolerance and understanding of others
Goals specific to families
Identify indirect or covert styles of communication
Enhance communication
Clarify or modify reinforcement patterns
Validate or invalidate attributions
Clarify motives, intentions, and effects
changes reflect pragmatically desirable progress. As described previously

(Gunderson et al. 1993), the continued growth of knowledge about ther-
apeutic effectiveness for borderline patients depends on specifying
increasingly discrete and more time-limited indices of change in which
outcome can be measured.
To develop treatment goals and to assess whether existing treatments
for BPD are making timely progress, it is useful to have a conceptual
framework for processes of change. Emerging from experience (Gunder-
son 1984) and research (Gunderson et al. 1993; Waldinger and Gunder-
son 1989) and buttressed by a review of related literature (Kopta et al.
1994; Lanktree and Briere 1995), a sequence in which changes can be ex-
pected is proposed (Table 3–3). Of particular value is the conclusion by
Kopta et al. (1994) that although patients’ subjective states can change
within weeks, characterological traits and self-concepts cannot be ex-
pected to change before a year in therapy. This schema is supported in a
meta-analysis of the effectiveness of psychosocial therapies for BPD. Perry
and Bond (2000) noted that subjective complaints, mood states, and glo-
bal function improved more in the first year of treatment than did social
function and interpersonal relationships. Converging evidence from ex-
isting research on schizophrenia supports the general validity of the
sequence and timetable for changes suggested here for BPD. With schizo-
phrenia patients, symptom remission, diminished family conflict, and im-
proved social skills function can be accomplished within a year (Hogarty
et al. 1986). This treatment still, however, leaves the successfully treated
patients interpersonally isolated and anhedonic, thus setting the stage
for individual therapies (Hogarty et al. 1997).
Table 3–3 and Figure 3–2 elaborate on the sequence of and approxi-
mate timetable for changes that are expectable in successful treatments of
BPD. It is important to recognize that this timetable is schematic—that
there are significant variations, depending on the stage from which bor-
derline patients start treatment (e.g., some are very unaware of anger, some
are successfully employed). This account of expectable changes is revisited
elsewhere in this book in the sequence of therapeutic functions (earlier in
this chapter) and in describing levels of care (Chapter 5), the sequence of
changes in family intervention (Chapter 8), and the sequence of changes
within psychotherapies (Chapters 10 and 12). The clinical value of identi-
fying the sequence and timetable for expectable changes is that therapists,
patients, and families can make more discerning judgments about thera-
peutic effectiveness (Sidebar 3–1). Failure to see the “expected” change
does not mean that such therapies are not being beneficial. It means that
the question should be raised whether the therapeutic services could be
improved. The best way to address these issues is by consultation.
Sidebar 3–1: Should Consumers Receive

Progress Reports?
Progress reports may feel critically important for consumers
(patients themselves or their loved ones), but they are often
hesitant to request them. Clinicians can expect that in-
creased consumer education (notably via the Web) and
empowerment will expand the frequency of such requests
and also will magnify the frequency and urgency with
which consumers will question a clinician’s answers when
they are based on the assumption of professional authority.
When progress is obvious, inquiring about it is less likely to be
TABLE 3–3. The framework for expectable changes

EXPECTABLE
AREAS OF RELEVANT TIME FOR
DISTURBANCE CHANGES INTERVENTIONS CHANGE
Subjective states Dysphoric Concerned Weeks

feelings: attention,
anxiety, validation, reality
depression, testing, problem
irritability solving
Behavior Deliberate Pattern recognition, Months
self-harm, chain analyses,
impulsivity, clarification of
rages defensive purposes
and maladaptive
consequences
Interpersonal Devaluation, Confrontation, 6–18
style assertiveness pattern months
recognition, here-
and-now,
interpretations
Intrapsychic Self-esteem, Defense and >2 years
organization trust transference
analysis, corrective
relationships, and
new experiences
secondary to prior
changes
Source. Adapted from Gunderson JG, Gabbard GO: “Making the Case for Psy-
choanalytic Therapies in the Current Psychiatric Environment.” Journal of the
American Psychoanalytic Association 47(3):679–704, 1999.
needed. But progress often is not obvious, and treaters of-

ten find it very difficult to answer inquiries. Psychotherapists
can be so involved in the meaningful interactions they
have with their patients that they are unprepared to de-
scribe whether those interactions are leading to improve-
ments in their patients’ social function or symptoms.
Therapists should feel a responsibility to answer such
questions. If the question comes from the borderline pa-
tient, it is cause for a thoughtful discussion. If the question
comes from anyone else, the therapist may hide behind
the constraints of confidentiality. But, these constraints of
76
❘
BORDERLINE PERSONALITY DISORDER: A CLINICAL GUIDE
FIGURE 3–2. Sequence and timetable for expectable changes in patients with borderline personality disorder dur-
ing therapy.
Arrows following the spheres of change indicate continuation from prior year. Arrows following the five therapeutic functions show the years
of treatment in which the functions should be performed.
aParticularly notable markers of change.
confidentiality rarely pose a real obstacle to giving

progress reports. Almost always, the patient only needs re-
assurance that the therapist’s response to others involves
only a disclosure of his or her views about progress—views
the patient will already have previewed—and not a disclo-
sure of confidences shared in sessions.
Therapists who will not respond to inquiries about treat-
ment progress will justifiably magnify a consumer’s con-
cerns. Consultation becomes desirable. Most experienced
therapists will use such inquiries as opportunities to hear
why the patient or the patient’s significant others have
concerns about progress. Clinicians should, of course, be
proactive in addressing their own concerns about insuffi-
cient progress with their patients. Treaters themselves may
welcome the opportunity to address reasons that progress
has in fact been disappointing when they believe it is a
result of treatment-interfering behaviors (e.g., missed
appointments, substance abuse). Such inquiries from a pa-
tient’s significant others set the stage for subsequent dis-
cussions that will involve the patient and can help create a
better holding environment. Consumers should be encour-
aged to learn about the overall sequence of expectable
changes, described in this chapter and illustrated in Figure
3–2. More public information and greater consumer advo-
cacy (see Chapter 13) are positive developments that cli-
nicians need to become comfortable with.
Significant variations are seen between patients in how changes in

one of the spheres shown in Figure 3–2 affect changes in others (e.g.,
some patients may become very dependent on their therapist during
times when they are unemployed or may become suicidal years after start-
ing treatment when feeling acutely abandoned). Still, the progression of
changes is reasonably predictable. Following are more detailed accounts
of what changes involve within each of the four spheres of change shown
in Figure 3–2. This figure also shows how these changes relate to the pri-
macy of the different therapeutic functions described earlier.
Changes Within Four Spheres

Affects or Emotions
Some negative affect states (anxiety, despair, anger, fears) in borderline
patients are among those that can change the soonest. Most notably, the
states of desperation or panic engendered by abandonment and alone-
ness can be dramatically reduced (within hours even) by involvement in
an adequate holding environment (Winnicott’s [1965] term, as mentioned
previously in this chapter, connoting a situation in which a person feels

contained and safe). Although the rapid relief of dysphoric affects has al-
ways characterized the borderline patient’s response to hospitalization,
the relief can be sustained by adequate aftercare treatment or by medi-
cations (Chapters 6 and 7) or sometimes by increased familial supports.
Similarly, states of rage will be dramatically reduced by the presence of
holding relationships. Here, however, the rages give way to ongoing oc-
currences of ready irritability and impatience, traits whose change in re-
sponse to medications or other treatments is not very predictable.
The depressed symptoms of hopelessness, worthlessness, or despair
about the future remain intermittent states that can gradually diminish in
intensity. Depressed moods may persist or actually become more evident
when borderline patients correct the behavior problems that have served
defensive functions. Thus, the use of carbamazepine (Chapter 7) or dialec-
tical behavior therapy (DBT; Chapter 11) could change behaviors without
improving depressed symptoms. It is a step forward when these depressive
mental states become connected to longings for care or feelings of loneli-
ness. Both psychoanalytic partial hospital programs (Chapter 5) and
Adler/Kohut self psychological psychotherapy (Chapter 12) have effected
improvements in the first year. Still, persistent dysphoric states typify many
borderline patients (Zanarini et al. 1998) and, in my experience, only give
way significantly when patients resolve their splitting—that is, when they
learn to own their own hostilities comfortably and accept them as an ap-
propriate part of their relationships. The earliest I have seen this occur is in
the fourth year of treatment (see Chapter 12 for discussion).
Of all the negative affects typifying borderline patients, emptiness
seems to be the most resistant to change. After years with improved func-
tioning, some borderline patients do report that it bothers them less or
that they feel it less often. This change seems to be a gradual process that,
in theory, relates to the internalization of good experiences of being cared
for, either within intensive therapy or in relationships outside therapy.
Behaviors: Impulse/Action Pattern

Medications can help improve behaviors in borderline patients. Usually
the benefits are not dramatic. As noted in Table 11–1, DBT identifies sui-
cidal behavior changes as the highest priorities for change—for obvious
reasons of safety and survival and to sustain the therapy that will, it is
hoped, then improve the quality of life. Traditionally, psychodynamic
therapies have accorded behavior change secondary status, believing that
self-destructive or suicidal behaviors will diminish by themselves when pa-
tients acquire either insight into their motivations or stable relationships.
In both types of therapy, the process of behavior change begins by helping
patients to recognize—or learn—that behaviors that have been habitual

and may have previously had adaptive functions are counterproductive.
This recognition occurs through having unwanted consequences clari-
fied (e.g., being rejected after being too needy) or enacted (being given
less attention after cutting oneself) or through learning new coping strat-
egies (see Chapters 9 and 11). The primary targets for behavior change
involve self-destructive behaviors such as cutting, substance abuse, or eat-
ing disorders, but others involve alienating behaviors such as yelling,
demanding, or withdrawal. Readers will note throughout this book that
almost every modality has now empirically documented the ability to
bring about behavior change within the first year of treatment.
Social Function: Impairment

The clearest evidence of borderline patients’ social impairment is their
unemployment rate, a rate similar to that seen in individuals with schizo-
phrenia (Gunderson et al. 1975)—despite apparent social and intellec-
tual abilities that should enable individuals with BPD to do better. It has
become clear that even for borderline patients whose symptoms systemat-
ically remit, serious impairment in their social function persists (Skodol et
al. 2002, 2005). Structure is needed to diminish an elaboration of affects
or a decreased reality sense (Singer and Larson 1981). The simplest and
most common form of structure involves a steady job. But remaining vo-
cationally or otherwise socially dysfunctional has many determinants.
Some patients resist attaining employment because it threatens secondary
gains (e.g., attention, sympathy, low expectations) and generates aban-
donment fears as well as fears of failure. To make employment as palatable
as possible, work options with a high likelihood of success should be en-
couraged. Low-competition, low-demand, and high-structure work set-
tings are optimal starting points (see Sidebar 5–3). At present, there has
been no BPD-specific vocational rehabilitation, and of the existing empir-
ically validated BPD-specific therapies, none has shown particular advan-
tages for vocational performance. Nonetheless, as shown in Figure 3–2,
my experience is that progress in this area should be evident in the course
of effective therapies by the second year for most borderline patients.
Social impairment in interpersonal relationships is also a focus of treat-
ment. An early goal in this area is to develop a network of friendly but some-
what superficial acquaintances. Group living situations may be more
feasible and less high risk than are new romances, in which too many needs
or expectations are ignited. When behavior problems diminish, many bor-
derline patients become quite socially phobic. To avoid expected rejections,
they isolate themselves, even to the point of qualifying for the diagnosis of
avoidant personality disorder (Zanarini et al. 2007). True friendships, with-
out dependency but built on shared interests and depth of caring, are tri-
umphs that signal someone as no longer having BPD. Although this may
occur by the third year of treatment, such progress would be unusual.
Relationship With Treaters

More will be said about the borderline patient’s relationship with treaters
throughout this book, and certainly this aspect of treatment has received
extensive coverage in the literature. Initially, borderline patients’ relation-
ships with treaters are distrustful or split (i.e., idealized or devalued)
(Agrawal et al. 2004; Butler et al. 2002; Gunderson and Lyons-Ruth 2008;
Shedler and Westen 2004). Idealization is helpful and can be promoted
by validation and the promise of relief from dysphoric moods. The proac-
tive “I can help you” approach offered by psychopharmacologists or cog-
nitive-behavioral therapists encourages hope and perhaps idealization.
More sustained trust can be engendered by reliability, availability, and re-
silience in the face of challenges. Clinicians of all sorts need to establish
their trustworthiness. This sets the stage for emotional dependency—a
good basis for case management, psychopharmacology, or exploratory
psychotherapy. The subsequent changes seen in Figure 3–2 are relevant
primarily to long-term psychotherapy (Chapters 10–12). Even in DBT, the
third-stage targets of increased self-respect and pursuit of individual goals
(Linehan 1993) involve intrapsychic changes that are consistent with the
overall sequence of changes expectable from both generic and BPD-
specific observations about change described in this chapter.
The Initial Structuring of Treatment

The key concept in starting treatments is to establish structure (goals,
roles, organization). Specific ways this can be implemented involve iden-
tifying a primary clinician, setting short-term goals, establishing ade-
quately supportive context, and providing psychoeducation.
Primary Clinician
It is essential that a primary clinician be identified who will assume respon-
sibility for each patient’s safety and treatment. This person’s role inevitably
involves serving as case manager (see Chapter 4). The role also may include
being the patient’s psychotherapist, but only if the clinician has suitable
training and the patient indicates an interest in change (see Chapter 10).
Short-Term Goals
Short-term goals establish a task orientation for any therapy: it is for the
purpose of change. Realistic goals such as diminished anxiety and suicid-
ality, asking for help, balancing a checkbook, developing alliances, or

scheduling time should be targeted with time-limited treatment plans. At
the same time, long-term goals (e.g., tolerating aloneness, developing in-
timate relationships, and achieving career satisfaction) should be en-
couraged as possibilities.
Least Restrictive Safe Treatment Setting

Identifying the least restrictive safe treatment setting not only is cost ben-
eficial but also allows the most effective treatment. See Chapter 5 for a
discussion of this topic.
Psychoeducation
Although psychoeducational approaches for borderline patients had
been proposed many years ago (Benjamin 1993; Brightman 1992), this
approach is still not widely practiced or even seen as desirable (Ruiz-
Sancho et al. 2001). As noted in Chapter 1, I suggest that BPD patients
and those they live with should uniformly be familiarized with the diag-
nosis, including its expectable course, responsiveness to treatments, and
known pathogenetic factors. Psychoeducational methods are appropri-
ate and are generally welcomed by both patients and their families.
The psychoeducational approach is based on the hope that patients
diagnosed as having BPD or the people who live with or love them will
benefit from learning about the disorder (Table 3–4). It rests heavily on
the medical model of BPD as an illness. In this model, the behavior prob-
lems associated with BPD are sequelae to underlying neurobiological ab-
normalities over which they can exert only weak or inconsistent control.
This message is usually welcome to patients (G. Rubovszky, J.G. Gunder-
son, I. Weinberg, “Patients’ Reactions to Disclosure of the Borderline Per-
sonality Disorder Diagnosis,” unpublished manuscript, November 2007).
It is reassuring to know they are not alone with their disorder and that a
body of knowledge is available about this disorder and its treatment. It
also conveys hope, insofar as BPD generally has a good prognosis. In-
forming patients about their disorder can make them more aware of how
their feelings, behaviors, and thinking can cause problems. In my expe-
rience, it encourages intellectualization (a form of “mentalizing,” as in
Fonagy et al. 1991) and with this, a type of valuable constraint on action.
As noted elsewhere (Chapters 5 and 12), I am very explicit in making pre-
dictions about how BPD patients can expect to respond to forthcoming
situations (e.g., a vacation or a step-down in level of care). If priority is be-
ing given to treating comorbid Axis I conditions, such as substance abuse,
depression, posttraumatic stress disorder, or eating disorders (see Chap-
ter 2), useful cautions about the overall prognosis for that Axis I condi-
TABLE 3–4. Rationale for psychoeducation of patients with

borderline personality disorder
Patient’s right to know
Increases awareness of disorder—demystifies and destigmatizes
Diminishes sense of unique, unknown problems: knowledge that others
have similar problems
Enlists intellectual strengths and curiosity
Invites active participation in treatment planning
Establishes realistic hopes for change
Source. Adapted from Ruiz-Sancho et al. 2001.
tion can be given. As noted in Chapter 1, the simplest and most common
psychoeducational intervention involves the diagnosis itself. Patients usu-
ally welcome reading the DSM-IV-TR text (American Psychiatric Associa-
tion 2000) and describing how the criteria do or do not apply to them
(see Chapter 2). In a more general way, it helps to demystify and destig-
matize the diagnosis. By informing patients about treatment options and
the potential for change, psychoeducation helps establish realistic expec-
tations for treatment and a greater likelihood of complying with treat-
ment (G. Rubovszky, J. G. Gunderson, I. Weinberg, “Patients’ Reactions
to Disclosure of the Borderline Personality Disorder Diagnosis,” unpub-
lished manuscript, November 2007). Psychoeducation is often done dur-
ing early sessions with case managers (Chapter 4) or with individual
psychotherapy when a treatment plan or “contract” is being developed
(see Chapter 10).
Types and Sequence of Therapeutic Alliance

The concept of a therapeutic alliance helps frame the discussion of both the
initial engagement of borderline patients in all forms of therapy and the
subsequent longer-term processes within therapies. The concept of alliance
has special significance for BPD: at one time, an alliance was considered a
prerequisite for dynamic psychotherapy, which, if true, would in theory ren-
der many such patients unsuitable for that modality (Sidebar 3–2).
To guide our usage of the term alliance, Table 3–5 adopts definitions
of three types that occur sequentially (Greenspan and Sharfstein 1981;
Luborsky 1976). Defining roles and goals and establishing a concrete
framework for the treatment (schedule, fee, confidentiality) constitutes
the earliest form of alliance, the contractual alliance. It is relevant to all
TABLE 3–5. Three forms of therapy alliance

Contractual (behavioral). This form refers to the agreement between
patient and therapist on treatment goals and their roles in achieving
them. This type can be established in the first session, but it often takes
two or three.
Relational (affective/empathic). Emphasized by Rogerian client-
centered relationships. This form refers to patient’s experience of the
therapist as caring, understanding, genuine, and likable. This type
develops in the first 6 months.
Working (cognitive/motivational). The psychoanalytic prototype. In
this form, the patient is a reliable collaborator who can recognize
unwanted observation by a therapist as being well-intended. This type
forms gradually, vacillates within sessions, and is unlikely to be reliably
present for several years.
modalities of treatment. To a considerable extent, the problem of drop-

outs can be diminished by giving special attention to mutually agreed-on
expectations for the therapy.
Sidebar 3–2: Myths About Alliance With

Borderline Patients
A report provocatively titled, “The Myth of the Alliance With
Borderline Patients” (Adler 1979), following L. Friedman
(1969), argued that a working alliance (as in Table 3–5) de-
velops as an outcome only late in psychoanalytic psycho-
therapy with borderline patients—so late that they may in
fact no longer have a borderline personality. This thesis con-
trasted with an earlier analytic theorem, that the presence of
or capacity for a working alliance was prerequisite to engag-
ing in an exploratory, insight-oriented, transference-based
psychotherapy (Greenson 1965; Sterba 1934; Zetzel 1956).
This separation of the alliance from transference has
served conceptual purposes, but it is intrinsically mythical.
The ability to observe oneself collaboratively while sitting
with a therapist is itself based on a transference wherein the
patient’s suspension of disbelief and suspiciousness is based
on acquired expectations about caregiving relationships. It
is unlikely to be an expectation that has been “earned” by
virtue of experience with the therapist. Indeed, a good
working alliance within psychotherapy is based on a trans-
ference, presumably derived from early childhood experi-
ences, in which there was a secure attachment with
sufficient opportunity for self-expression and nonpunitive re-
sponse that positive expectations of such a relationship

can be operative with a relative stranger (Brenner 1988; Gill
1979; Hoffman 1998; Langs 1976).
The relational alliance (discussed in Chapter 12) is a type that is cen-

tral to most individual therapies, including cognitive-behavioral types,
and is often an important, albeit adjunctive, element of a psychopharma-
cologist’s, family therapist’s, or group therapist’s functioning.
The working form of alliance is the classical psychoanalytic model of
alliance. In this mature form, the patient and the therapist are joined by
their mutual interest in and attentiveness to a common task: the under-
standing of the patient. As described in Chapter 12, the presence of this
form evolves slowly and can dramatically disappear in sessions with bor-
derline patients, even years after therapy has begun.
Countertransference
No report about treating BPD can fail to note the strong countertransfer-
ence responses that such patients evoke and the frequency with which
those responses are destructive to therapies of all kinds. The classic paper
on countertransference hate by Maltsberger and Buie (1974) was written
from experience with borderline patients. Gabbard and Wilkinson
(1994) provide a comprehensive and clinically valuable guide to this es-
sential topic. So strong is this feature that even the diagnosis itself carries
countertransference weight (as described in Sidebar 1–5). A distinction
can be made between emotional or attitudinal responses to characteris-
tics of borderline patients (e.g., neediness or anger) that may determine
whether a clinician will want to work with them and the emotional or at-
titudinal responses that are evoked as an outgrowth of getting involved
with a patient. The latter are what can greatly affect whether a clinician
will find that involvement personally rewarding and effective.
As is evident throughout this book, no clinical role offers a safe retreat
from potential transference-countertransference enactments with bor-
derline patients. Having said this, the more central one’s responsibilities,
the more intensive the contracts, and the more involving one’s interac-
tional style, the more likely it is that transference-countertransference
problems will arise. Psychopharmacological and cognitive-behavioral in-
terventions enhance early positive transferences by their explicit and
structured efforts to relieve subjective distress. Psychoanalytic therapies
invite more negative transference in that they emphasize the role of in-
terpretive rather than supportive interventions and in that they invite
projections by virtue of their usual lack of structure, neutrality, and en-
couragement of a patient’s self-disclosure.
It is an important aspect of caring for BPD patients that clinicians not

work alone. The need for supervision, consultation, case discussions, and
communication with other members of a team are all safeguards against
countertransference enactments. These interactions with other clinical
professionals also provide a type of supportive relief that transforms such
reactions into understandable, commonplace, and clinically valuable ex-
periences. The advantages of this teamwork are central to the value of
split treatment (Chapter 4) and even to having cotherapists for groups
(Chapter 5 and 9). It is not coincidental that the four psychotherapies
with significant empirical support—DBT, schema-focused therapy, men-
talization-based therapy, and transference-focused psychotherapy (Chap-
ters 11 and 12)—all involve heavy doses of supervision, consultation, and
discussions between treaters.
Summary
This chapter offers an overview of the processes of change and sequence
of treatment modalities involved in the treatment of BPD. It offers cli-
nicians and patients a conceptual infrastructure by which they can orga-
nize treatment plans and by which they can determine whether progress
is occurring—in essence, a structure for deciding whether a treatment
program is well suited to the patient’s changing goals and needs. It estab-
lishes the road map for the rest of the book, which follows the progress of
borderline patients from the issues of being so severely impaired and sui-
cidal that their lives are tenuous to eventually addressing psychological
conflicts about issues such as competition and intimacy that interfere
with their life’s quality.
References
Adler G: The myth of the alliance with borderline patients. Am J Psychi-
atry 136:642–645, 1979
Adler G: The borderline patient in the general hospital. Gen Hosp Psy-
chiatry 3:297–300, 1981
Adler G: Borderline Psychopathology and Its Treatment. New York, Jason
Aronson, 1986
Agrawal HR, Gunderson JG, Holmes BM, et al: Attachment studies with
borderline patients: a review. Harv Rev Psychiatry 12:94–104, 2004
Bateman A, Fonagy P: Psychotherapy for Borderline Personality Disor-
der—Mentalization-Based Treatment. Oxford, England, Oxford
University Press, 2004

Brenner C: Working alliance, therapeutic alliance, and transference, in
The Mind in Conflict. New York, International Universities Press,
1988, pp 137–157
Brightman BK: Peer support and education in the comprehensive care of
patients with borderline personality disorder. Psychiatr Hosp 23:55–
59, 1992
Butler AC, Brown GK, Beck AT, et al: Assessment of dysfunctional beliefs in
borderline personality disorder. Behav Res Ther 40:1231–1240, 2002
Fonagy P, Steele H, Moran G, et al: The capacity for understanding men-
tal states: the reflective self in parent and child and its significance for
security of attachment. Infant Ment Health J 13:200–217, 1991
Friedman HJ: Some problems of inpatient management with borderline
patients. Am J Psychiatry 126:299–304, 1969
Friedman L: The therapeutic alliance. Int J Psychoanal 50:139–153, 1969
Gabbard GO, Wilkinson SM: Management of Countertransference With
Borderline Patients. Washington, DC, American Psychiatric Press, 1994
Gill MM: The analysis of the transference. J Am Psychoanal Assoc
27(suppl):263–288, 1979
Greenson RR: The working alliance and the transference neurosis. Psy-
choanal Q 34:155–181, 1965
Greenspan SI, Sharfstein SS: Efficacy of psychotherapy: asking the right
questions. Arch Gen Psychiatry 38:1213–1219, 1981
Gunderson JG: Defining the therapeutic processes in psychiatric milieus.
Psychiatry 41:327–335, 1978
Gunderson JG, Gabbard GO: Making the case for psychoanalytic thera-
pies in the current psychiatric environment. J Am Psychoanal Assoc
47:679–704, 1999 (discussion 704–740)
Gunderson JG, Lyons-Ruth K: BPD’s Interpersonal Hypersensitivity Phe-
notype: A Gene-Environment-Developmental Model. J Personal Dis-
ord 22(1):22–41, 2008
Gunderson JG, Carpenter WT, Strauss JS: Borderline and schizophrenic
patients: a comparative study. Am J Psychiatry 132:1257–1264, 1975
Gunderson J, Waldinger R, Sabo A, et al: Stages of change in dynamic psy-
chotherapy with borderline patients: clinical and research implica-
tions. J Psychother Pract Res 2:64–72, 1993
Hoffman IZ: Ritual and Spontaneity in the Psychoanalytic Process: A Di-
alectical-Constructivist View. Hillsdale, NJ, Analytic Press, 1998
Hogarty GE, Anderson CM, Reiss DJ, et al: Family psychoeducation, so-
cial skills training, and maintenance chemotherapy in the aftercare
of schizophrenia, I: one-year effects of a controlled study on relapse
and expressed emotion. Arch Gen Psychiatry 43:633–642, 1986
Hogarty GE, Kornblith SJ, Greenwald D, et al: Three-year trials of per-

sonal therapy among schizophrenic patients living with or indepen-
dent of family, I: description of study and effects on relapse rates. Am
J Psychiatry 154:1504–1513, 1997
Jones M: Social Psychiatry: A Study of Therapeutic Communities. Lon-
don, Tavistock, 1952
Jones M: The concept of a therapeutic community. Am J Psychiatry
112:647–650, 1956
Kaysen S: Girl, Interrupted. New York, Random House, 1993
Kopta SM, Howard KI, Lowry JL, et al: Patterns of symptomatic recovery
in psychotherapy. J Consult Clin Psychol 62:1009–1016, 1994
Langs R: The Bipersonal Field. New York, Jason Aronson, 1976
Lanktree CB, Briere J: Outcome of therapy for sexually abused children:
a repeated measures study. Child Abuse Negl 19:1145–1155, 1995
Lewin RA, Schulz CG: Losing and Fusing: Borderline Transitional Object
and Self Relations. New York, Jason Aronson, 1992
Linehan MM: Cognitive-Behavioral Treatment of Borderline Personality
Disorder. New York, Guilford, 1993
Links PS: Psychiatric rehabilitation model for borderline personality dis-
order. Can J Psychiatry 38 (suppl 1):35–38, 1993
Luborsky L: Helping alliances in psychotherapy, in Successful Psycho-
therapy. Edited by Claghorn JL. New York, Brunner/Mazel, 1976, pp
92–116
Main TF: The hospital as a therapeutic institution. Bull Menninger Clin
19:66–70, 1946
Maltsberger JT, Buie DH: Countertransference hate in the treatment of
suicidal patients. Arch Gen Psychiatry 30:625–633, 1974
Masterson J: Treatment of the adolescent with borderline syndrome (a prob-
lem in separation-individuation). Bull Menninger Clin 35:5–18, 1971
Perry JC, Bond M: Empirical studies of psychotherapy for personality dis-
orders, in American Psychiatric Press Review of Psychiatry, Vol 19. Ed-
ited by Oldham JM, Riba MB. Washington, DC, American Psychiatric
Press, 2000, pp 1–31
Ruiz-Sancho A, Smith GW, Gunderson JG: Psychoeducational ap-
proaches for personality disorders, in The Handbook of Personality
Disorders: Theory, Research, and Treatment. Edited by Livesley J.
New York, Guilford, 2001, pp 460–474
Shedler J, Westen D: Refining personality disorder diagnosis: integrating
science and practice. Am J Psychiatry 161:1350–1365, 2004
Singer MT, Larson DG: Borderline personality and the Rorschach test.
Arch Gen Psychiatry 38:693–698, 1981
Skodol AE, Gunderson JG, McGlashan TH, et al: Functional impairment

in patients with schizotypal, borderline, avoidant, or obsessive-com-
pulsive personality disorder. Am J Psychiatry 159:276–283, 2002
Skodol AE, Pagano ME, Bender DS, et al: Stability of functional impair-
ment in patients with schizotypal, borderline, avoidant, or obsessive-
compulsive personality disorder over two years. Psychol Med 35:443–
451, 2005
Sterba R: The fate of the ego in analytic therapy. Int J Psychoanal 15:117–
126, 1934
Waldinger RJ, Gunderson JG: Completed psychotherapies with border-
line patients. Am J Psychother 38:190–202, 1984
Waldinger RJ, Gunderson JG: Effective Psychotherapy With Borderline
Patients: Case Studies. Washington, DC, American Psychiatric Press,
1989
Winnicott DW: The Maturational Process and the Facilitating Environ-
ment: London, Hogarth, 1965
Zanarini MC, Frankenburg FR, DeLuca CJ, et al: The pain of being bor-
derline: dysphoric states specific to borderline personality disorder.
Harv Rev Psychiatry 6:201–207, 1998
Zanarini MC, Frankenburg FR, Reich DB, et al: The subsyndromal phe-
nomenology of borderline personality disorder: a 10-year follow up
study. Am J Psychiatry 164:920–935, 2007
Zetzel ER: Current concepts of transference. Int J Psychoanal 37:369–
376, 1956
Zetzel ER: A developmental approach to the borderline patient. Am J Psy-
chiatry 127:867–871, 1971
Chapter 4
CASE MANAGEMENT
The Primary Clinician
EVEN AS THIS BOOK EMPHASIZES THAT MINIMAL—and certainly

optimal—treatment plans for patients with borderline personality disor-
der (BPD) require complementary modalities, and thus a coordinated
team, it is essential to recognize that one member of the clinical team
needs to assume the primary responsibility for the patient’s care. This
point needs special emphasis because of the natural and universal ten-
dency for members of a team to want to avoid or reduce negative coun-
tertransference reactions or to avoid responsibilities when criticism or
liability can be expected. The person who takes on primary responsibility
essentially is the final decision maker for the numerous questions that
borderline patients pose about the who, what, where, and when regard-
ing therapies. Traditionally, this role has been assumed by someone who
defines himself or herself as the patient’s therapist. However, there are in-
herent problems in trying to administer the selection and implemen-
tation of a therapeutic effort while remaining in the noncontrolling,
exploratory, and empathic stance required of those therapists who prac-
tice psychodynamic psychotherapy.
The clinician who oversees decisions and implements a treatment
plan (including delegation of responsibilities), hereafter referred to as
the primary clinician, must appreciate what is entailed in his or her role,
and this role must be explicitly agreed to by other members of the pa-
tient’s treatment team. For example, the dialectical behavior therapy
(DBT) skills training group leader depends on a primary clinician to re-
inforce the contingencies that therapy requires, the family therapist of-
ten needs a primary clinician to manage the patient’s crises, and the
psychopharmacologist needs a primary clinician to monitor compliance,
89
help assess benefits, and the like. The primary clinician may evolve into a
role that is primarily psychotherapeutic, but initially the required re-
sponsibilities include administrative functions (e.g., monitoring safety,
implementing treatment recommendations) that may in themselves be
therapeutic but will often take precedence over traditional psychothera-
peutic activities (e.g., self-disclosure, insight, affect recognition).
Qualifications
Any mental health clinician who is experienced with borderline patients
and who combines good judgment and a readiness to communicate with
others can fulfill the primary clinician role. Even mental health workers
without professional degrees who have years of experience in inpatient
or residential treatment settings can become very skilled. Nonetheless,
the expectable safety issues, the judgment questions around level of care,
and the potential legal complications of the required decisions mean that
there are definite advantages in having psychiatrists fill this role. Psychi-
atrists generally have more training in making these judgments (and ex-
perience shows that even when their assigned role is modest, they will
probably be included in any legal action to collect damages). The psychi-
atrist’s advantages are outweighed, however, when his or her contacts
with the patient are limited and an experienced and capable clinician
from another discipline, usually a psychologist or social worker, is seeing
the patient more intensely, knows the patient better, or has a stronger al-
liance. Regardless of discipline, no one should undertake the role of pri-
mary clinician without significant experience or, in its absence, without
ongoing supervision by an experienced clinician.
Responsibilities
As outlined in Table 4–1, the responsibilities assumed by a borderline pa-
tient’s primary clinician involve complicated clinical judgments. The
issue of monitoring safety is the most important and is given extended
discussion later in this chapter.
The first task is to establish a contractual alliance (as described in Chap-
ter 3). This is often begun by educating both the patient (Chapters 1 and
3) and his or her family (Chapter 8) about the diagnosis. Regarding rec-
ommended therapies, the contractual alliance is established through dis-
cussion with the patient about what roles will be played and what the
goals of therapy will be. Both DBT (Linehan 1993) and transference-
focused psychotherapy (Kernberg et al. in press) recommend an exten-
sive process in which the patient’s motivation for the treatment is assessed
Case Management ❘ 91
TABLE 4–1. Responsibilities of a primary clinician

Establish a therapeutic framework (a contractual alliance)
Identify needs and develop treatment plan
Level of care (e.g., hospitalization, day or night care)
Modalities (e.g., group, rehabilitation, family, or cognitive-behavioral
therapy)
Comorbidity
Monitor safety
Monitor progress and effectiveness
Use (attendance, involvement)
Benefits (e.g., is patient learning? changing?)
Coordinate therapies
Communication and collaboration
Provide psychoeducation
Between treaters
With those responsible for financing treatment
(and tested) and in which the limits of the clinician’s role are elucidated
(e.g., contingencies for continuation, unavailability except for true emer-
gencies). No doubt the resulting selectivity is important when conduct-
ing research intended to confirm the value of therapies. However, this
selectivity is not usually available to clinicians who are assigned or re-
ferred patients who need treatment and for whom they are assuming re-
sponsibility, including making judgments about the appropriateness of
any type of treatment. For primary clinicians, including those who do not
assume a psychotherapeutic role (e.g., psychopharmacologists; see Chap-
ter 6), the development of an alliance is a mandate, not an option.
The primary clinician combines administrative (i.e., management
and assessment) tasks with alliance-building therapeutic activities (e.g.,
engagement, support, and, when necessary, confrontation). Clinicians
who accept responsibility for the care of borderline patients during or af-
ter a crisis must provide what the patient needs, if possible; these clini-
cians do not have the privilege of saying, “I offer this type of therapy, and
if it isn’t suitable, goodbye.” At the same time, experience and good judg-
ment are necessary to know when a treatment is inappropriate or un-
workable. As described throughout this book, it requires skill to manage
the tasks of being a primary clinician without enacting transference

wishes or fears that cause flight or otherwise undermine the patient’s par-
ticipation in growth-enhancing therapies. Hence to do these tasks well re-
quires a good understanding of the major borderline issues. Of note (as
noted repeatedly in this book) is the important role of using consultants
or supervision or of otherwise avoiding the exclusivity that borderline pa-
tients often crave—at their own risk.
In this book, I summarize most of the knowledge on which primary
clinicians can base their recommendations about therapeutic options.
The most desirable options are often not accessible (e.g., a well-informed
and capable cognitive-behavioral therapist or a halfway house for nonpsy-
chotic patients); the primary clinician then needs to judge whether the
available therapies are potentially effective or are likely to make things
worse. The primary clinician must be aware that any therapy, poorly con-
ducted, is likely to have harmful effects (e.g., increased crises, self-harm,
regressions, rages). Thus, it is unwise to triage a borderline patient to
therapists who are inexperienced and unsupervised or to clinicians who
dislike working with these patients.
It is sometimes difficult to know a priori whether available therapies
can work, and the primary clinician’s role involves ongoing assessment of
effectiveness. Here too the sequences and timetable for expectable
changes described in this book can provide a framework (see Chapter 3
and Table 3–3). Frequently, the primary clinician’s assessment of ineffec-
tiveness becomes very simple—for example, the patient is not attending
groups, the patient keeps abusing his or her medications, the psychother-
apist is clinically depressed, and the psychopharmacologist is inaccessi-
ble. The more serious problems involve implementing the changes that
are consequences of this conclusion.
Liability Issues
Gutheil (1985, 1989) noted that borderline patients are particularly
likely to involve their treaters in liability suits. Without question, this is re-
lated to these patients’ ongoing suicide risks, their tendency to project
malevolence, and the fact that borderline patients are usually—and op-
timally—treated by teams. Psychiatrists usually carry a disproportionate
level of liability risk, but the principles that help diminish such risk are
relevant to all members of a team. Sidebar 4–1 offers some guidelines on
how primary clinicians can conduct their tasks and minimize the dangers
of having liability suits. Sidebar 6–2 addresses the more specific liability is-
sues related to split treatment—when a psychiatrist assumes primarily a psy-
chopharmacologist role.
Sidebar 4–1: Guidelines to Avoid Liability

• Know what usual practices are. If you plan to do anything
innovative (e.g., Amytal [amobarbital] interviews, regres-
sive psychotherapy), obtain consultation, and be sure
that the patient’s consent is informed and documented.
• Do not see a patient more than twice a week without
having significant prior experience or a qualified (expe-
rienced, credentialed) supervisor (see Chapters 10–12).
• Use consultants whenever treatment has an extended
impasse or the patient is getting worse (develops new
behavior problems, becomes more self-injurious).
• When implementing treatment changes against a
patient’s wishes (e.g., discontinuing therapy), seek con-
sultation and document your reasons (see section “Imple-
menting Changes” later in this chapter, and Chapter 5).
• If you are a psychiatrist who is not assuming primary
responsibility for monitoring treatment implementation
or safety monitoring, 1) be sure the responsible others
are credentialed and capable, and 2) explain the
agreement about your role to the patient (see section
“Splits, Splitting, and the Virtues of Split Treatments” later
in this chapter, and Chapter 5).
• In the face of significant risk of suicide or violence, sus-
pend any agreements about confidentiality. (See sec-
tion “Managing Safety” later in this chapter, and Chap-
ters 8 and 10.)
• Do not agree to participate in therapies that you
believe are unworkable (e.g., seeing an alcoholic ado-
lescent who refuses to enter a day hospital for outpa-
tient pharmacotherapy) without first advising the
patient and having his or her significant others accept
that your participation is a time-limited trial to deter-
mine whether treatment is possible.
Relationship Management
Dawson and MacMillan (1993) made a significant contribution to the
treatment wisdom for borderline patients with their book Relationship
Management and the Borderline Patient. Unlike most books that emphasize
ways to interpret or confront borderline patients’ relational problems
with treaters, Dawson and MacMillan move into operational ways to side-
step these problems and have borderline patients be responsibly involved
in their own treatment—or otherwise not be in treatment at all. Central
to their thesis is that the traditional proactive approaches of psychiatrists
and institutions (e.g., prescribing, directing, controlling) expected by—

indeed, welcomed by—most patients are approaches that provide the
materials with which borderline patients destroy their therapies and
make themselves worse. Hence the wise clinician will step back and wait
for borderline patients first to identify what they want, even though the
clinician’s inaction may be protested.
One useful principle of relationship management is that the primary
clinician shifts (i.e., “demedicalizes”) the focus of discourse from diagno-
sis, pills, and suicide risk to social competence—for example, employ-
ment, budgeting, and self-care. As noted in Chapter 3, therapies to
address social competence issues have been slow to develop.
A second principle involves practicing what Dawson and MacMillan call
“no-therapy therapy.” Thus, in response to the borderline patient’s wish for
psychotherapy, a regular time for sessions may readily be offered, but with
the caveat that the therapist is not sure how she or he can be helpful. This
“contract” is well suited to primary clinicians. The patient sets the agenda;
or, as often happens, concerns about the patient voiced to the primary cli-
nician by others become an agenda (e.g., “Joyce missed her last three ap-
pointments”; “Joan stays up until 2:00 A. M. and sleeps all morning”).
Sessions explicitly are not intended to be therapeutic. The atmosphere is
informal; sharing coffee or discussing public events is commonplace. The
primary clinician, despite his or her responsibilities (noted in Table 4–1),
does not pursue these except as they are “forced” on him or her by the pa-
tient or by others in the patient’s life. Even then, the primary clinician ac-
cepts responsibility reluctantly, with explicit statements that he or she will
be likely to make mistakes unless the patient provides direction.
In my experience, Dawson and MacMillan’s approach is generally use-
ful and is very helpful for orienting trainees whose instincts are to “do
things.” However, this approach is more easily implemented within a
health care system in which the patient is assigned a clinician than in a
system in which the patient selects the clinician. In the private-practice
sector, Dawson and MacMillan’s approach—unless buttressed by expla-
nations for the patient and the patient’s significant others—will evoke de-
valuation and a search for a therapist who evokes more hope of the
patient being helped.
Managing Safety
Assessing Suicidality
Because of apprehensions about the legal, administrative, and psycholog-
ical consequences should a suicide occur, mental health professionals
feel highly anxious about distinguishing true suicidal intentions from
self-harming behavior without lethal intent. In the absence of such a dis-

tinction, mental health professionals are likely to assume the worst; this
covers their liability and offers the bonus of allowing them to feel that
they are fulfilling one of the most dramatic and perhaps alluring roles of
a caregiver: saving a life.
The statistics documenting the high rates of suicide (about 9%; see
Chapter 1, section “The Behavioral Specialty: Self-Injurious Behavior”)
can be used to vindicate clinicians who attempt to prevent borderline pa-
tients from performing suicidal acts. For primary clinicians, this can
mean involuntary hospitalization of patients, but more often it entails de-
cisions such as giving prescriptions for only small quantities of medica-
tions, enlisting family members to help monitor patients’ suicidality, and
making oneself available for crises. From another perspective, the statis-
tics documenting the low percentage of suicidal acts that are serious at-
tempts vindicate clinicians who are primarily concerned about the
secondary gain and manipulative intentions related to borderline pa-
tients’ self-destructive acts. Interventions by these clinicians are typically
directed toward diminishing the secondary gains from self-destructive
acts by, for example, staying uninvolved with hospitalizations or being un-
available between sessions. On balance, the statistics about borderline pa-
tients’ suicidality offer little comfort.
Of only modest additional help are a series of studies examining pre-
dictors of suicidality in borderline patients. These studies provide evi-
dence that patients with BPD are at increased risk for suicidal behavior
and, perhaps, suicide if the following are present (Kolla et al. 2008; Links
and Kolla 2005):
• Worsening of a major depressive episode

• Worsening of substance use disorders
• Recent (within several weeks) discharge from a psychiatric hospital
• Recent negative life events such as loss of immediate family support
and legal troubles
• Presenting in a highly regressed, uncommunicative state
The painful truth is that a clinician working with borderline patients

must make thoughtful judgments about the patients’ suicidality with con-
sideration of their motives and intentions; access to lethal means; the
complexity of the patients’ relationship to significant others, including
the clinician; and the past responses from those others, including the cli-
nician. From a medicolegal point of view, clinicians making these judg-
ments should document their assessment and considerations regarding
suicide risk.
An acute-on-chronic model of suicide risk assessment can assist clini-

cians in making judgments about a patient’s risk for suicide and can pro-
vide a clear method to communicate about these judgments (Links and
Kolla 2005). Figure 4–1 shows a way of assessing and communicating the
suicidal risk. Patients with BPD and a history of repeated suicide attempts
typically have a chronically elevated risk for suicide much higher than that
of the general population. This chronic risk can remain elevated above
that of the general population for prolonged periods, even for years. The
patient’s level of chronic risk can be estimated by taking a careful history
of the previous suicidal behavior and focusing on those events when the
patient made attempts with the greatest intent and medical lethality. By
documenting the patient’s most serious suicide attempt, the clinician can
estimate the severity of the patient’s ongoing chronic risk for suicide.
In patients with BPD, the acute-on-chronic level of risk is related to
several factors, as shown in Figure 4–1. The borderline patient who is at-
tempting to evoke a response from his or her environment is at less risk
than the borderline patient who presents in a highly regressed, uncom-
municative or dissociative state. In these states, interventions have to be
implemented quickly to reduce the risk of suicide attempts or self-harm.
In addition, patients with BPD are known to be at risk for suicide around
times of discharge. The clinical scenario of a patient presenting in crisis
shortly after discharge from an inpatient setting illustrates a time when
the risk assessment must be very carefully completed and documented to
ensure that a proper disposition is made.
Using the acute-on-chronic model can be very effective for communi-
cating decisions regarding interventions.
Vignette
Ms. B, a 27-year-old woman with BPD, had been pleading for admission.
On learning that discharge was recommended by the consulting psychia-
trist, the on-duty emergency physician was surprised. He asked the psychi-
atrist why he was discharging the patient because he thought for sure that
she would be admitted. The consultant replied, “I know Ms. B pretty
well.... let me see if I can explain this” (he drew a diagram similar to Fig-
ure 4–1). “Ms. B is clearly a risk for suicide and much above the risk of
someone in the general population. But this risk is chronic. I have seen
her here in the emergency department several times over the last year,
and the chronic level of risk seems unchanged. I don’t see any point in
placing her in the hospital. A short stay in the hospital won’t change her
chronic level of risk. But I did try to motivate her to do something about
her drinking. I think a lot of her suicidal feelings are related to her drink-
ing and told her so. She wasn’t ready to hear this from me, but she did say
she would speak with her regular physician tomorrow. I will let him know
she was here.”
FIGURE 4–1. Acute-on-chronic suicide risk in patients with bor-

derline personality disorder (BPD).
In patients with BPD, the acute-on-chronic level of risk (curved arrow) can
change more quickly and will be modified by several factors: the black arrow in-
dicates evidence-based factors causing an acute exacerbation of risk; the white ar-
row indicates factors that might reduce an acute exacerbation.
If a patient is thought to be at a chronic but not acute-on-chronic risk

for suicide, a clinician should communicate that a short-term hospital ad-
mission will have little or no effect on a chronic risk. However, if the pa-
tient has an acute-on-chronic level of risk, then several things could be
considered. As shown in Figure 4–1, the patient should be encouraged to
use crisis management skills (see Chapter 9 on skills training), a short
trial of low-dose antipsychotic medication might reduce the acute symp-
toms (see Chapter 7 on pharmacotherapy), an inpatient admission might
well be indicated, or the patient might be able to state his or her needs
and problem-solve for alternative solutions (see section “Giving, Receiv-
ing, and Participating in Supervision” later in this chapter). A short-term
admission usually will cause the level of risk to return to chronic pread-
mission levels. Recommendations regarding the chronic level of suicide
risk in patients with BPD should involve referral to the various therapies
described elsewhere in this book.
A Preventive Stance
Primary clinicians should early and often advise borderline patients that
the clinicians view suicidal acts as dangerous distractions from the pa-
tients’ work—that of attaining a better life. To make this message mean-

ingful, it is essential that—while never ignoring any hints of suicidality—
primary clinicians do not proactively look for evidence of it. This ap-
proach contrasts with the approach used in DBT, in which the therapists
systematically inquire about self-destructiveness at the start of each session
(see Chapter 11). However, recurrence of suicidal thoughts and impulses
should be assumed and specifically predicted whenever the borderline pa-
tient is about to lose some source of support. Anticipating such recur-
rences is part of an important process during the first year of treatment
(phase 2 of psychotherapy; see Chapter 12). It gives an essential meaning
to these ideas, impulses, or behaviors—as responses to feeling insuffi-
ciently cared for. Interpretation of this meaning can then allow discussion
of alternative, more adaptive responses (much as DBT therapists do).
Two embellishments of this simple stance can often be useful. The first
is to suggest that suicidality is motivated by anger, not depression. The sec-
ond is to link patients’ history of counterdependence to family dynamics,
wherein wanting to be cared for was unacceptable—thus, extreme behav-
iors and illness became the vehicle for getting needs met. When the cli-
nician is proactive with such interpretations—often presented in an
educational way—borderline patients will hear that their anger or their
wishes for care are understandable feelings that can be talked about.
Responding to “Feeling Unsafe”

In conjunction with increased concern in the current mental health com-
munity about suicide prevention, patients have increasingly been advised
to talk about suicidal ideas or impulses—rather than acting on them. In
this context, patients will now report that they “feel unsafe.” The following
vignette illustrates those issues that require a primary clinician to manage:
Vignette
Ms. C, a 28-year-old woman, began treatment while hospitalized for an
overdose. Her primary aftercare clinician, also her therapist, had seen her
once weekly while she attended an interpersonal group and attempted to
reenter graduate training.
Two weeks after discharge from the hospital, she called the primary
clinician at 11:00 P.M. on a weeknight:
Ms. C (in a weak voice): I’m sorry to call, but I’ve been feeling strange,
unsafe, out of control.
Therapist (waits, then asks): When did this start?
Ms. C: I don’t know.... I’ve been getting worse for awhile.
Therapist (no response).
Ms. C: You’re not saying much.
Therapist: I’d like to help, but I’m unclear about what I can do. Did
you have some ideas?
Ms. C: No.
Therapist: Hmm.
Ms. C: What does that mean?!
Therapist (no response).
Ms. C: I guess you can’t help me.
Therapist: That’s what I was fearing too.
Ms. C: Then what should I do?
Therapist: I do hope you will take good care of yourself and make use
of the emergency services if necessary.
Of note in this vignette are the clinician’s modest expressions of cu-

riosity or alarm but equally modest, albeit reassuring, expressions of con-
cern. The clinician seems aware that the patient could be in danger of
self-harm but does not ask for reassurance that she will not act on such
impulses. The clinician appears confident that, should the patient need
more active help, she knows how to obtain it. This assumption contrasts
with the practice of contracting for safety (Sidebar 4–2). Not noted in this vi-
gnette is that the therapist had trouble getting to sleep after the call. It is
notable that this fact (trouble sleeping) was told to the patient in the fol-
lowing session. This example illustrates some of the general principles re-
lated to monitoring safety in borderline outpatients shown in Table 4–2.
Sidebar 4–2: Is Contracting for Safety1 Safe?

Although borderline patients are capable of distorting real-
ity, many borderline patients—probably most—would find it
abhorrent to knowingly lie, especially to anyone who ap-
pears caring. Thus has evolved the practice of contracting
for safety, in which clinicians ask patients for reassurance
that they will not harm themselves, formally inviting them to
be explicit about this, sometimes even writing it out. This pro-
cedure encourages patients to remain safe to uphold the
honor of their promise once given; it is also an exercise that
affirms the clinician’s concern and acts as a deterrent to
impulsive action. Hence its popularity.
What’s the downside? First, special conditions are re-
quired before contracting for safety is appropriate. The
contracting borderline patients must have a value system
that values honesty or keeping their word. Alternatively,
even patients who lack these values may perceive their cli-
nician as so “good” (e.g., not motivated by interest in as-
1
Contracting for safety is very different from contracting for therapy,
although the latter may include expecting patients to take care of safety
issues (for example, see Kernberg et al. 1989 or Plakun 1994).
TABLE 4–2. Guidelines for management of safety

During a crisis
1. Express concern after the patient alerts you to suicidal or other safety
issues.
2. Allow patients to ventilate. It will relieve tensions around suicidality.
3. Avoid taking actions to prevent potential suicidal behaviors when
possible:
a. Ask patients to be explicit about wanting help.
b. Ask patients to be explicit about what help they hope you can
offer.
c. Assume, unless told otherwise, that the patient can use
community-based emergency services.
After the crisis
1. Follow up by discussing all safety issues, including their effect on you,
within the context of scheduled appointments.
2. Actively interpret the nonspecific reasons that can and did provide
relief—for example, the perception of being cared for.
3. Identify the infeasibility of depending on your being constantly
available; work on problem solving about available alternatives.
4. Actively address the patient’s anger toward you whenever it becomes
apparent.
serting control) that they would not want to betray the

clinician’s trust. Another condition that allows contracting
to help is that contracting patients must be sufficiently re-
flective to control self-destructive impulses. When these
conditions are met, the contract will help secure a holding
environment: the patients will be safer.
But even when these conditions exist, there can be a
more subtle downside to contracting for safety. Contract-
ing will alter (I think damage) the process of establishing a
psychotherapeutic working alliance (see Chapter 3). Alli-
ances for treatment—the contractual and relational
forms—are created to keep an illness at bay, and these
can rest on a businesslike, explicit verbal agreement. A
working therapeutic alliance depends on a patient’s com-
mitment to change and growth. Contracting for safety im-
plicitly moves the act of preventing self-destructiveness into
the therapy: the therapist is actively trying to become a
barrier to self-destructive behaviors. By this activity, the ther-
apist enacts a borderline patient’s powerful transference
wish for the therapist to be the patient’s protector and
caregiver. The patient’s internal experience is that he or she

is safe or alive because of his or her therapist—rather than
because he or she believes that he or she can change.
Such a transference, once established, can be hard to
undo and is easily disrupted by the inevitable vicissitudes of
the relationship. Thus, the patient may agree to contract for
safety as a way to please the idealized therapist; but the
contract itself, once made, reifies the therapist’s transfer-
ence role as an omnipotent protector. Thus, because the
clinician is ready to have the patient’s life rely on the con-
tract, the patient’s idealized transference can become in-
tensified. Because of this intensified transference, the
patient’s inevitable subsequent disillusionments with the
therapist may be all the more dangerously life-threatening.
Is contracting for safety safe? It depends.
Of course, the type of clinician response shown in the last vignette

does not always get the same type of response from the patient. In that vi-
gnette, the patient fortunately seemed somewhat reflective, wanting a dis-
cussion. Other principles identified in Table 4–2 are illustrated by a
different version of that vignette. In the version following, the therapist’s
responses evoke an angrier, more threatening response:
Vignette
Patient: I’m sorry to call, but I’ve been feeling strange, unsafe, out of
control.
Therapist (waits, then asks): When did this start?
Patient: I don’t know. ... (Irritably): How the hell can I think about
that?! Didn’t you hear me say that I feel unsafe, out of control?! I’m stand-
ing here with a bottle of pills.
Therapist (waits, then): You need me to respond to the fact that
you’re at risk?
Patient: Yes!
Therapist: This is a crisis?
Patient: Yes!
Therapist: How can I help?
Patient: I don’t know! You’re the fucking doctor; you should know.
Therapist: I wish I did.
Patient: Are you telling me you don’t know how you can help me?!
Therapist: (no response).
Patient: Is that what you’re saying?!
Therapist: I hope you won’t hurt yourself.
Patient (starts weeping): Oh God, I just don’t know what to do. I feel
so awful.
Therapist: I know you do. I can hear that.
Patient: It all started yesterday.. . . (begins to narrate a detailed re-
counting of the intervening events).
In this instance, the patient’s desperation and anger move the thera-
pist to address more directly the patient’s at-risk behaviors. Even here,
however, he enlists the patient in articulating that she is in a crisis and in
identifying how she wants him to respond. As is usually the case, she as a
borderline patient has trouble saying what is wanted. Unspoken is that
she wants concerned attention. Also unspoken is what the primary clini-
cian should know in this situation: that if the patient is given license to
ventilate, the immediate danger of self-harm will dissipate. After the ex-
change in this vignette, the therapist slept comfortably.
In the session following the exchange above, the therapist insisted,
against the patient’s protests, on discussing what had transpired. He iden-
tified how “surprisingly angry” the patient had become when he did not
immediately express concern for her safety and when he said that he
didn’t know what she wanted him to do to help.
Patient (irritably): I’m sorry for getting so angry, but it was a crisis.
I guess you’ve never been through what I go through.
Therapist (sidestepping her anger): When you began to talk about
what was bothering you, that seemed to have helped.
Patient: Yes, it really did. I appreciated that you listened.
Therapist: That was most interesting to me. What seemed to help was
just having someone listen. I didn’t do anything. How can that be?
The patient then discussed how rare it had been to have someone lis-
ten to her. The therapist used this exchange to educate the patient about
theory, suggesting that some people (with BPD) get overwhelmingly pan-
icked when they do not have someone available to offer comfort and that
such people find aloneness intolerable. That led to a discussion of the pa-
tient’s living situation and alternative sources of comfort. He added that
although he was glad to have proved useful, it was dangerous for her
safety to depend on his availability. Moreover, providing comfort was not
a function he could serve too often without disrupting his own life. (He
thereby actively drew attention to his own limits, as opposed to setting
limits on the patient—as detailed later in this chapter, section “Bound-
aries, Violations, and Setting Limits.”)
Responding to Recurrent Suicidality:

The Principle of False Submission
Encounters with borderline patients who voice active suicidal intentions
present other problems. Frequently, the borderline patients who voice
such intentions have histories of chronic suicidality and multiple at-
tempts. For clinicians, this history makes it difficult to judge the serious-
ness of the intentions and creates moral and ethical dilemmas (Fine and
Sansone 1990; Frances and Miller 1989). The clinician usually feels that
questioning the seriousness of the patient’s suicidal intentions could
magnify the likelihood and lethality of an attempt. Beyond this, the clini-
cian will know that hospitalizations—the usual response to suicidality—
can rarely address the underlying causes of the suicidality and might in
fact perpetuate the borderline patient’s allegations of suicidality (as a re-
sult of the secondary gains of being rescued, getting attention, and avoid-
ing the problems of living in the community).
Vignette
Ms. D, a 35-year-old, disheveled, agitated, overweight, single woman, ap-
peared for her first clinic appointment. She promptly stated that she was
grateful to “now have a therapist,” and that she had needed one for 3
years. The evaluating clinician felt uneasy about the role of “therapist”
that he had been assigned by the patient, but before he could address this
the patient went on to say that she felt very suicidal. In response to the cli-
nician’s inquiries, she reported that she had been suicidal “off and on for
many years” and had already had 31 hospitalizations.
Clinician: What has caused you to become suicidal now?
Ms. D: I don’t know; what difference does it make? (now becoming ir-
ritated and defensive)
Clinician: Has anything happened in your life recently? (Clinician is
skeptical about the patient’s lethality and hoping to isolate specific events
that can be addressed but already is feeling highly anxious about the pa-
tient’s volatility and potential flight.)
Ms. D: All I know is that I visited my parents and became very upset
and had to leave. No, I don’t know why. No, they didn’t say anything. Yes,
it’s happened before, and last time I nearly killed myself.
Clinician: What happened?
Ms. D: I drank a quart of vodka and then took any fucking pills I could
find.. .. I would have been dead if my landlord hadn’t noticed that the
television was on all night.
Clinician (now convinced that the patient is dangerous, but still feeling
coerced into suggesting hospitalization): Are you feeling that way again?
Ms. D: I just want to get control of myself. If I can’t, I’m going to slash
my neck. This time I don’t want to fail.
Clinician: Would you like to go into the hospital?
Ms. D: I need to.
This vignette illustrates a common situation, a relatively unknown pa-

tient presenting with agitation and suicidal ideation, and a usual interven-
tion, the patient getting hospitalized. This hospitalization proceeds despite
the clinician’s doubt about the seriousness of the intention and despite an
expectation that another hospitalization (in Ms. D’s case, the 32nd!) is un-
likely to help and may even be reinforcing a self-defeating pattern. The cli-
nician will usually feel coerced, manipulated, and helpless. Still, in the
absence of alternatives that can surely safeguard the patient, he is offering
her the safest and most expedient response by suggesting hospitalization.
In a thoughtful disquisition on this borderline-specific dilemma,
Behnke and Saks (1998) argued that an extended informed-consent pro-
cess (using contracting in DBT as an example in which the patient com-
mits to treatment goals) can redirect such patients’ intentions. This is true
for patients soliciting treatment, a situation in which clinicians have the
choice of saying that they cannot help an unmotivated patient. But the di-
lemma stated in the previous paragraph was felt very acutely by Ms. D’s cli-
nician, who did not have the choice of turning down the patient’s request
for help. The problem for the field of therapeutic jurisprudence is whether
the law can protect clinicians who keep such patients out of hospitals—
basing their decision on the patient’s welfare and acknowledging what
Maltsberger (1994) referred to as a “calculated risk” of death—rather than
hospitalizing the patient because it protects the clinician’s welfare. Having
identified the problem, Behnke and Saks could offer no remedy.
My own approach to this situation starts by making the dilemma ex-
plicit. I tell patients such as Ms. D that hospitalization would be the safest
option but that it is not likely to be helpful and probably would be harm-
ful to her longer-term welfare. I explain that hospitalization involves in-
viting others to assume control of the patient’s life and that this can
discourage learning self-control. Moreover, I say that for many patients
such “rescues” become a way of feeling cared for and that being hospital-
ized feels like being adopted, although that is not actually what hospital-
izations mean. I tell these patients, “To me, offering hospitalization to you
primarily represents a way to avoid my being legally liable should you
commit suicide. I actually believe the more caring response would be to
try to keep you out of a hospital despite the potential risk to me.” I then
tell them that in my judgment the best way to proceed would be to take
the time needed to see why they are recurrently suicidal and to develop a
treatment plan that addresses those reasons. Patients are often unsur-
prised by such statements, and a different negotiation then occurs, as
seen in the following vignette:
Vignette
Patient: Are you saying that you really think it’s a mistake to go into
the hospital?
Clinician: Not if you’d otherwise kill yourself, but if you stay alive
you’d be better off without it.
Patient: Are you saying you won’t put me in a hospital?
Clinician: No, of course not. It would be “suicidal” for me to try to pre-
vent a potentially suicidal patient like you [note that therapist does not
question her suicidal potential] from entering a hospital if you want to

[note that therapist moves the patient’s impulse for action into the arena
of the patient’s wanting, giving the patient agency for whatever happens].
I just don’t believe it will be good for you. If you were to make a suicide at-
tempt after leaving here, it could be difficult for me personally and pro-
fessionally; potentially, I could even be sued. So if you tell me that you
intend to kill yourself, that’s very powerful. Then in you have to go. But, if
you go, don’t go thinking that I’ve done what I think is the right thing for
you—or that it’s because I care for you. It doesn’t mean either of these
things. I would think that you are just hoping for an adoption.
This illustrates the principle of false submission: by ostensibly giving the

patient what he or she wants but disarming it of its hoped-for meaning,
the cycle of repeated admissions can be broken. This change will not usu-
ally happen the first time: the patient will almost always go into the hospi-
tal after first having this exchange. But the action now has a different
meaning: the patient is going because he or she wants to go, not because
the doctor said so. When this stance is followed up and reinforced by oth-
ers on the patient’s treatment team, it diminishes the treaters’ sense of be-
ing manipulated or coerced by the patient and breaks down the patient’s
fantasies of rescue or love. The therapist “gives in” but robs the patient of
much of the expected satisfaction. It is particularly important that the staff
on the inpatient unit be aware of and feel comfortable “being used” this
way. If they are unaware, they may offer unnecessary secondary gains. If
they are aware but angry, they are likely to provoke a hostile control strug-
gle that unnecessarily extends the duration of the hospitalization.
In Ms. D’s case, I would want to involve the patient’s family and her
previous treaters in the decision about being hospitalized. Such involve-
ments take time and may for practical reasons prove infeasible, but the
principle behind advocating this process is to underscore the clinician’s
wish to do the right thing, and this involvement will encourage the pa-
tient to consider alternatives. With my patients, ultimately I declare that
I agree to their going into the hospital only because they “insist” on it
(now changing “want” to “insist” to underscore their agency, but still care-
fully avoiding suggesting that it is because they are in my view suicidal).
This helps move the discourse from medical necessity into the patient’s
agency. In essence, this approach extends the principles described earlier
by Dawson and MacMillan (1993).
Implementing Changes
When implementing changes in treatment of BPD, the primary clinician
must proceed with sensitivity and caution. It is easy to unwittingly evoke a
response in which the borderline patient desperately or defiantly clings
more tightly to the ineffective therapy. The mechanism for this angry re-
sistance to a proposed change often involves evoking a split, whereby the
primary clinician is seen as cruelly depriving the patient rather than try-
ing to help. Therefore, the way in which the need for change is commu-
nicated to a patient is very important. Autocratic announcements usually
will evoke resistance, but even when accepted, they can be harmful be-
cause they do not improve the patient’s self-awareness about his or her
needs or about the ways in which these needs can be communicated. Cer-
tainly, recommendations for change—especially if they involve changes
to less-intensive services—should be accompanied by empathic anticipa-
tion that the changes will be difficult. Giving “you can do it” assurances
causes borderline patients to feel that the therapist is minimizing their
difficulties. It is also of critical importance that the primary clinician ini-
tiate communications with the collaborating member(s) of the therapy
team, and with the patient, to ensure that everyone is aware of and in-
volved in all treatment planning. Most clinicians who like working with
borderline patients learn to do this quite comfortably; clinicians who are
hesitant about addressing problems usually avoid this sector of psychiatry.
Boundaries, Violations, and Setting Limits

Boundaries refer to the agreed-on differentiation of the patients’ and the
clinicians’ or therapists’ roles. Both patients and clinicians are capable of
boundary transgressions—that is, stepping out of their roles. When pa-
tients do this—expressing their wish for a therapist to depart from his or
her role—it reveals issues to be clarified and explored. When therapists
do this behaviorally (e.g., having lunch with the patient or giving stock
tips), it is called a boundary violation. Without doubt, transgression by ei-
ther patient or clinician is likely to encourage the transgression of the
other, but as Gabbard (2004) pointed out, patients do not commit “vio-
lations”; their transgressions require examination and may reflect fail-
ures on the part of the therapist.
Although many of the pioneering psychoanalytic leaders practiced
what would now be considered boundary violations (Gutheil and Gab-
bard 1993), the acceptable norms of practice are now better established.
Still, clinicians are often warned that borderline patients do not respect
the boundaries of a professional relationship and that, as a result, clini-
cians need to be prepared to set limits. In fact, therapists who are working
with borderline patients are most likely to violate acceptable norms
(Gutheil 1989). Because professionals’ consciousness about boundaries
is so tinged by concerns about professional ethics and liability, highlight-
ing this issue runs the risk of increasing therapists’ anxieties. As a result,
therapists may adhere too rigidly to “professional” rules and become un-
duly hesitant to become deeply involved with borderline patients. This
anxiety may be manifest in disproportionate impatience about a minor
transgression such as lateness or an intersession contact. On the contrary,
increased consciousness about this topic may underscore the importance
of extensive supervision, use of consultants, and attention to counter-
transference.
Colson et al. (1985) noted that the psychotherapies with the most
negative outcomes in the Menninger Psychotherapy Research Project
were those in which therapists were content to interpret acting-out be-
haviors without setting limits. Table 4–3 identifies a sequence of re-
sponses that usually sidestep the necessity of setting limits. Limits are
sometimes valuable, but usually they reflect impatience or fearfulness on
the part of therapists who are uninformed about or do not trust the pro-
cess described in this table. It can be very difficult to insist that patients
talk about the meaning behind their undesirable behaviors, but this dis-
cussion is essential for patients to understand and respect the limits on a
therapist’s availability, support, or knowledge. Such discussions provide
the cornerstone for resolution and prevention of boundary transgres-
sions.
Central to the process described in Table 4–3 is that the clinician rec-
ognize his or her own limits. These limits should be compatible with com-
passion and with accepting a responsible role in monitoring patient
safety. But, having said this, it must be added that limits also should be
compatible with the clinician’s personal and professional welfare. When
they are at risk, the limits should be introduced as originating in oneself
(steps 4 and 5 in the table). Being clear that it is the clinician’s limitation,
while remaining empathic about the patient’s wishes, rather than hostile,
is almost always accepted by borderline patients.
As a further note, clinicians must set only limits that he or she can re-
inforce (e.g., it would be fruitless to set a limit on when a patient goes to
bed) and ensure that contingencies are proportional to the problem
(e.g., it would be unfair to view lateness as incompatible with a therapy’s
goals). Diminishing the number or duration of appointments, requiring
discussion with the patient’s family, or obtaining a consultation should
precede the more extreme limit of terminating the treatment.
Splits, Splitting, and the Virtues of Split Treatments

Splitting is a term used to describe both an interpersonal and an intrapsy-
chic phenomenon. Within psychoanalytic terminology, splitting refers to
a defensive process identified by Melanie Klein (1946) as originating early
TABLE 4–3. Clinician responses to boundary transgressions

(possible phrasings of responses appear in
parentheses)
1. Identify it as a problem after it occurs. (“Let’s talk about…”)
2. Investigate what the patient wants. Don’t assume the behavior was
based on needs. (“I was unclear how you hoped I could help.”)
3. Validate how that wish is understandable. (“Yes, I could see how that
would help you.”)
4. Discuss how the behavior can be harmful to the therapy or the
therapist. (“I found it troubling because…”)
5. Apologize for one’s limitations. (“I wish I [could give, be] what you
want.”)
6. If the behavior recurs under circumstances when points 4 and 5
have been discussed—that is, when the therapist’s disapproval has
been established—inquire why. (“Did you misunderstand? Are you
angry? Did you want me to disapprove?”)
7. Note that whatever motivated the recurrence should be discussed,
not enacted, for therapy to benefit the patient.
8. If it still recurs, set a limit—preferably when the limit setting can be
processed, not when either you or the patient is angry.
in life that allows a child to ignore or dissociate (split off) negative hostile
perceptions of his or her needed other, thereby preserving a “good,” al-
beit distorted, representation (a part object) of that other. Within the larger
mental health community, this defense became identifiable by the border-
line patient’s tendency to perceive others in dichotomous, “all-good” or
“all-bad” terms and then to treat others very differently (idealized or de-
valued, respectively), depending on which side of the internal split they
occupied. Because of this tendency to split, prior generations of clinicians
have been warned to beware of splitting lest they develop antagonistic
views toward the member(s) of a treatment team who are on the opposite
side of the patients’ split or lest they otherwise get involved in counter-
transference enactments (Gabbard 1989, 1994).
As described elsewhere (Gunderson 1984), the splitting between ob-
jects is not simply a product of the borderline patients’ splits—that is,
their projections—but is predictably based on whether the other is in fact
frustrating or supportive. In this way, the “projections” are well suited to
the recipients (i.e., are based on real characteristics of the objects). As
such, the splitting reflects an interpersonal as opposed to a purely intra-
psychic process. Moreover, such splits otherwise serve desirable functions

within treatment teams by helping borderline patients grow aware and
tolerant of both frustrations and support (“kicks and kisses”). The critical
issue is that members of a treatment team recognize that whether they
are providing frustration or support (or being seen as providing them), the
other component is necessary and desirable for others to provide. Unfor-
tunately, when clinicians are aware that validating borderline patients’
projections of badness (i.e., agreeing with their devalued view of another
treater) can lead to splits, this awareness can lead members of teams to
bond together by invalidating the borderline patient’s attributions (i.e.,
being protective about the other treater’s goodness). Such responses ne-
gate the partial reality of the borderline patients’ perceptions. Moreover,
the idea that, to prevent splits, staff members need to protect one an-
other against negativity confers too much power on the patients’ hostility.
This, too, is harmful.
The principle of split treatment is that despite the dangers of splitting,
treatment plans for borderline patients should routinely involve at least
two treaters, two modalities, or any two components. When coordinated,
two components in a treatment can provide a container for the splits and
projections that keep the borderline patient in treatment. To be specific,
split treatment means that patients receive two different and, in some re-
spects, independent services. Whatever the components (e.g., hospital
and psychotherapist, psychopharmacologist and family therapist, or pri-
mary clinician and self-assessment group), the governing principle is that
having two relatively independent but complementary therapies allows
the inevitable frustrations with any particular treatment to be contained
without necessitating the patient’s flight. Borderline patients’ inability to
experience frustrations without assigning malevolence and taking angry
or fearful flight is the reason that they so frequently drop out of therapies
(Waldinger 1987). This tendency is why selecting appropriate psycho-
therapeutic techniques relates to borderline patients’ level of care. When
borderline patients have a second component to discuss their frustra-
tions with, they retain a “good object” who will urge them to voice their
complaints to the frustrating therapist (e.g., psychopharmacologist,
group leader) rather than leave.
The usual setting in which the advantages of split treatment are met is
by the combination of a psychopharmacologist and a psychotherapist
(usually a psychologist; see Chapter 6), with one or the other serving as
the primary clinician. The containing function of split treatment is,
I believe, one reason that DBT had such a low dropout rate (16%; see
Chapter 10). Linehan (1993) nicely operationalized the response that cli-
nicians or therapists should make when confronted about the alleged
failures, cruelties, and so forth of the other “bad” therapist. The “good”
therapist should neither agree with the patient nor defend the other—simply en-
courage the borderline patient to express complaints directly to the ob-
ject of the complaints. Split treatments are advantageous to borderline
patients if provided by knowledgeable and mutually respectful clinicians.
If not, split treatments may be harmful and may increase liability risks
(see Sidebars 4–1 and 6–2).
Giving, Receiving, and Participating in Supervision

In the earlier description of principles for structuring treatment with BPD
patients, careful attention to countertransference feelings and setting a
low threshold for seeking consultation or supervision was noted. Supervi-
sion should support the therapist, give another perspective to problem-
solve difficult clinical dilemmas, bolster a theoretical understanding to
comprehend the patient’s current issues, and assist the therapist to main-
tain a benevolent, caring, and curious attitude to the patient’s vicissitudes.
Effective supervising related to therapy with BPD patients should provide
a safe place for therapists to disclose feelings and attitudes that wrong
them. Therapists are unwise to undertake ongoing psychotherapy with
BPD patients if good supervision, or at least consultation, is not readily
available from someone experienced with these patients.
Summary
To clarify and simplify the process of clinical decision making, someone
needs to be clearly identifiable as a borderline patient’s primary clinician,
sometimes referred to as case administrator. The person in this role also
may fill other roles, but insofar as the borderline patient still requires lim-
its, safety interventions, or unwanted confrontations, it is difficult for the
primary clinician also to serve as a patient’s dynamic psychotherapist (at
least within the transference-focused psychotherapy model, described in
Chapter 12) until the patient progresses into a second phase of treatment
(see Chapters 3 and 12). The role of primary clinician is compatible with
being a patient’s family group therapist, cognitive-behavioral therapist,
or psychopharmacologist. Central to a primary clinician’s tasks is the abil-
ity to communicate with the patient’s significant others (both family and
treaters) and to make good clinical judgments about whether a patient is
progressing or is safe and, if not, to implement solutions effectively. A
stance about safety issues that involves much inquiry and minimal action
is suggested.
References
Behnke SH, Saks ER: Therapeutic jurisprudences: informed consent as
a clinical indication for the chronically suicidal patient with border-
line personality disorder. Loyola Law Review 31:945–982, 1998
Colson DB, Lewis L, Horwitz L: Negative effects in psychotherapy and
psychoanalysis, in Negative Outcome in Psychotherapy and What to
Do About It. Edited by Mays DT, Franks CM. New York, Springer,
1985, pp 59–75
Dawson D, MacMillan HL: Relationship Management and the Border-
line Patient. New York, Brunner/Mazel, 1993
Fine MA, Sansone RA: Dilemmas in the management of suicidal behav-
ior in individuals with borderline personality disorder. Am J Psy-
chother 44:160–171, 1990
Frances AJ, Miller LJ: Coordinating inpatient and outpatient treatment
for a chronically suicidal woman. Hosp Community Psychiatry 40:
468–470, 1989
Gabbard GO: Splitting in hospital treatment. Am J Psychiatry 146:444–
451, 1989
Gabbard GO: Treatment of borderline patients in a multiple-treater set-
ting. Psychiatr Clin North Am 17:839–850, 1994
Gabbard GO: Long-Term Psychodynamic Psychotherapy: A Basic Text.
Washington, DC, American Psychiatric Publishing, 2004
Gutheil TG: Medicolegal pitfalls in the treatment of borderline patients.
Am J Psychiatry 142:9–14, 1985
Gutheil TG: Borderline personality disorder, boundary violations, and
patient-therapist sex: medicolegal pitfalls. Am J Psychiatry 146:597–
602, 1989
Gutheil TG, Gabbard GO: The concept of boundaries in clinical prac-
tice: theoretical and risk-management dimensions. Am J Psychiatry
150:188–196, 1993
Kernberg O, Selzer M, Koenigsberg HW, et al: Psychodynamic Psycho-
therapy of Borderline Patients. New York, Basic Books, 1989
Kernberg O, Yeomans F, Clarkin JF: Transference focused psychother-
apy: overview and update. Int J Psychoanal (in press)
Klein M: Notes on some schizoid mechanisms. Int J Psychoanal 27:99–
110, 1946
Kolla NJ, Eisenberg H, Links PS: Epidemiology, risk factors, and psycho-
pharmacological management of suicidal behavior in borderline
personality disorder. Arch Suicide Res 12:1–19, 2008
Links PS, Kolla N: Assessing and managing suicide risk, in The American
Psychiatric Publishing Textbook of Personality Disorders. Edited by
Oldham J, Skodol AE, Bender DS. Washington, DC, American Psy-
chiatric Publishing, 2005, pp 449–462
Maltsberger JT: Calculated risks in the treatment of intractably suicidal
patients. Psychiatry 57:199–212, 1994
Plakun EM: Principles in the psychotherapy of self-destructive border-
line patients. J Psychother Pract Res 3:138–148, 1994
Waldinger RJ: Intensive psychodynamic therapy with borderline pa-
tients: an overview. Am J Psychiatry 144:267–274, 1987
Chapter 5
LEVELS OF CARE
Indications, Structure, Staffing
IN THIS CHAPTER, I offer condensed statements about the indications

for deploying four different and decreasingly intensive levels of care (Fig-
ure 5–1):
• Level IV: Hospital

• Level III: Residential/partial hospital/day treatment
• Level II: Intensive outpatient
• Level I: Outpatient
Experienced clinicians know that the capability of programs on levels

IV and III (hospital and residential programs) to help borderline pa-
tients varies greatly. This chapter may help those who administer or de-
velop such services to provide them more effectively. Emerging from the
cumulative clinical experience in the 30 years since the borderline diag-
nosis was established is awareness that effective therapies at any level of
care require modifications of traditional practices that can be effective
for other diagnostic groups (most notably, psychoses or depression).
With each decrement in level of care, the need for patients to have spe-
cialized (i.e., borderline personality disorder [BPD]–specific) services is
increased. The possibility for borderline patients to make significant and
enduring changes usually rests with therapies residing in outpatient set-
tings (levels I and II). Just as with patients or families (see Chapter 3), ed-
ucation of staff about BPD by talks or reading material at all levels of
treatment can alter attitudes in ways that help these patients (Miller and
Davenport 1996).
113
FIGURE 5–1. Levels and functions of care for borderline personal-

ity disorder patients.
Selecting or Changing a Level of Care

My discussion of each level of care includes descriptions of the structures,
staffing, and clinical processes that best serve the needs of borderline pa-
tients. Table 5–1 summarizes some aspects of this with special reference
to the five basic therapeutic processes described earlier in this book (see
Chapter 3). It is important in selecting the appropriate level of care to
consider the following general principles: 1) the least restrictive level of
care is best—this encourages use of the patients’ skills and capabilities in
community living; 2) having all or most levels of care is clinically desirable
and probably cost beneficial (Gunderson et al. 2005; Quaytman and
Sharfstein 1997). Although I suggest in this chapter that all borderline
patients progress by moving sequentially through all levels of care, this
process is rarely feasible and, fortunately, is rarely necessary. Most border-
line patients use brief hospitalizations (level IV) for crises but otherwise
participate in nonintensive outpatient care (level I). The presence and
use of specialized partial hospitalization/day treatment (level III) or in-
tensive outpatient treatment (level II) offer opportunities for behavior
change and rehabilitation that are highly beneficial options that, regret-
tably, are often unavailable. The absence of either of these intermediary
levels of care extends the frequency and duration of time needed at the
hospital level. Similarly, stays in residential/partial hospital (level III)
Levels of Care ❘ 115
programs need to be greatly extended either in the absence of an inten-

sive outpatient service (level II) or in the absence of the skilled profes-
sional help needed for level I care. Hence health care systems that do not
have all the levels available and/or do not have experienced BPD special-
ists will be cost-ineffective (Quaytman and Sharfstein 1997).
Ideally, borderline patients may progress from settings that provide
greater safety and security to settings that foster independence and new
capabilities. When progressing from one level to a lesser one, many bor-
derline patients have angry or regressive reactions to the step-down, even
if they have urged such a step. Because of these predictable reactions, it is
absolutely essential that administrators hold steady with expectations of
patients and that continuity of key treaters (i.e., primary clinicians and
therapists) is maintained as patients change levels of care. Step-downs of-
fer excellent opportunities for clinical staff in all roles to help borderline
patients realize and accept that they want more nurturance, caregiving,
and attention than they might otherwise recognize or—and this is com-
mon—might otherwise wish to acknowledge. Understanding this dy-
namic and its implications is one of the central psychological themes for
borderline patients in the early phase of treatment, regardless of the level
of care or the type of modality.
Level IV:
Hospital Treatment—Makes Therapy Possible
Until the 1990s, long-term hospitalizations were feasible and considered
to be highly desirable options for treating BPD. Although there remains
a role for long-term hospitalizations (at least 6 months), the indications
for it are rare (Sidebar 5–1). Hospitalizations now are usually 2–14 days in
duration, and the following discussion focuses on such short-term stays,
which are both the norm and usually more beneficial (Gunderson 1984;
Gunderson et al. 2005; Nurnberg and Suh 1978; Sederer and Thorbeck
1986; Silk et al. 1994).
Sidebar 5–1: Can Long-Term Hospitalization

Be Desirable for BPD?
Long-term hospitalizations (meaning a minimum of 6
months and usually a year or more) have become infre-
quent, largely as a result of the cost consciousness of the
managed care environment. Yet even when long-term
hospitalizations were more available, they were in my expe-
rience rarely useful for borderline patients: too much con-
tainment led to a regressive dependency or a paranoid
combativeness. These were nicely documented in the
116
TABLE 5–1. Four levels of care for borderline patients
LEVEL TYPE AND ITS COMPONENTS GOAL AND ITS COMPONENTSa EXPECTABLE LENGTH OF STAYb PROCESSc
❘
IV Hospital Crisis management 2–14 days C
Medication Assessments
Psychoeducation Treatment planning
III Residential/partial hospital Social rehabilitation

Night care Daily living skills (eating, sleeping) 2 weeks–3 months C, St, Su
Day treatment Social skills 2 weeks–6 months St, Su, I
Alliance building (relational) 2 weeks–6 months St, Su, I
II Intensive outpatient Social (behavioral) adaptation

Self-assessment group Socialization 1–3 months St, Su
Case management Anticipation, planning 6–18 months St, Su, I
Dialectical behavior therapy Impulse/affect control 6–12 months St, Su
Family prescription Alliance building (relational) 2–6 months St, Su, I
Levels of Care
TABLE 5–1. Four levels of care for borderline patients (continued)
LEVEL TYPE AND ITS COMPONENTS GOAL AND ITS COMPONENTSa EXPECTABLE LENGTH OF STAYb PROCESSc
I Outpatient Psychological growth

Group prescription Skills training 5–12 months St, Su
Psychotherapy Interpersonal skills 1–2 years I
Intrapsychic change 1–6 years I, V
Alliance building 6 months–4 years St, I, V
aEach goal component is aligned with the expectable length of stay for that goal (in the next column to the right), as well as with the processes
involved in the goal component (in the column furthest right). Type components do not relate one for one to the goal components, lengths
of stay, or processes.
bLengths of stay are estimates based on the author’s experience when the appropriate step-down level of care is available.
c
C =containment; I= involvement; St= structure; Su= support; V=validation (see Chapter 3 for explanation of terms).
❘117
book and movie Girl, Interrupted (see Appendix). It proved

unrealistic to hope that analysis of either of these negative
reactions could have much meaning as long as the non-
specific factors of being held, taken care of, and given un-
solicited attention were sustained.
Evidence from Henderson Hospital in England indi-
cated that borderline patients selected for long-term
(1-year) hospitalization did get better in such settings and
that their improvement was positively correlated with the
length of stay (Dolan et al. 1997). About 43% of the patients
showed significant improvement on a borderline symptom
self-report, compared with about 18% of an otherwise
comparable sample who were referred but did not receive
funding. These results are suggestive, but they do not indi-
cate whether a well-designed, less-intensive (such as resi-
dential) level of care might have been more effective.
Long-term hospitalization of borderline patients may be
considered desirable by the patient’s significant others, who
otherwise feel overwhelmed by issues of the patient’s safety
and health. Specifically, it will often seem desirable for par-
ents or spouses to have their family member with BPD
treated apart from the family to avoid their being exposed
to the patient’s disruptive or dangerous behaviors. Although
long-term hospitalization can be good for the welfare of the
family, they still should use this respite to develop better strat-
egies to cope and assist with their family member’s recovery
(for more about family interventions, see Chapter 8).
Exceptions to the generally negative effects of long-
term hospitalization can occur with exceptional patients or
exceptional hospital treatment conditions. The excep-
tional borderline patients for whom long-term hospitaliza-
tions (or residential care) are desirable include those in
need of in loco parentis functions—specifically, adoles-
cents with chaotic or excessively punitive home environ-
ments. Another group for whom long-term hospitalizations
can help are women whose borderline personality traits
emerged from the stress of too much child-care responsi-
bility. A third group are borderline patients whose failure to
control their impulses is frequent and dangerous enough
that the initial task of learning self-control requires many
months of containment.
The exceptional hospital treatment conditions that
can make long-term stays effective involve a clinical ad-
ministration that can give—or insist on—privileges for the
patients to engage in community-based or vocational ac-
tivities that allow their longings for attention and depen-
dency to be meaningfully awakened, identified, and
understood. Long-term hospital units typically become so
inner-directed and totalitarian that it is very difficult to do
this. The apparent success reported at Henderson Hospital

was attributed to a nonauthoritarian therapeutic commu-
nity (Jones 1952) that expected the borderline patients to
be actively involved in decisions and self-governance. Sim-
ilar therapeutic community assets can be found in long-
term residential care facilities. Such therapeutic milieus re-
quire skilled staff. The issues of dependency and attention
usually will be better addressed by the structures and off-
unit requirements of day treatment programs (level III) or
even less containing (level II or I) programs.
Within the framework of the borderline patient’s overall treatment,

hospitals serve as the place for initiating or changing therapies or for
managing crises. Whereas patients without personality disorders who are
in crisis can benefit from a community outreach intervention that avoids
hospitalization, hospitalization may have some added benefits for those
with a personality disorder (Rosenbluth and Silver 1992; Tyrer et al.
1994). Almost all borderline patients do best with brief (1- to 3-day) to no
more than 2-week hospitalizations. It is wise to establish this time frame at
the point of admission to discourage regressive, idealized, or dependent
attachments (Sederer 1991). Usually, longer stays in a hospital occur not
because of their therapeutic value but because appropriate step-down
services (levels III or II) are unavailable (Lewin and Sharfstein 1990).
The importance of graduated and careful discharge planning is brought
home by the observation that many suicides occur just after discharge or
just before a mandatory discharge (Kullgren 1988).
Vignette
Ms. E, a 28-year-old woman with a history of substance abuse and promis-
cuity, has been prescribed increasing doses of tranquilizing and sedating
medications. She presented herself to her state mental health depart-
ment caseworker as affectively blunted and mentally dull. She angrily dis-
missed outright, or failed to follow through on, all treatments arranged by
her caseworker. It was unclear whether her obtunded mental state was
due to misuse of her prescriptions. Hospitalization was recommended to
adjust medications, evaluate her problems (e.g., hostility, missed appoint-
ments) with her psychiatrist and caseworker, and assess whether her di-
vorced parents could offer more consistent supports, including possibly
residence.
Goals: Contain Patients for Safety, Assessments, and

Treatment Planning
The containment (see Chapter 3) offered by hospitalization (level IV)
provides opportunities to do evaluations and treatment interventions
that would be impossible elsewhere. Following are the major goals and
the usual time required for meeting them:
• Crisis interventions. Hospitalization is responsive to acute agitation or

danger of suicidal or violent acts. (2–6 days)
• Extensive medical, neurological, or psychological evaluations. These evalua-
tions are more easily coordinated in hospital settings and, for some
borderline patients, may only be feasible there. (2–6 days)
• Development of a treatment plan and personnel. Such plans usually require
arranging for treatment continuity through appropriate step-downs
and through assessing the suitability of new therapy personnel. An es-
sential part of these processes is to identify the primary clinician (see
Chapter 4) who will be responsible for the patient’s treatment. For
primary clinicians, an essential first step is to define roles and goals
(i.e., establish a “contractual alliance”; see Chapters 3 and 10) and to
contract with the patient about participation in aftercare services. (3–
14 days)
• Changes in prior therapies. These changes are often indicated, but they
may require expert consultation and the introduction of new thera-
pists. Inpatient settings can provide an essential and even lifesaving
role in recognizing and intervening in therapy impasses and therapies
doing more harm than good. If the changes are considered undesir-
able by the patient, working through resistance may be possible only
in the hospital, where the options for flight from the proposed
changes are reduced. Hospitalizations may allow therapists to review
prior impasses or establish a clearer treatment framework for their
ongoing work (Chapter 10). For many borderline patients, hospital-
ization serves as an environment to evaluate medication benefits and
to initiate medication changes (Chapter 6). (3–14 days)
Structure
To establish the businesslike, practical orientation that allows the above
goals to be reached efficiently, it is useful to have clarity and simplicity in
the inpatient units’ structures. This means a clear hierarchy, fixed roles,
and consistent policies. Each patient needs to have a case manager or co-
ordinator who processes the patient’s wishes and makes the administra-
tive decisions. The administratively responsible psychiatrist during the
hospitalization needs to assess the patient, preside over treatment plans,
delegate tasks, and prescribe medications. (Sidebar 5–2 provides discus-
sion of how psychotherapeutic technique relates to levels of care.) A so-
cial worker or the case manager or coordinator needs to assess with the
patient what are the available social supports, especially family, and in-
volve these support persons in planning aftercare. Psychoeducation of

families can begin (see Chapter 8). Family meetings can serve to improve
communication and avoid incendiary responses; both of these processes
may be useful in establishing a viable aftercare plan and may even help to
prevent future hospitalizations (Lansky 1989). Borderline patients who
resist such family involvement often pose significant problems for after-
care planning.
Sidebar 5–2: How Psychotherapeutic Technique

Relates to Level of Care
The function of containing splits often has been performed
by the staff of inpatient units who have helped borderline
patients appreciate the more benign significance of what
the patients often experienced as “angry” and “cruel” in-
terpretations, confrontations, or frustrations from their psy-
chotherapists. It is not accidental that Kernberg’s (Kernberg
1968; Kernberg et al. 1989) advocacy of these techniques
(see Chapter 12) arose from experience working at the
Menninger Clinic, where long-term (level IV) containment
was then being provided. In contrast, Adler (1986), who
worked in his level I private office with a short-term hospital
backup, became a champion for validating and empathic
interventions. Similarly, the successful outpatient therapies
advocated by Linehan (1993), Fonagy and Bateman
(2006), and Stevenson and Meares (1992) emphasized the
need for supportive techniques. More generally, a recipro-
cal relation exists between the amount of contextual (ex-
ternal) structure and support and the degree of structure,
directedness, and activity required by psychotherapists.
Note that short-term hospital programs should not offer community

meetings or group therapies that encourage cohesion or bonding.
Rather, groups oriented toward coping, crisis skills training, and psycho-
education are more valuable. Also, attending outpatient aftercare groups
while still in the hospital is very desirable even though managed care re-
imbursement policies often will not pay for these.
Staff
The ideal staff within hospital programs are comfortable but impersonal
about setting limits, recognize (preferably even enjoy) but do not enact
provocations (see Sidebar 5–2), and focus on the patients’ community liv-
ing situations and needs rather than on the patients’ in-hospital behav-
iors. Although staff can be selected with these attitudes in mind, the
development of this desirable approach often is acquired only by consid-
erable experience. This means that units should consciously avoid having
too many inexperienced staff and should actively inculcate these atti-
tudes for those who are new. The primary danger with inexperienced
staff is that they are often hesitant to actively direct the borderline pa-
tients’ attention to their precipitating situational crises and to the need to
plan their aftercare.
Generic inpatient units are capable of fulfilling very well the goals of
hospitalization suitable for borderline patients, but the staff of such units
need to be attuned to the special needs of borderline patients for clear
structure, treatment goals, and the staff supervision or meetings required
to safeguard against splits. Generic units that are too medically oriented
or too organized around the low-stimulus needs of psychotic patients will
be likely to foster—unwittingly—staff hostility toward the emotional and
time demands typically made by borderline patients. At best, hostility may
result in strict limits and early discharges, but this is less than optimal.
Units that do not welcome the challenges posed by borderline patients
are likely to aggravate the problems they dread.
Level III:
Residential/Partial Hospital Care/Day Treatment—
Basic Socialization
Level III includes residential care per se, meaning round-the-clock psy-
chiatric services in settings that are less intensively monitored and less
restrictive than are hospitals. Level III also includes two divisions: day
treatment and night care (usually a halfway house). These types of level
III services offer sufficient holding of the patient to reduce suicidality to
a degree that allows extramural activities (Stone 1990). During the pe-
riod that borderline patients spend in level III, they establish a contractual
alliance with their primary clinician by defining and agreeing on roles
and goals and begin work on a relational alliance with the primary clini-
cian and/or with a therapist (see Chapter 3 for discussion of types of al-
liance). Also in this level of care, the medication changes introduced
during hospitalization, or during recent outpatient upheavals, can be sta-
bilized, and both the benefits and the use of medications (e.g., compli-
ance) can be monitored. For patients in full residential care, it is critical
for the patient’s primary clinician or case manager to actively help the pa-
tient arrange for room and board (night care) if the patient will be stay-
ing in day care or to help arrange for structured community activities that
will enable the patient to leave day care while continuing to need night
care. As noted in Table 5–1, the primary goals of level III services involve
social rehabilitation.
Vignette
Arthur, a 16-year-old boy who lived with his mother, used both head bang-
ing and violence (breaking dishes, threatening to strike her) to intimi-
date and control her. His mother’s retreat into excessive use of anxiolytics
escalated his threats to the point that he threatened her with a knife.
School counselors were impressed by his aptitude and likeability but con-
cerned about his deteriorating school performance and identified a need
for a consistent, structured living situation to enable him to get to school
on time, to help him control his anger, and to help his mother develop
better coping strategies. His mother and his therapist agreed enthusiasti-
cally. An adolescent residential program that could allow him to com-
mute to school proved unavailable, so he went to a halfway house with
young adults.
Goals
• Teach or stabilize daily living skills (e.g., eating, sleeping, hygiene). The need
for this goal varies, as does the optimal approach to achieving it. Most
borderline patients need consistent monitoring and education about
the importance of eating and sleeping in regular patterns. Introduc-
tion of sleep medications may prove useful for borderline patients
who often will have trouble getting to sleep because of fearfulness.
• Initiate vocational rehabilitation. This goal is typically the most likely to
be overlooked. Borderline patients do not introduce it or welcome it,
even though these patients are typically underachievers with inconsis-
tent work histories. Young or inexperienced staff may have little con-
sciousness of the value and importance of these issues. In contrast,
this goal is often profoundly important to parents (and for anyone
with public health considerations). Program administrators or pri-
mary clinicians usually determine whether it is addressed.
An important component of level III services is the availability of
vocational rehabilitation services. The feelings and actions of border-
line patients so often preoccupy clinical staff that they can easily over-
look enduring impairments in social function (Sidebar 5–3).
Sidebar 5–3: Vocational Counseling: Should a

Borderline Patient Return to School, Pursue a
Career, or Become a Caregiver?
Returning to School
A common problem concerns the young borderline pa-
tient whose self-destructiveness has necessitated a hospi-
talization that interrupts a school term. Often, especially if
school performance was fine or if the student says he or she
wants to return to school, it becomes reflexive to support a

return. After all, the education itself is a valuable asset, and
the threat of being set apart (as well as lagging behind)
one’s peers is significant. Unfortunately, returning to school
virtually never succeeds if the youth’s parents or treaters
support this reflexively, and especially if they appear to be
optimistic. Returning to school occasionally succeeds
when the parents or treaters are explicitly opposed but
yield to the youth’s insistence. Their resistance may show
that the self-destructive behaviors are taken seriously and
conveys concern for the youth’s welfare above his or her
achievements. Support for returning to school fuels the bor-
derline patient’s fears that his or her caregivers want to ex-
pel or abandon him or her. For parents or treaters to “forbid”
a return may be a relief to some patients—primarily to those
who welcome a prolonged and regressive return to depen-
dency and flight from autonomy. To others, being forbid-
den may simply amplify rebellious, defiant behaviors and a
sense of alienation from “overcontrolling,” anti-indepen-
dence authorities.
Returning to school, especially when the person has
had prior school problems, always should be preceded by
holding down a steady job. This ability demonstrates
whether the person with BPD has motivation, concentra-
tion, conscientiousness, ability to accept external author-
ity, and willingness to complete what may at times be
undesirable tasks. These are prerequisites to succeed in
school, and they should be established in a context that
does not involve competition or the fear that advance-
ment means separation or autonomy.
Pursuing a Career
When borderline patients have been out of the work or
school marketplace for a sustained period—6 months or
more—the reentry (or entry) should be gradual. A general
principle is that this process should begin with a job that
does not test their capabilities and that does not have too
many implications for their eventual career (and sub-
sequent independence). Thus, the initial job placement
should be carefully attuned to their past work experience,
aptitudes, and long-term goals. The jobs that are recom-
mended should be ones with a high likelihood of success
and tolerable levels of anxiety. For the 30-year-old who has
not previously held a paying job, it is better to start as a vol-
unteer. For the future dress designer, it is better to start as a
clothing salesclerk than as a designer’s aide. For the son
who was expected to take over his father’s business, it may
be better to work alongside a supportive friend or relative.
It may be beneficial for a son or daughter who has an un-
derinvolved parent to work with or for that person.
Becoming a Caregiver
Many BPD patients will want to pursue or return to work that
involves caregiving functions. It is wise to caution them that
such work is invariably stressful for people who themselves
need caregiving and/or who perceive that they have not
gotten adequate care in the past. If patients nevertheless
insist on this field, encourage them to move slowly: plants
before pets before people. Certainly, the ability to assume
responsibility for a pet is a useful indicator of aptitude. Here
too there is a hierarchy: start with fish, go to rodents, then
cats, and finally dogs. For patients who insist on pursuing
the delivery of human services, the likelihood of success will
be inversely related to the likelihood of negative (hostile or
critical) feedback and the level of responsibility. Thus, work-
ing with people who have dementia is better than working
with adolescents, and working as an aide is better than
working as a nurse.
• Identify/modify gross maladaptive behavioral (impulse-control) and interper-

sonal (affect recognition/tolerance) traits. To identify maladaptive behav-
iors and traits requires that staff of the residential program repeatedly
clarify or confront BPD patients about the dysfunctional and undesir-
able aspects of themselves (e.g., behaviors that are attention seeking;
traits, such as maladaptiveness, that prevent patients from achieving
their goals—for example, of being praised or preventing rejections).
This work is done to make behaviors or traits ego-dystonic that have
previously been ego-syntonic (e.g., bullying, hiding feelings, or pro-
crastinating). In the vignette earlier in this section, the mother’s
behavioral responses to Arthur’s tyranny were unwittingly positive re-
inforcement, the modification of which could be initiated (and re-
peatedly reinforced) within the structures and professional attitudes
of a well-run level III service.
Staff
To facilitate attachment, identification, and transferences, it is desirable
for the staff of level III services to have a mixture of gender, age, levels of
experience, and even attitudes (see discussion of splitting, Chapter 4).
Regular staff meetings are needed to facilitate communication, examine
countertransference, address splits, retain focus on goals, assess progress,
provide education, and develop a case formulation (a way of understand-
ing the sources and meaning of the patients’ symptoms).
As in hospitals (level IV), each patient needs a staff case manager or
coordinator, preferably a full-time nonprofessional mental health
worker, whose responsibility is to implement the treatment plan, monitor
progress, and help patients address the how-to issues of coping with daily
life and goal attainment. This person needs to be in regular communica-
tion with the patient’s primary clinician (usually the therapist) to imple-
ment or change treatment plans. When, as is usually the case, the level III
service has its own administrative personnel, the case manager or coordi-
nator needs to clarify whether or when the patient’s primary clinician or
therapist defers decision-making authority to the program administrator.
This, I think, should depend on the expected length of stay: for residen-
tial stays of less than 1 month, the primary clinician or therapist is best left
in charge; for longer residential stays, the program administrator should
at least share authority. Failure to clarify these roles often renders the
level III care useless, if not harmful.
The meetings of staff coordinators with patients should be frequent,
brief, and as needed, not sit-down “pseudotherapy” (inviting disclosure
of secrets or expression of feelings) sessions. When patients are very an-
gry or frustrated by the case manager’s message or style, they can and
should know who the case manager’s supervisor is and be encouraged to
take up the problem there. This principle of triangulation means that bor-
derline patients who are in institutional treatment should always—or at
least in the early phases of treatment—have an identifiable means of ap-
pealing their case. This prevents splits, diffuses rage, and offers useful
holding and learning opportunities. Within level III or IV programs, the
program administrator can perform this function if there are problems
with a primary clinician or therapist. Similarly, the primary clinician or
therapist can perform this function when there are problems with the
program administrator.
Structure
GROUP MEETINGS
The most important structures of residential/partial hospital/day treat-
ment programs involve group meetings. These can be divided into those
for the entire community and more focused or time-limited types. Com-
munity meetings are for all patients and staff, whereas the group ther-
apies with staff leadership are for patients assigned by virtue of their
problems, and recreational/expressive groups are elective.
• Community meetings. Typically scheduled in mornings to maximize at-

tendance by patients and staff, these meetings help establish a sense
of community among patients by focusing on ward administrative is-
sues, such as policies or disruptive events, and inviting discussion of
feelings or opinions. We believe there are advantages in 1) having
such meetings three or more times a week, 2) making attendance
mandatory for everyone who is not in seclusion and who has no au-
thorized nonhospital-based activities (e.g., a job interview), and 3)
having meetings led or co-led by the clinician in charge of the unit.
The effectiveness of the long-term inpatient unit at Henderson Hos-
pital in England rested heavily on this form of therapeutic commu-
nity, emphasizing patients as collaborators (Dolan et al. 1997). In level
III programs, the lower level of external containment requires more
staff leadership.
• Group therapies. Membership in a group will be based on whether the
group’s goals have relevance to patients, and thus assignment is not
controlled by the group leaders. These groups should meet three
times a week to allow cohesion and depth. Because of the potential for
borderline patients to overwhelm psychotic patients, it is usually best
not to put the two groups together. All groups in day- or night-care set-
tings require active, directive affect and anxiety-controlling leader-
ship. Good topics for these groups include family issues, vocational
issues, skills of daily living, mentalization or dialectical behavior ther-
apy (DBT) skills training, and the use of social skills training modules
(see Chapter 11).
• Recreational/expressive groups. Participation in these groups is often
elective. They invite borderline patients to be active participants. Rec-
reational outlets, such as exercising, cooking, carpentry, or even at-
tending community events, enhance social skills and encourage the
development of friendships over a common task. The expressive (e.g.,
collage, pottery, dance) groups can enhance self-esteem and offer op-
portunities for symbolic communication of conflicts and hopes. Be-
cause of the emotional expressiveness invited, the leaders of such
activities need to encourage verbalization and to be alert to the poten-
tial loss of control.
DAY TREATMENT (P ARTIAL HOSPITAL)/NIGHT CARE (HALFWAY HOUSE)

Note that day services and night services are not all—or necessarily—con-
nected to hospitals and therefore should not all be referred to as partial
hospital services, except in the sense that they are clearly a transitional
level of care between hospital and outpatient care. Moreover, the term par-
tial hospital almost always refers to day services. Transition from full resi-
dential care to either day treatment or night care alone is best initiated by
1) advising patients that the transition will be difficult insofar as they will
miss their supports and will need to assume added responsibilities for
themselves and 2) encouraging them to make the transition on a trial basis
because to suggest that transitions are irreversible will beget angry, panicky
behavioral responses. (For borderline patients needing this level of care,
learning to contain and verbalize their responses to less support often will
become an indicator that a lower level of care is becoming possible.)
Day treatment (living in the community while receiving structured
treatment 3 or more hours a day for 3–5 days each week) is the usual step-
down from residential (day and night) care. It allows the social rehabili-
tation goals of the day program to continue. Day treatment alone is for
patients who have a reasonable place to live, meaning they are safe from
lethal, self-destructive acts and are able to take care of the basic tasks of
self-care such as eating, sleeping, and responsibly using medications. The
clinical value of a long-term, well-organized day hospital program has im-
pressive empirical support (Sidebar 5–4). Most day treatment programs
are offered for too brief a period to achieve social rehabilitation goals.
Stabilization can be achieved in a matter of weeks, but social learning re-
quires at least a month and usually a minimum of 2–6 months.
Sidebar 5–4: Empirical Support for a Specialized

“Mentalization-Based” Day Hospital
Welcome confirmation of the clinical effectiveness of a day
hospital program has come from the English team of Bate-
man and Fonagy (1999). BPD patients (N =44) were ran-
domly assigned to receive 18 months of day hospital
service or to receive “general hospital” care. The latter con-
trol condition involved as-needed use of hospital and day
services with medications and community/outpatient fol-
low-ups. In contrast, the experimental condition provided a
more continuous care system in which medications were
used in conjunction with psychoanalytically oriented group
psychotherapy three times a week, psychoanalytically ori-
ented individual psychotherapy (by nurses, occupational
therapists, or psychiatrists) once a week, and expressive
therapy of a psychodrama type once a week. These spe-
cific modalities were coordinated in daily staff meetings,
senior psychotherapy consultation, and periodic case con-
ferences.
The 22 patients in the day hospital condition showed
significantly more improvement than did the 19 control pa-
tients (3 dropped out because of suicidality) in depressive
symptoms, suicidal and self-destructive acts, number of
hospital days (reduced), and social and interpersonal
functioning. These advantages were already evident by 6
months, and these distinctions grew at the 12- and 18-
month assessments. In follow-up of the 38 study subjects 18
months after the treatments, those who were in the day
hospital program continued to show improvement,
whereas the 16 control subjects did not (Bateman and Fon-
agy 2001). The investigators concluded that the initial 18
months of treatment set in motion longer-term rehabilita-

tive changes.
This study was notable in that, like the better-known
study of DBT (see Chapter 11), it reported very impressive
benefits from a coherent and well-informed approach to
borderline patients when compared with the usual treat-
ment offered to such patients. The symptomatic and inter-
personal benefits from the partial hospital treatment seem
considerably stronger than those observed after a year of
DBT. Whether this observation is due to the psychodynamic
focus of this treatment (as opposed to DBT’s behavioral fo-
cus), or to sample differences, or to the considerably
greater intensity and longer duration of the therapy are
empirical questions yet to be answered.
Although less specific to borderline patients, a 4-month
day treatment program at the University of Alberta Hospital
showed confirmatory results. That program is a psychody-
namically based therapeutic community, with emphasis
on group therapies, that is designed to treat patients with
personality disorders. Piper et al. (1993) showed that 60 pa-
tients who received treatment in this program had signifi-
cantly better outcomes than did 60 waiting-list control
subjects in four areas: interpersonal functioning, symptoms,
self-esteem, and satisfaction with their lives. These gains
were sustained on follow-up 8 months later. Although sup-
portive of Bateman and Fonagy’s conclusions, this study in-
volved only 14 BPD subjects, whose outcomes were not
broken out. These limitations in Piper’s study gain signifi-
cance insofar as results from a Norwegian study of partial
hospital effectiveness found that borderline patients are
likely to be less responsive than those with other diagnoses
(Wilberg et al. 1998).
Night care (usually in a halfway house) is important for BPD patients

who are able to work or go to school but who do not have sufficient sup-
ports or are otherwise unable to safely take care of themselves at night.
Symptomatic of the intolerance of being alone (Chapter 1), these prob-
lems make solitary living unacceptable, and the need for night care can
extend for considerable periods. The night care function is often per-
formed by community-based halfway houses or cooperative apartments,
but enlightened residential (day and night) programs will offer this as a
step-down. Listed in descending order of holding capability are an on-site
residential step-down service, an off-site halfway house, and an off-site co-
operative apartment. The selection between such options should be
based on the patient’s expected tolerance—the least “holding” option
being best. To make such a selection requires an understanding of the pa-
tient’s ego functioning, level of community supports, and relation to the

treating agencies. For counterdependent patients, it may be better to
“recommend” a more monitored setting than is actually needed. For re-
gressive patients motivated by secondary gains, it may be better to “rec-
ommend” less monitored settings (i.e., level II or even level I) than is
judged to be needed. In both instances, the eventual compromise may be
what works best.
Level II:
Intensive Outpatient Care—Behavioral Change
The intensive outpatient level of care is for patients who are able to man-
age some social role, such as some part-time school or work, and who
have adequate room and board. This level of care is still often unavailable
but is proven efficacious and is highly beneficial for many borderline pa-
tients, either as a direct step-down from hospitalization or as a step-up
from outpatient care. DBT involving 5 hours of treatment per week and a
system of coverage that contains crises should be classified as a specific
type of intensive outpatient care (see Chapter 11). It successfully dimin-
ishes hospitalizations. Clinical experience indicates that intensive outpa-
tient care (other than DBT) greatly reduces the need for residential/
partial hospital programs, in much the same way that residential/partial
hospital programs can reduce the need for hospitalizations. The success
of intensive outpatient programs depends on their offering sufficient
holding to counter regressive flights and to support sustained community
living. This holding function is directly related to the degree to which a
patient’s treatment is coordinated via frequent communications between
the clinical team members (Ruiz-Sancho et al. 2001). In the absence of
good, preferably standardized, communications, destructive romances
or rivalries can form between patients who share clinic therapists.
Links (1998) reviewed studies in which patients with personality dis-
orders received a type of intensive outpatient care called assertive commu-
nity treatment (ACT). ACT uses proactive interventions involving visits to
the patient’s place of residence, with a focus on assistance with the tasks
of daily living and in vivo counseling about relationships and work. Only
the original study of ACT by Stein and Test (1980) had a large enough co-
hort of patients with personality disorders (26) to examine their out-
comes. That study showed diminished use of hospitals, better compliance
with treatment, and decreased legal problems. Links concluded that ACT
has considerable promise for more severely impaired BPD patients but
that this model of intensive outpatient care, like the one being described
and advocated in this chapter, awaits empirical testing.
Vignette
Ms. F was referred for treatment because the small town she lived in had
no psychiatric facilities. Her primary care physician thought that hospi-
talization would provide the containment needed so that she could make
use of medications and psychotherapy—therapies that until then she had
been noncompliant with—and could terminate her bulimic and rageful
behaviors, which were “destroying her family.” During the intake inter-
view, it was apparent that Ms. F was frightened of hospitalization. She pro-
tested that it wasn’t necessary, that she could live with her aunt, and that
the only reason she saw for relocating for treatment was that it could offer
opportunities to become a dancing instructor. It was recommended that
she come to daily self-assessment groups and have an extended psychiat-
ric evaluation. During the next week, it became apparent that she could
responsibly use the program, that instruction in dancing could be ar-
ranged, and that living with her aunt stabilized what was an 8-year pattern
of bulimia.
Goals
• Vocational. Enlist in needed training and/or obtain work (see Sidebar

5–3).
• Interpersonal. Recognize anger and dependent needs as part of self and
others.
• Behavioral. Improve abilities to contain impulsive expressions of feel-
ings or attitudes or impulsive acts.
Components
Figure 5–2 diagrams the components of level II outpatient services and
the relative lengths of time during which patients participate. Discussion
of these components, except for medications, follows:
• Self-assessment groups. Such groups, available every day (a minimum of

three times a week), are the backbone of an intensive outpatient ser-
vice. They function well for as few as 3 patients and as many as 10. The
group leader needs to be an experienced clinician who is comfortable
with both group dynamics and crisis interventions. Coleadership, as
with other groups having predominantly borderline members, serves
to diminish burdens and countertransference enactments and to
ensure continuity. Participants’ attendance is flexibly arranged: they
can come as often as every weekday or as infrequently as once a week.
They should expect to remain from 2 to 8 weeks. Crises are referred to
the patient’s primary clinician.
Self-assessment groups provide a social network because most mem-
bers will be dealing with similar immediate issues of transition into com-
Self-assessment groups
FIGURE 5–2. Level II intensive outpatient services for borderline

personality disorder.
DBT= dialectical behavior therapy; IPG= interpersonal group.
munity living. The frequency of meetings allows participants to become

familiar quickly with the details of one another’s current lives. Common
issues include the attitudes encountered on returning to families or co-
workers, feelings related to losing prior supports (e.g., school peers or
treatment staff), and the difficulties involved in working with new men-
tal health staff and new therapeutic modalities.
Groups meet daily, but participants attend with variable and usually
gradually diminishing frequency. Each group begins with a go-around,
in which each participant reports on recent events or plans about which
he or she receives feedback from others. Because members all have one
foot in community living, they are vigilant to signs of retreat or regres-
sion in one another. Groups meet for an hour, preferably in the early
morning or late afternoon so as to minimize conflicts with work or
school responsibilities.
• Case management. At the intensive outpatient level of care, the clinician
responsible for the patient’s care, the primary clinician or therapist,
needs to be active and direct in administrative decisions. There is no
longer a separate administrative structure such as exists on levels IV
and III. Now the primary clinician, in principle, is responsible for im-
plementing plans, assessing safety, communicating with family, seek-
ing consults, and changing the treatment (including shifts in level of
care). In practice, this is often not done well, either because 1) the pri-
mary clinician or therapist is not geographically present and immedi-
ately accessible or because 2) he or she may be trying to establish a

role as a dynamic therapist who can see the patient’s behavior prob-
lems as a subject for interpretation rather than for management. Both
of these problems can lead to splits by the patient.
Leaders of self-assessment groups will encounter occasions when
they need to inform the primary clinician about what actions are
needed. Sometimes this is not possible, and they themselves need to
get involved in crisis management. The primary clinician can actively
help a patient anticipate problems and select coping methods that are
adaptive. Usually, borderline patients at this level of care are able to
do psychotherapeutic work such as identifying feelings and how they
relate to current issues, including the clinician’s or therapist’s func-
tioning. Either the primary clinician should be able to accommodate
and conduct such work or a psychotherapist needs to be added (see
Chapter 9).
• Other activities. The minimal essential components of intensive outpa-
tient care are self-assessment groups and the case management and
therapy activities cited earlier. Preferably, these are complemented by
functions that introduce and stabilize the longer-term modalities that
patients will need for continued progress when they enter level I (out-
patient) care. This level of care is particularly suited to family interven-
tions designed to identify and diminish the triggers that lead to crisis
(see Chapter 8). As noted, the staff members who offer level I services
need to include someone who can do psychotherapy. In addition, a
second collaborating and complementary modality (in accord with the
principle of split treatment, discussed in Chapter 4) is best initiated
while the patient is still attending self-assessment groups. Either a DBT
or other skills training group or an interpersonal group (see Chapter
9) is an ideal complement to psychotherapy. Other group options that
can begin as part of the intensive outpatient level are vocational reha-
bilitation, family issues, and either trauma or alcoholism recovery
groups. Finally, this is an excellent time to get involved in self-help or-
ganizations (e.g., Alcoholics Anonymous or Overeaters Anonymous)
or community organizations (e.g., churches or volunteer groups).
Level I:
Outpatient Care—Interpersonal Growth
Outpatient care (level I) is when critically important changes in interper-
sonal and intrapsychic functioning can occur. For most borderline pa-
tients, the first year also involves continued work on significant behavior
problems; however, at the outpatient care level, this continued work takes
place concurrently with the development of a relational alliance with the

primary clinician or therapist (see Chapters 10–12). A major function of
the primary clinician is to facilitate the patient’s transition from a higher
level of care (usually level IV, but often from level III or II) to the less-
intensive outpatient care. This process begins by identifying a suitable
second modality to accompany the primary clinician’s or therapist’s on-
going work. At this level of care, the split treatment (Chapter 4) often in-
volves the combination of a psychiatrist overseeing medications and
another mental health professional doing psychotherapy. When a psy-
chiatrist is doing both of these tasks, the second modality should include
a social rehabilitative component—cognitive-behavioral therapy (see
Chapter 11) or skills training group (see Chapter 9) and/or some con-
tinuation of family involvement (see Chapter 8). These therapies ideally
begin while the patient is still in a higher level of care.
Vignette
Ms. G, a 28-year-old single woman, recently lost her job and returned to
live with her mother and stepfather after a year’s absence. She had lost
her job because of absences necessitated by hospitalizations for suicidal
impulses. After the last of these, her hospital psychiatrist had referred her
for aftercare. At intake, she noted that, since returning, she had gotten a
new job, and her suicidality had diminished. She was hesitant to relate
these changes to returning home insofar as she reported long-standing
conflicts with her mother. She also reported that her problems were re-
lated to several romances that “ended badly” because of her “losing her
own identity” (e.g., opinions, interests) and that were followed by a pro-
gressive reluctance to socialize because of fears of rejection. Her ability to
work and her history of nonpromiscuous romances were judged to be
strengths that should be encouraged. It was recommended that Ms. G be-
gin once-weekly psychodynamic therapy and an interpersonal group. She
was encouraged to have her parents read materials about BPD (see Ap-
pendix at the end of this book) to better understand her problems.
The outpatient level of care is the most extended. It is also the most
unpredictable in its expectable duration because the motivation for
change, the vicissitudes of life, and the goals and skill of the outpatient
therapists are so variable. At this level of care, basic personality change is
sometimes possible over a period of years.
Summary
In this chapter, I describe distinctions in the goals, functions, structures,
and lengths of stay for four levels of care that can be appropriate for bor-
derline patients. Most clinical sites do not offer the intermediary levels of
care (III and II) yet are still able to do well for many borderline patients.
For those who need a prolonged social rehabilitative experience (e.g.,

many adolescents and those with sustained dangerous habits such as sub-
stance abuse), the absence of levels III or II is costly and ineffective. The
principle of split treatments (Chapter 4) is again introduced, meaning that
two complementary forms of treatment can diminish flight and enhance
effectiveness by containing splits.
In the background of the discussion about levels of care is the need
for a primary clinician or therapist who works in conjunction with those
responsible for administering services or modalities to make thoughtful
judgments about 1) what level of care a borderline patient needs, 2)
when it is time to change that level, and 3) who facilitates these changes.
Moreover, the clinician who will be part of the patient’s longer-term out-
patient care, preferably along with others (e.g., DBT group leader, family
therapist), should anticipate that any changes to lower levels of care will
be experienced as losses and should therefore be prepared to help bor-
derline patients put the related issues of abandonment and dependency
into words. Borderline patients need to be involved in reaching judg-
ments about levels of care, but until they no longer need such services, it
is incompatible with their psychopathology to expect them to make these
judgments wisely.
References
Aronson, 1986
Bateman A, Fonagy P: The effectiveness of partial hospitalization in the
treatment of BPD: a randomized controlled trial. Am J Psychiatry
156:1563–1569, 1999
Bateman A, Fonagy P: Treatment of borderline personality disorder with
psychoanalytically oriented partial hospitalization: an 18-month fol-
low-up. Am J Psychiatry 158:36–42, 2001
Dolan B, Warren F, Norton K: Change in borderline symptoms one year
after therapeutic community treatment for severe personality disor-
der. Br J Psychiatry 171:274–279, 1997
Fonagy P, Bateman A: Progress in the treatment of borderline personality
disorder. Br J Psychiatry 188:1–3, 2006
Gunderson JG, Gratz KL, Neuhaus EC, et al: Levels of care, in The Amer-
ican Psychiatric Publishing Textbook of Personality Disorders. Ed-
ited by Skodol AE, Bender DS, Oldham J. Washington, DC, American
Psychiatric Publishing, 2005, pp 239–256
Jones M: Social Psychiatry: A Study of Therapeutic Communities. Lon-
don, Tavistock, 1952

Kernberg O, Selzer M, Koenigsberg HW, et al: Psychodynamic Psycho-
Kullgren G: Factors associated with completed suicide in borderline per-
sonality disorder. J Nerv Ment Dis 176:40–44, 1988
Lansky MR: The subacute hospital treatment of the borderline patient:
management of suicidal crisis by family intervention. J Clin Psychia-
try 11:81–97, 1989
Lewin R, Sharfstein SS: Managed care and the discharge dilemma. Psy-
chiatry 53:116–121, 1990
Links PS: Developing effective services for patients with personality dis-
orders. Can J Psychiatry 43:251–259, 1998
Miller SA, Davenport NC: Increasing staff knowledge of and improving
attitudes toward patients with borderline personality disorder. Psy-
chiatr Serv 47:533–535, 1996
Nurnberg HG, Suh R: Time-limited treatment of hospitalized borderline
patients: considerations. Compr Psychiatry 19:419–431, 1978
Piper WE, Rosie JS, Azim HFA, et al: A randomized trial of psychiatric day
treatment for patients with affective and personality disorders. Hosp
Quaytman M, Sharfstein SS: Treatment for severe borderline personality
disorder in 1987 and 1997. Am J Psychiatry 154:1139–1144, 1997
Rosenbluth M, Silver D: The inpatient treatment of borderline personality
disorder, in Handbook of Borderline Disorders. Edited by Silver D,
Rosenbluth M. Madison, CT, International Universities Press, 1992,
pp 509–532
Ruiz-Sancho AM, Smith GW, Gunderson JG: Psychoeducational ap-
proaches, in Handbook of Personality Disorders: Theory, Research,
and Treatment. Edited by Livesley WJ. New York, Guilford, 2001, pp
460–474
Sederer LI (ed): Inpatient Psychiatry: Diagnosis and Treatment, 3rd Edi-
tion. Baltimore, MD, Williams & Wilkins, 1991
Sederer LI, Thorbeck J: First do no harm: short-term inpatient psycho-
therapy of the borderline patient. Hosp Community Psychiatry 37:
692–697, 1986
Silk KR, Eisner W, Allport C, et al: Focused time-limited inpatient treat-
ment of borderline personality disorder. J Personal Disord 8:268–
278, 1994
Stein LI, Test MA: Alternative to mental hospital treatment. Arch Gen
Psychiatry 37:392–397, 1980
Stevenson J, Meares R: An outcome study of psychotherapy for patients

with borderline personality disorder. Am J Psychiatry 149:358–362,
1992
Stone MH: Multimodal therapy: applications to partial hospitalization in
the light of long-term follow-up of borderline patients. Int J Partial
Hosp 6:1–14, 1990
Tyrer P, Merson S, Onyett S, et al: The effect of personality disorder on
clinical outcome, social networks, and adjustment: a controlled clin-
ical trial of psychiatric emergencies. Psychol Med 24:731–740, 1994
Wilberg T, Friis S, Karterud S, et al: Outpatient group psychotherapy: a
valuable continuation treatment for patients with borderline person-
ality disorder treated in a day hospital? a 3-year follow-up study. Nord
J Psychiatry 52:213–221, 1998
Chapter 6
PHARMACOTHERAPY
Clinical Practices
History
The role of psychotropic medications was extremely peripheral to discus-
sions of treatment of borderline personality disorder (BPD) in the 1970s,
when psychoanalytic perspectives predominated. The role of these med-
ications began to be actively explored in the early 1980s as a result of the
existence of standardized criteria and reliable assessments, the medical-
ization of psychiatry, and a growing appreciation for the value of medica-
tions for other disorders. The initial considerations about medications
reflected the question of whether BPD was an atypical form of another
disorder, schizophrenia (see Chapter 2). The issue of whether BPD was
an atypical form of schizophrenia was originally examined by Brinkley et
al. (1979), whose pioneering but noncontrolled account encouraged use
of low-dose neuroleptics. The issue rapidly switched to the boundary of
BPD with depression in response to provocative accounts by Akiskal
(1981), Klein (1975, 1977), and Stone (1979) (Chapter 1). All three of
these psychiatrists had clinical and empirical experiences suggesting that
BPD was an atypical form of depressive disorder that might prove respon-
sive to antidepressant medications. The initial series of controlled studies
investigating these boundaries suggested that the response of BPD to
antipsychotics or antidepressants was not as impressive as would be ex-
pected were BPD an atypical offspring of either of these parent con-
ditions (Cowdry and Gardner 1988; Goldberg et al. 1986; Soloff et al.
1986). These studies did, however, show that both types of medication
can be helpful in BPD and thereby opened up an exciting and still ongo-
ing era of pharmacotherapeutic optimism.
Even as the early projects made it clear that neither traditional anti-
psychotics nor antidepressants offered very strong answers to the ques-
139
tion of “borderline to what,” these medications also had some more

specific, surprising, and important results. For example, the studies by
Cowdry and Gardner (1988) and Soloff et al. (1989) showed that anti-
psychotics were as effective in diminishing depression as were antidepres-
sants. A second finding was that the dramatic effects that the first few
weeks of hospitalization can have on reducing presenting symptoms
made it impossible to discern the effects of medications initiated in that
context (Siever and Davis 1991; Soloff et al. 1989). A third finding was
that borderline patients’ judgments about the benefits of a medication
could differ dramatically from judgments made by professionals. Medi-
cations most favored by borderline patients appeared actually to make
them worse, in the judgment of others, whereas the type of medication
they disliked most was judged most beneficial by others (Cowdry and
Gardner 1988). A fourth finding was that although many types of medi-
cations could be helpful, no type proved consistently beneficial.
Overall Role of Medications

Whereas a previous generation of clinicians worried about whether, or
under what circumstances, medications should be added to therapy, such
concerns (for better or for worse) are now rare. Even in the 1980s, only
about 10% of psychiatrists treated BPD without medications (Cole et al.
1984; Pope et al. 1983; Skodol et al. 1983; Soloff 1981; Waldinger and
Frank 1989a; Zanarini et al. 1988). Such treatment is now truly rare. For
many psychiatrists, the question has become whether it is ever reasonable
to forgo pharmacotherapy. Undervaluation of medications is unlikely,
and many psychiatrists even consider it unethical to withhold drugs. As a
group, psychologists are only slightly less prone to advocate their use. An
ongoing prospective follow-up study of a sample of personality disorder
patients selected from many clinical sites in four northeastern cities
found that at baseline, 90% of the borderline patients had received psy-
chotropic medications—a significantly higher percentage than a com-
parison sample of patients with major depression (Bender et al. 2001).
The use of medication in the care of borderline patients has transformed
from being occasional to being expected to frequently administering
multiple medications. Zanarini and colleagues (2004) documented that
polypharmacy is now very common. They found that over 6 years of fol-
low-up, 40% of the patients with BPD had been taking three or more con-
current psychotropic medications, 20% had been taking four or more,
and 10% had been taking five or more medications.
Medication effects are difficult to assess in borderline patients, for
three very basic reasons:
Pharmacotherapy: Clinical Practices ❘ 141
1. Many of the symptoms that are the targets of medications are very de-
pendent on context. As a result, psychiatrists who lack experience
with borderline patients can easily attribute too much benefit to med-
ications (e.g., hospitalized patients whose depression disappears), or
too little (e.g., discharged patients who cut themselves), when pa-
tients’ symptoms are really the product of predictable changes in
their level of care.
2. Medications are used as vehicles for projection. It is very easy for bor-
derline patients to attribute changes in their moods to their medi-
cation. If they feel bad, the medications offer an easily discernible
and less painful explanation than, for example, the patients’ being
rejected. More will be said about this.
3. Medications are rarely, in my experience, dramatic in their effective-
ness. Their effect is almost always partial and modest.
If a borderline patient’s symptoms respond dramatically, by the
patient becoming essentially nonborderline as a result of medica-
tions, the borderline diagnosis was probably mistaken (Sidebar 6–1).
Such experience is illustrated in the vignette following Sidebar 6–1.
Sidebar 6–1: Listening to Prozac: Can Selective

Serotonin Reuptake Inhibitors Cure BPD?
Peter Kramer’s best-selling book Listening to Prozac (Kramer
1993) reported that selective serotonin reuptake inhibitors
(SSRIs) changed the personality (i.e., the attitudes, expec-
tations, level of energy, and overall mood) of his patients.
His report raised the expectations of many patients—and
their psychiatrists too. High expectations are also encour-
aged by drug trials that report borderline patients who stop
meeting criteria for the diagnosis after only 1–2 months of
taking medications.
Alas, Kramer’s hopeful message is not borne out by lis-
tening to borderline patients. Medications can decrease
the frequency of symptomatic acts and affects so that
DSM criteria can appear to have remitted. Thus, medica-
tions appear to have affected borderline personality .
These changes are not insignificant, and the prescribing
psychiatrist can easily exaggerate the benefits. Still, the
more core personality features of affective instability, mis-
attributions, and pathological object relations await other
means of intervention.
There remains hope that Kramer’s optimistic message
might one day be fulfilled, but the type of medication for
curing BPD has not yet been developed.
Vignette
A 30-year-old obese married woman, with a highly dependent relation-
ship with her husband (she called him four times daily), was given the di-
agnosis of BPD when she began self-mutilating activities. They occurred
in the context of the couple’s having decided, at the husband’s urging, to
apply for adoption. Eight months later, having taken an SSRI, she no
longer met DSM-IV (American Psychiatric Association 1994) criteria for
BPD. Indeed, she had stopped cutting, was working full-time, and had
ceased needing excessive reassurances from her husband. This remission
originally was thought by her psychiatrist to exemplify a medication cure
(like those that can be seen with depression or anxiety disorders). On
closer examination, the patient’s recovery did not actually begin until 4
months after starting the SSRI—too long a delay to assume that the SSRI
accounted for the changes. Moreover, the improvements began shortly af-
ter she and her husband decided to withdraw the adoption application.
This vignette illustrates the temptation to credit medications with

changes that might better be understood by examining life events. The
patient may well have had BPD, but the reasons for her regression and re-
covery were unlikely to be explained by her serotonin metabolism.
Because medications are primarily intended to reduce subjective dis-
tress (i.e., symptoms considered undesirable to the patient) and help
contain behavior problems (often considered undesirable by others),
they are a type of intervention that can have relatively rapid and desirable
effects (see Chapter 3 on the sequence of change). Medications also may
help with longer-term and later goals of treatment. In my experience,
medications are often, perhaps even usually, helpful for borderline pa-
tients. But, in the absence of knowledge about long-term risks and bene-
fits, and in the presence of the very real dangers of misuse, it remains
clinically and scientifically important to recognize that we lack empirical
justification for our usual practices.
The overall role of medications in the long-term care of borderline
patients still needs to be determined; however, there is consensus that
medications will not cure this disorder but can lead to significant reduc-
tion in symptoms and improvement in mental functioning. Overall, the
studies suggest that medications are mildly to moderately effective for an-
ger, particularly impulsive aggression, and modestly effective for depres-
sion. Beyond the usual, but often uncertain, helpfulness for these
symptoms, the continued proliferation of new types of medications fuels
the hope that their role may become more significant and certain. More-
over, as is pointed out throughout this chapter, the role of medications is
not static; it can be expected to change as the borderline symptoms im-
prove.
Getting Started
An important role of psychopharmacology involves its usefulness in engag-
ing and allying BPD patients and their families in treatment. By first
anchoring BPD psychopathology within medicine and biology, the psy-
chopharmacological approach underscores the “illness” (see Chapters 1,
3, and 13). This approach usefully diminishes unrealistic expectations that
the patient can willfully “get over it.” A survey reported by Waldinger and
Frank (1989b) showed that most borderline patients feel pleased and im-
pressed by the doctors who prescribe medications and that 92% of the psy-
chiatrists or therapists believed that prescribing strengthened the alliance.
Anchoring BPD within medicine and biology also prompts a less defen-
sive, more supportive posture by families regarding treatment (see Chap-
ter 8). Moreover, this approach conveys a proactive and hopeful attitude
about diminishing immediate symptoms that, if not oversold, is always wel-
come and helps establish the relational alliance (Chapter 3) needed for
longer-term goals involving psychological change. However, it is also criti-
cally important to convey to borderline patients and their families the
overall limitations of expectable benefits from medications to set the stage
for appropriate (i.e., multimodal) treatment. The expert consensus guide-
lines for BPD developed by the American Psychiatric Association (2001)
indicated that the “primary treatment of BPD is psychotherapy, comple-
mented by symptom-targeted pharmacotherapy.”
The Prescribing Psychiatrist’s Role

With the Borderline Patient
It is important for the prescribing psychiatrist to begin by clarifying what
the patient can expect from him or her, as well as from medications. Re-
garding the expectable benefits of medications, important messages to
convey to patients appear in Table 6–1. The overall message is to guard
proactively against too high expectations and to insist proactively on the
borderline patient’s collaboration in selecting targets and outcomes for
treatment and on his or her being an alert, well-informed consumer. The
following vignette illustrates how this can be elaborated into almost a sci-
entific N-of-1 experiment:
Vignette
A 38-year-old single woman agreed to a trial of citalopram to decrease her
undesirable impulsive and aggressive behaviors. Although she presented
with a variety of impulsive behaviors, her regular participation in sado-
masochistic sexual practices that often led to physical injuries was the mu-
tually agreed-on target for the citalopram “experiment.” We established a
TABLE 6–1. Guidelines for psychopharmacological treatment of

borderline personality disorder
1. Medications can be helpful, but their overall role is adjunctive. They
should not be expected to be curative. Convey cautious optimism
about expectable benefits.
2. Require the patient’s collaboration in identifying target problems/
symptoms that medications might reasonably benefit (e.g.,
stabilizing affects, undesirable behaviors, or distorted perceptions).
Choose an outcome (e.g., the amount of decrease in the
undesirable behaviors or symptoms) that would reflect the desired
response to the medication.
3. Outline the expectable time course by which benefits might occur.
4. Inform the patient about the possible adverse side effects and about
alternative medications. This helps patients to be active in decisions
and more frankly communicative. Before prescribing the
medication, evaluate the patient for symptoms that might possibly
be side effects of the proposed medication. This evaluation provides
a baseline for monitoring the appearance of new symptoms during
the trial of new medications.
5. Encourage the patient to read about whatever medications are
prescribed.
6. Stress that effects are difficult to evaluate, and enlist the patient as
an ally in this process.
7. Because noncompliance is common, stress the necessity for
meticulous and responsible use to evaluate effectiveness.
baseline rate of the frequency and severity of her impulsivity, including

the sadomasochistic practices. The patient developed her own self-report
format to track her behaviors. The patient was followed up weekly over
the next 12 weeks and documented that the sadomasochistic behaviors
were less frequent and much less severe in terms of risk of injury. We
agreed that the improvements might be explained by mechanisms other
than citalopram’s effect on selective serotonin reuptake inhibition, so the
patient was motivated to withdraw herself from the medication and to
continue to monitor her sadomasochistic behaviors. From this “experi-
ment” of being on and then off the medication and carefully monitoring
the target symptoms, the patient and I both concluded that the trial of ci-
talopram had been successful.
The guidelines in Table 6–1 establish an atmosphere of pragmatic em-

piricism not unlike that advocated for cognitive-behavioral therapies (see
Chapter 11).
As the Primary Clinician (Psychiatrist/Therapist)

If the prescribing psychiatrist is also the primary clinician (i.e., therapist),
his or her role among treaters working as a team is not usually a problem.
Moreover, the psychiatrist in this role has the opportunity for more fre-
quent and longer appointments, necessitated by his or her expanded
role and consequent greater knowledge of the patient’s contextual fac-
tors (e.g., forthcoming examinations or separations). This position on
the team offers the psychiatrist enormous benefits in assessing the role of
medications, and it can also offer an excellent opportunity to explore the
meanings assigned to the medications (see the following vignette).
Vignette
Ms. H, a 20-year-old woman with a history of recurrent sexual abuse and of
alcohol abuse, has long-standing and severe anxiety and depression. She
received treatment in the outpatient clinic with Klonopin, which re-
quired escalating dosages over a period of months and led to increasingly
desperate and angry calls that she had run out of the medication and
needed refills. Her psychiatrist feared that this pattern emerged because
she was giving the Klonopin to either her mother or her boyfriend. This
psychiatrist also did not want to undermedicate her real needs. He re-
ferred Ms. H to another psychiatrist, a resident, for psychotherapy. The
patient rapidly formed a strong attachment with the therapist. She began
to talk about breaking up with her drug-abusing boyfriend.
The prescribing psychiatrist now felt more able to set limits on Ms. H’s
requests for refills, and he began to taper her dose of Klonopin to a mod-
est level, which was maintained. At this point, the therapist and psycho-
pharmacologist discussed turning the medication management over to
the therapist. They concluded that doing so would offer more time and
better motivation by Ms. H to investigate the role that the medication plays
in her life. Moreover, they concluded that the patient’s potential anger
about not receiving more Klonopin could be processed within the therapy
without endangering their alliance or having the patient drop out.
The downside of the prescribing psychiatrist also serving as the ther-

apist is that the transference arising from the directive caregiver role can
become more heated. If unsuccessfully explored, this more intense trans-
ference can increase the likelihood of misuse of medications (e.g., non-
compliance, overdosing). As I discuss later in this chapter (see section
“Transference-Countertransference Issues”), the transference to the pre-
scribing therapist also can influence the borderline patient’s apparent re-
sponsiveness to the medications and the therapist’s prescribing behavior.
Waldinger and Frank (1989b) noted that when the prescribing psy-
chiatrist is also offering psychotherapy (i.e., when the psychiatrist is the
primary clinician, as described in Chapter 4), medication abuse takes
place about 50% of the time. The important empirical question remains
whether this rate is higher or lower when the medications are prescribed
by someone who manages only pharmacotherapy (in the arrangement
often called split treatment).
As a practical matter, the issue of medications is often set apart to be
dealt with in the first or last 10–15 minutes of sessions. Obviously, the
need to devote time to this issue should diminish considerably over a pe-
riod of 2–4 months unless problems with compliance or usage persist.
Such persistence may be an indication for splitting the treatment.
As the Psychopharmacologist Only (Split Treatment)

As noted in the earlier vignette about Ms. H, it is increasingly normative for
psychiatrists to split treatment responsibilities with other mental health
professionals who conduct the psychotherapy. This practice stems from the
cost-saving mandate of managed care—saving ostensibly by limiting the
amount of time psychiatrists spend with patients. This practice then dimin-
ishes the training, experience, and comfort that many psychiatrists have in
filling other roles—most notably, psychotherapy. When the prescribing
psychiatrist’s role does not include being the borderline patient’s primary
clinician (i.e., therapist), the relationship usually will be less intense, but
psychological splitting of the treaters (into “good” and “bad”) is more eas-
ily enacted.
It often becomes complicated to establish and maintain a clear defini-
tion of what the psychopharmacologist’s responsibilities are. Problems often
result from unstated role expectations. One of the psychopharmacologist’s
responsibilities, easily neglected when he or she is not the patient’s primary
clinician, is to communicate and discuss interventions with whoever is re-
sponsible for the patient’s overall care (the primary clinician)—often the in-
dividual psychotherapist. This communicative role is especially necessary
when medications are being changed and the effects of those changes need
to be monitored. Communication is equally important for assessing
whether changes in a patient’s symptoms are related to the prescribed med-
ications. For example, the isolated psychiatrist acting as only a psychophar-
macologist might attribute too much benefit to drugs for the patient whose
depression disappears after a boyfriend moves in or too little benefit to
drugs because that same borderline patient resumes cutting after the boy-
friend moves out. In both instances, the patient’s symptom changes are re-
ally the product of changes in interpersonal life.
A second complication involves coordination of roles and responsibil-
ities around safety issues. Because the prescribing psychiatrist is often the
legally (i.e., medically) responsible member of a team, a crisis plan needs
to be developed that includes him or her in assessing safety and in mak-
ing decisions about changing the patient’s level of care (Sidebar 6–2). In
any event, the psychopharmacologist needs to clarify how and when he or

she can be contacted for medication questions.
Sidebar 6–2: Liability Hazards of Split Treatment

Psychiatrists whose responsibilities are confined to psycho-
pharmacological management and who, as a result of
managed care incentives or personal preference, see their
borderline patients only briefly, often have disproportionate
legal responsibilities and liability risks. They often represent
the “deep-pocket” member of the treatment team. The
American Psychiatric Association’s (1980) Guidelines for
Consultative, Supervisory, or Collaborative Relationships
With Nonmedical Therapists state that the psychiatrist must
spend sufficient time to ensure that proper care is given and
warns against psychiatrists being used as figureheads (An-
notation 5, Section 3). Psychiatrists therefore should recog-
nize that most liability exposure is due to negligence of two
kinds: The first involves communication that is inadequate to
ensure that the psychiatrist understands and approves of
what other members of the treatment team are doing. The
second involves inadequate participation in decisions
about changes in the treatment plan. Adequate participa-
tion may require only that the psychiatrist’s approval is at-
tained—but the nonphysician primary clinician needs to
accept that treatment changes require such approval.
Gabbard (2000) noted that explicit discussion and
agreement between the prescribing psychiatrist and other
clinicians (usually the therapist) is the best defense against
liability. Explicitly, both should have the following:
• Adequate liability insurance

• Competence and credentialing in the treatment they
provide
• An agreement about whether the psychiatrist has su-
pervisory responsibility
• The patient’s agreement that patient and therapist will
discuss changes or concerns with each other
The disproportionate liability risk of psychiatrists may in-

crease their fears about getting involved with borderline
patients. The real implication of these risks is that the
greater the involvement with and knowledge about a pa-
tient a psychiatrist has, the less the psychiatrist’s liability risk.
Still, greater knowledge about the patient will not neces-
sarily diminish the risk associated with a psychiatrist’s need
to know and approve of the practices used by others who
are members of a borderline patient’s team.
Symptom Chasing
Symptom chasing with borderline patients can, at its worst, involve multiple
unsustained medication trials in pursuit of alleviating a patient’s transi-
tory, dramatized symptoms. It results in little relief of the underlying
problems impelling the patient’s complaints and in little learning about
whether medications could be useful. It may further result in a patient
who is chronically overmedicated.
At its best, though, symptom chasing is a reasonable extension of the
pragmatic, empirical approach cited earlier. The prescribing psychiatrist
should be aware that the borderline patient’s needs for medication
change over time. The patient who is overly constricted but intermit-
tently explosive may profit from a regimen different from the one he or
she will need later in treatment, when he or she may be depressed and
fearful of abandonment. Within an even more transient time frame, the
borderline patient who is reentering school may have needs for sleeping
medications or anxiolytics that were previously unnecessary and may be
unnecessary again in a few months. Adjusting medications is good psy-
chopharmacological practice. It is responsive to the patient’s changing
needs, and it sustains an ongoing collaborative alliance (see later in this
chapter, section “Contraindications and Discontinuance”).
Attitudes, Meanings, and Attributions

Table 6–2 identifies the dichotomous (split) thinking about both the
medication and the prescribing physician that borderline patients often
bring to bear when medications are prescribed. Psychopharmacological
interventions should be accompanied by an awareness of such possibili-
ties. This table chronicles what Koenigsberg (1994, 1997) described as
the important meanings that borderline patients can attribute to medi-
cations. These attributions confound a patient’s compliance and also a
clinician’s interpretation of the actual value of the medications. When
positive meanings are attributed to medications, as is most often the case,
this reaction should be accepted gratefully by the clinician but watched
carefully lest it inflate the patient’s evaluation of benefits. When negative
attributions about drugs are present, they need to be taken seriously. The
patient’s subsequent report of side effects or lack of benefit will be col-
ored by this kind of attribution.
First and foremost in the patient’s thinking are the issues that involve
control. One fear is that medications will take control of the patient’s
mind—a fear that is worse in borderline patients who have felt exploited
by prior caregivers. Closely related is the fear that by taking medications,
the borderline patient will feel too controlled by the therapist. A patient
TABLE 6–2. Dichotomous attitudes of borderline personality

disorder patients that affect medication use
POSITIVE ATTRIBUTIONS NEGATIVE ATTRIBUTIONS
About medications I’m ill; meds are needed. Meds are irrelevant.
Meds reduce pain. Meds control mind.
Meds can cure. Meds are addictive,
cause disability.
About prescriber He or she has medical M.D.’s are only
training. interested in illness.
He or she wants to He or she is not
alleviate suffering. interested in me.
He or she will do He or she thinks I have a
everything possible. chronic illness.
I can depend on him or He or she wants to
her. control me.
who is very distrustful may passively be noncompliant or even deliberately

store the medications for possible overdosing. Because noncompliance
and overdosing may occur as often as half the time (Waldinger and Frank
1989a), it is very important that psychopharmacologists initiate medica-
tions cautiously, actively inquiring about how a suspicious or quiet patient
feels about these issues, and being particularly respectful of the patient’s
hesitations or concerns. Under these circumstances, it is particularly im-
portant for a prescribing psychiatrist who is not a patient’s therapist to
have close communication with others who know the patient better.
Transference-Countertransference Issues
Even though psychopharmacologists often try to define a quite narrow
and limited role for themselves as a way to maintain a cool, professional re-
lationship, they too are vulnerable to the same intense countertransfer-
ence responses to borderline patients (Table 6–3) that psychotherapists
are familiar with: being overinvolved in alleviating patients’ pain (i.e., in
“rescuing”) or being overly frustrated by patients’ resistances (i.e., becom-
ing angry). Rescuing is the more common hazard, induced by the “doc-
torly” role of psychopharmacologists. In wanting to alleviate suffering, they
become objects of idealizing transferences. That idealization can further
encourage their wish to be helpful, which further encourages idealization,
which encourages more ambitious, special efforts to help, and so on.
TABLE 6–3. Three common transference-countertransference

enactments
1. Solicitous attention evokes increased dysfunction (including
noncompliance) and exacerbates symptom complaints.
2. Wary, worried prescribing evokes hostile, secretive use.
3. Medicalization (i.e., telling a patient that he or she has a biological
or brain disease) encourages a lack of sense of agency.
Often impelled by a focus on treating the borderline patients’ depres-

sive symptoms, psychopharmacologists may embark on extensive and
heroic searches for curative changes. The countertransference issue is
that such searches overlook how all these prescriptive activities enact the
borderline patient’s transference wish to retain the physician’s caregiving
attention and to believe that creative changes can occur without the pa-
tient changing himself or herself. Such physicians may respond to the
perception of the patient’s needing them by assuming an increasing, and
increasingly inappropriate, role in the patient’s life within—or even out-
side—treatment.
Vignette
A 35-year-old BPD patient gained 75 pounds while taking lithium plus di-
valproex (Depakote), amitriptyline, perphenazine (Trilafon), and fluox-
etine (Prozac). She showed little evidence of improvement, but the
patient was grateful for her psychopharmacologist’s earnest, kindly, re-
sponsive care. Her mother sought out the physician to complain that as a
result of the medications her daughter was increasingly dazed, somnolent,
and short of breath. The psychiatrist recognized that this could be due to
medications but did not believe that he could discontinue any of them
without risking the patient’s increased suicidality. Moreover, the patient al-
ways protested efforts to diminish her regimen. On her way to the ensuing
appointment, the patient fell, was too weak to move, and was eventually
taken to an emergency department. She died of pulmonary emboli.
Another common transference-countertransference enactment oc-

curring around medications arises when prescribing psychiatrists are too
wary of borderline patients. Undue wariness arises when the prescribing
psychiatrist’s countertransference involves viewing borderline patients as
primarily deceitful, litigious, and angry. Of course, borderline patients
can be any of these, but such qualities are far less likely to become signif-
icant problems when clinicians are aware of such apprehensions but are
discerning in assessing them. Wariness leads to overly cautious prescrib-
ing and inadequate efforts to build an alliance. Borderline patients are
sensitive to what they perceive as withholding, dislike, or distrust, and

their resentment sets the stage for noncompliance or misuse.
Psychiatrists who devalue psychosocial therapies can conceptualize
BPD as either being biological (and amenable to medication) or being
untreatable. This thinking can create a split with other treaters and ag-
gravate the split within the borderline patient’s mind about his or her ac-
countability (i.e., the medicalization can cause patients who feel overly
responsible for symptoms to feel that they are not at all responsible; see
Bolton and Gunderson 1996). Overmedicalizing a borderline patient’s
symptoms represents another of the common countertransference issues
(see Table 6–3). Medications easily multiply, side effects occur, and the
prescribing psychiatrist can respond either by mounting more heroic ef-
forts (e.g., Amytal [amobarbital] interviews) or by coming to dread the
patient’s next appointment. When the prescribing psychiatrist’s increas-
ingly desperate efforts to help fail, he or she may try to withdraw too
quickly and without explanation or referral.
Contraindications and Discontinuance

When a borderline patient is being prescribed four or more psychotropic
medications, it often signifies the absence of identifiable effectiveness.
With the growth of augmentation strategies, this conclusion now has more
exceptions, but multiple medications should always be a concern when
working with BPD patients. The clinician must be cautious about changing
medications until effects are clarified. For psychiatrists without special ex-
pertise, the prescription of more than four medications or the absence of
clear benefits after several trials warrants seeking expert consultation. Such
consultations are almost always desirable if a patient breaks through the in-
hibiting and muting effects of his or her psychopharmacological blanket
with a relapse requiring a higher, more intensive level of care. The out-
come of such consultations, when a borderline patient has already been
taking a heavy medication regimen for a sufficient duration, is typically
that the consultant advises a fresh start with trials of single medications.
Remarkably, despite the wide agreement that medications are adjunc-
tive and that they frequently offer only modest and uncertain benefits, no
guidelines, or even literature, are available regarding when to conclude
that use is contraindicated or should be discontinued. The exception
here is the finding by Soloff et al. (1993) that the value for neuroleptics
wanes rapidly, and they can therefore be discontinued after 6 weeks with-
out harm. In effect, however, the lack of advice or discussion means that,
in the current era, a patient who receives a borderline diagnosis will be
given medications. And, once medications are started, given their posi-
tive transferential significance (i.e., their role as transitional objects that

connect them with a powerful caregiver), it can be very difficult to dis-
continue the medications, even if they are not helpful (Adelman 1985).
In practice, the only times that exceptions arise are if the patient ex-
presses a fairly sustained or vehement wish to discontinue.
It is not uncommon to diminish without problems the number of
maintenance medications that borderline patients are taking. At the
least, efforts to diminish maintenance medications offer a useful way to
clarify the relative advantages of continuation versus discontinuation.
Vignette
A 26-year-old single woman had started taking paroxetine (Paxil) during
her initial hospitalization 4 years previously. After about 2 years, during
which she had resumed work and tapered her other outpatient care to on-
going group therapy, the dose of 40 mg/day was reduced to 20 mg/day.
She was stable on this dose, but she was increasingly disturbed by how her
taking paroxetine was used by her family as a reason to see her as weak
and to question or discount her judgments. Indeed, she had reason to
think her emotionality and her readiness to discuss problems rendered
her the healthiest member of the family. She wondered, “Could I be nor-
mal?” She wanted to discontinue the medication. With her physician’s
agreement, the dose was lowered to 10 mg/day. She felt more, including
more readily becoming tearful, but her feelings seemed to be appropri-
ately responsive to circumstances. After 2 months, her dose was reduced
from 10 mg to none. In about 10–14 days, her “sadness increased.” She be-
gan to feel empty. Her life seemed to lack substance or value. Feeling des-
perate and like a failure, she renewed a dose of 10 mg/day and began to
notice a change in 4 days. Within a week, she had stopped crying without
good reason, and her work again was a source of personal satisfaction.
It is possible to taper off a patient’s medications altogether, but in prac-

tice, tapering is rarely prompted by prescribing psychiatrists. A significant
number of borderline patients can, however, get by comfortably using as-
needed hypnotics or anxiolytics when they have progressed to what is
called phase 3 issues (Chapter 12). It is not clear to us that these medica-
tions are needed, but they clearly play a reassuring role. Sometimes, bor-
derline patients can get well enough or are suspicious enough about
whether medications are helpful that they will purposely stop taking
them. Borderline patients who develop new coping mechanisms, forgo re-
gressive flights, and/or have otherwise gained ego strength, resilience, or
maturity may exchange these traits for the ego-buttressing functions that
medications have otherwise served. Notably, it is more often the patient
than the prescribing psychiatrist who has the courage or optimism about
his or her ability to live with less medication. Less medication allows a pa-
tient to feel more and makes him or her feel more like a normal person.
Often, the issues of discontinuing medication are confounded by a

history of Axis I diagnoses about which the current psychiatrist has only
secondhand knowledge. If the patient was prescribed the medications for
an alleged Axis I disorder, a psychiatrist is often hesitant to discontinue
them even if there is good reason to doubt their value. Adding to the
doubts seeded by an unclear history and the current absence of an Axis I
condition is the fact that Axis I diagnoses in a managed care environ-
ment, where short evaluations and short-term treatments are mandated,
are often made for reasons of expedience (Zimmerman and Mattia
1999). In the absence of empirically based guidelines, psychiatrists must
balance what patients want with their own judgments about what the risks
and benefits of medications are. In the 1980s, the risk with borderline pa-
tients was that the benefits of medications would be underestimated; now
the risk is that their benefits will be overestimated.
Summary
The routine and long-term prescription of multiple psychoactive medica-
tions that are usually of modest and sometimes of uncertain benefit is a re-
ality of current psychiatric practice for borderline patients. Unfortunately,
polypharmacy is sometimes used as a marker of one’s sophistication with
managing such patients. This overvaluation can exist despite the patient’s
likelihood of misuse and potential for long-term dependency. Although it
is not unethical to treat BPD without medications, the widespread impres-
sion about their likely value renders it unwise to treat the disorder without
assessing whether patients can benefit from medications. Once medica-
tions are initiated, it is always wise to consider—and then reconsider—
whether the expected benefits are actually being derived. The evaluation
of medication effectiveness remains tied to the subjective responses of
both patients and clinicians. For this reason, skilled psychopharmacology
requires a psychotherapist’s appreciation of the meanings attached to the
pills as well as those attached to the prescriber. The prescribing psychia-
trist is subject to the same transference-countertransference problems
that beset psychotherapists. A pragmatic, empirical approach, consulta-
tions, cautious optimism, and actively engaging the borderline patient as
a co-investigator set the stage for meaningful trials—and reduced risks.
References
Adelman SA: Pills as transitional objects: a dynamic understanding of the
use of medication in psychotherapy. Psychiatry 48:246–253, 1985
Akiskal HS: Subaffective disorders: dysthymic, cyclothymic, and bipolar II
46, 1981
American Psychiatric Association: Guidelines for Consultative, Supervi-

sory, or Collaborative Relationships With Nonmedical Therapists.
Washington, DC, American Psychiatric Association, 1980
American Psychiatric Association: Practice guideline for the treatment of
patients with borderline personality disorder. Am J Psychiatry
158(suppl):1–52, 2001
Bender DS, Dolan RT, Skodol AE, et al: Treatment utilization by patients
with personality disorders. Am J Psychiatry 158:295–302, 2001
Bolton S, Gunderson JG: Distinguishing borderline personality disorder
from bipolar disorder: differential diagnosis and implications. Am J
Psychiatry 153:1202–1207, 1996
Brinkley J, Zeitman S, Friedel R, et al: Low-dose neuroleptic regimens in
the treatment of borderline patients. Arch Gen Psychiatry 36:319–
326, 1979
Cole JO, Salomon M, Gunderson J, et al: Drug therapy in borderline pa-
tients. Compr Psychiatry 25:249–254, 1984
Cowdry RW, Gardner DL: Pharmacotherapy of borderline personality
disorder: alprazolam, carbamazepine, trifluoperazine, and tranyl-
cypromine. Arch Gen Psychiatry 45:111–119, 1988
Gabbard GO: Combining medication with psychotherapy in the treat-
ment of personality disorders, in Psychotherapy for Personality Dis-
orders. Edited by Gunderson JG, Gabbard GO (Review of Psychiatry
Series, Vol 19; Oldham JM, Riba MB, series eds). Washington, DC,
American Psychiatric Press, 2000, pp 65–94
Goldberg SC, Schulz SC, Schulz PM, et al: Borderline and schizotypal
personality disorders treated with low-dose thiothixene versus pla-
cebo. Arch Gen Psychiatry 43:680–686, 1986
Klein D: Psychopharmacology and the borderline patients, in Borderline
1975, pp 75–92
line disorders, in Borderline Personality Disorders: The Concept, the
Syndrome, the Patient. Edited by Hartocollis P. New York, Interna-
tional Universities Press, 1977, pp 365–384
Koenigsberg HW: The combination of psychotherapy and pharmaco-
therapy in the treatment of borderline patients. J Psychother Pract
Res 3:93–107, 1994
Koenigsberg HW: Integrating psychotherapy and pharmacotherapy in
the treatment of borderline personality disorder. In Session: Psycho-
therapy in Practice 3(2):39–56, 1997
Kramer P: Listening to Prozac. New York, Viking Penguin, 1993
Pope HG Jr, Jonas JM, Hudson JI, et al: The validity of DSM-III borderline
personality disorder: a phenomenologic, family history, treatment re-
sponse, and long-term follow-up study. Arch Gen Psychiatry 40:23–
30, 1983
Skodol AE, Buckley P, Charles E: Is there a characteristic pattern to the
treatment history of clinic outpatients with borderline personality?
J Nerv Ment Dis 171:405–410, 1983
Soloff P: Pharmacotherapy of borderline disorders. Compr Psychiatry
22:535–543, 1981
Soloff PH, George A, Nathan RS, et al: Progress in pharmacotherapy of
borderline disorders: a double-blind study of amitriptyline, haloperi-
dol, and placebo. Arch Gen Psychiatry 43:691–697, 1986
Soloff PH, George A, Nathan S, et al: Amitriptyline vs haloperidol in bor-
derlines: final outcomes and predictors of response. J Clin Psycho-
pharmacol 9:238–246, 1989
Soloff PH, Cornelius J, George A, et al: Efficacy of phenelzine and halo-
peridol in borderline personality disorder. Arch Gen Psychiatry 50:
377–385, 1993
North Am 2:577–594, 1979
Waldinger RJ, Frank AF: Clinicians’ experiences in combining medica-
tion and psychotherapy in the treatment of borderline patients.
Hosp Community Psychiatry 40:712–718, 1989a
Waldinger RJ, Frank AF: Transference and the vicissitudes of medication
use by borderline patients. Psychiatry 52:416–427, 1989b
Zanarini MC, Frankenburg FR, Gunderson JG: Pharmacotherapy of bor-
derline outpatients. Compr Psychiatry 29:372–378, 1988
Zanarini MC, Frankenburg FR, Hennen J, et al: Mental health utilization
by borderline personality disorder patients and Axis II comparison
subjects followed prospectively for six years. J Clin Psychiatry 65:28–
36, 2004
Zimmerman M, Mattia JI: Differences between clinical and research prac-
tices in diagnosing borderline personality disorder. Am J Psychiatry
156:1570–1574, 1999
Chapter 7
PHARMACOTHERAPY
Selection of Medications
SINCE THE 1980s, pharmacological treatment in the management of

borderline personality disorder (BPD) has shifted from being an option
to an almost obligatory consideration. Studies have shown that patients
with BPD are frequently exposed to multiple psychiatric medications;
however, the evidence supporting this approach is clearly absent (Bender
et al. 2001; Zanarini et al. 2004a). As discussed in Chapter 6, the selection
of medications for patients with BPD can be difficult because of their
high levels of symptoms and distress and because of the attitudes, mean-
ings, and attributions given to prescribed medications.
In this chapter, we approach the pharmacological management of
BPD with a model targeting four symptom clusters:
1. Anger/impulsivity
2. Affective dyscontrol
3. Cognitive-perceptual dyscontrol
4. Anxiety
For each of these clinical presentations, we use the symptom-targeted

model and illustrate this with clinical vignettes. The available evidence is
summarized in Tables 7–1 through 7–4. The chapter concludes with a
caution about our current state of knowledge and a look ahead toward fu-
ture directions.
Pharmacotherapy Models
Personality disorders are best conceptualized as arising from genetic vul-
nerabilities that become manifest through interactions with exposure to
157
psychological and social factors during the person’s life course (Paris
2001). Because at least half of the total variance of personality traits is ex-
plained by genetic factors, we can anticipate that neurotransmitter sys-
tems and function will be closely tied to these personality traits (Coccaro
and Kavoussi 2001; Skodol et al. 2002). Therefore, we would predict that
some personality traits will be modifiable in response to pharmacother-
apy interventions. The strongest support of this tenet is the evidence that
central serotonin function is inversely related to impulsivity or aggressive-
ness. This association has been documented with many different re-
search methods and across many different populations (Coccaro and
Kavoussi 2001). However, beyond this one example, specific biological
functions have not been connected to observable personality traits. Cur-
rent pharmacotherapy practices for patients with BPD remain a clinical
art that at best can be based on a reasonable conceptual model.
As noted earlier, we use a symptom-targeted model, but two alterna-
tive models are worth noting:
1. A diagnostically focused model would treat BPD as a distinct disorder

that might respond to distinct medications. The model overlooks the
obvious overlap of BPD with other Axis II disorders and its inherent
heterogeneity. In any event, no BPD-specific medication has been
identified.
2. Another model, the subsyndromal model, asserts that BPD is best under-
stood as a subthreshold variant of Axis I disorders, and pharmacother-
apy should be targeted at the Axis I disorder (as reviewed in Chapter
2). However, Dolan-Sewell et al. (2001) reported that there were no
discrete associations between specific Axis I disorders and specific Axis
II disorders. Medications that are standard treatment for BPD’s pat-
tern of comorbidity are similar to that for other personality disorders
(Dolan-Sewell et al. 2001). Moreover, BPD fails to respond uniformly
or dramatically to any of the treatments for its major Axis I affiliates.
The Symptom-Targeted Model

The symptom-targeted model, developed by Paul Soloff (1998, 2000), was
adopted as a central aspect of the management of BPD in the American
Psychiatric Association (2001) guidelines. Soloff proposed that the targets
of pharmacotherapy could be based on three dimensions: cognitive-per-
ceptual, affective, and impulsive-behavioral—each of which was thought
to be regulated by the actions of specific neurotransmitter systems.
Before describing the modified symptom-targeted model, we propose
that the model’s limitations should be noted:
Pharmacotherapy: Selection of Medications ❘ 159
1. It is unclear that the regulation of cognitive-perceptual thinking, af-

fect, and impulses is specific to a single neurotransmitter that can be
easily identified as a target for pharmacotherapy. Medications typi-
cally have nonspecific effects in patients with personality disorders
(Kolla et al. 2008; Links et al. 1998). For example, in patients with
BPD, antipsychotic effectiveness is not restricted to improvement of
psychotic-like symptoms but also affects anxiety, obsessive-compulsive
symptoms, affective symptoms, and suicidal behavior.
2. The symptom-targeted approach often chooses targets that are only
broadly characterized and without good means of measurement.
3. The symptom-targeted approach in the algorithms as outlined by So-
loff and adapted by the American Psychiatric Association guidelines
increases patients’ exposure to untested interventions and fosters the
use of medications without adequate scientific justification (Paris
2002; Zanarini et al. 2004a). Many of the second, third, and fourth
steps in the algorithms are based only on evidence from case reports
or case series.
In what follows, we review evidence and use case examples to illustrate

the use of medications in their application to each of four symptom tar-
gets: anger/impulsivity, affective dyscontrol, cognitive-perceptual dys-
control, and anxiety.
Anger/Impulsivity
Impulsivity, expressed in actions such as recklessness, bingeing, promiscu-
ity, and impulsive suicide attempts, is a significant problem in patients
with BPD that is sometimes amenable to pharmacological treatment. An-
ger is linked with impulsivity for three reasons: 1) as noted, these symp-
toms are thought to have a common neurobiological base (Coccaro and
Kavoussi 2001); 2) anger often results in impulsive behavior; and 3) many
measures used as outcomes in previous drug studies connect anger and
impulsive dyscontrol (e.g., Coccaro and Kavoussi 1997; Nickel et al. 2004).
Vignette: Anger/Impulsivity
Ms. I appeared in the psychiatric emergency department following her
medical treatment for an overdose of a variety of medications. Ms. I was a
young-looking 48-year-old divorced woman who lived with her adult
daughter. She readily acknowledged that the overdose was impulsively
triggered by an argument with her daughter. Enraged, she went into the
bathroom and ingested a variety of pills from her medicine cabinet. Her
daughter was a witness to this and called the ambulance. The patient’s
chart documented numerous similar presentations and warned that Ms. I
was diagnosed with BPD. When we reached the point of deciding how
best to help Ms. I, she explicitly told me, “Doctor, I am less impulsive when
I take medication for my borderline personality disorder.” Giving pills to
a woman who impulsively overdoses; did this make any sense?
When deciding on medication for Ms. I, research suggests that mood

stabilizers are particularly useful against anger and impulsivity in BPD.
One randomized controlled trial offered evidence that lithium may be
useful against anger and suicidal symptoms (Links et al. 1990). In other
studies, carbamazepine reduced symptoms of impulsivity, alprazolam sig-
nificantly increased impulsivity (Cowdry and Gardner 1988; Gardner and
Cowdry 1986), and divalproex showed effectiveness for impulsivity (Hol-
lander 1999; Hollander et al. 2003). Some studies of valproate (Stein et
al. 1995), lamotrigine (Pinto and Akiskal 1998), and carbamazepine (De
La Fuenta and Lotstra 1994) have suggested their effectiveness; however,
some of the strongest evidence comes from studies of lamotrigine (Tritt
et al. 2005) and topiramate (Loew et al. 2006; Nickel et al. 2004).
The evidence on topiramate’s effects on anger and impulsivity is worth
reviewing. The purpose of the study by Nickel et al. (2004) was to deter-
mine the efficacy and safety of topiramate for the treatment of aggression
in women with BPD. The sample included 21 women ages 20–35 years who
met criteria for BPD. In this randomized double-blind, placebo-controlled
study, the maximum dose of topiramate was 250 mg/day by the sixth week
of the 8-week study. The main outcome was the change on a questionnaire
capturing various aspects of anger. The subjects receiving topiramate had
significantly decreased scores on the subscales of subjective anger, readi-
ness to act, and tendency for anger to be expressed outwardly, and an in-
crease in the sense of anger control. This study found that topiramate may
increase the patients’ control of their anger and modify an important part
of the impulsive/aggressive symptoms of BPD. The medication was well
tolerated, and the main side effect was weight loss, which is often seen as
an advantage by women with this disorder. The authors subsequently re-
peated the trial in men with BPD and replicated the above findings
(Nickel et al. 2005).
The same group of investigators completed a second study investigat-
ing the efficacy of topiramate for a wider variety of psychopathological
symptoms of BPD (Loew et al. 2006). This study reported that topiramate
was efficacious for various aspects of BPD, including anxiety, hostility, in-
terpersonal sensitivity, and the “aggressive-expansive traits” of the disor-
der. Topiramate was not helpful for the depressive and psychotic features
of BPD.
Overall, these findings are consistent with other studies of mood sta-
bilizers in patients with BPD in showing that these medications are quite
effective in modifying the anger, hostility, and impulsivity that character-

ize the disorder. However, these medications are less effective in modify-
ing the depressive features or stabilizing the mood of patients with BPD.
Ms. I was reluctant to consider a mood stabilizer because of the side ef-
fects but wanted to be prescribed a selective serotonin reuptake inhibitor
(SSRI), which she had taken previously. Some evidence supported Ms. I’s
choice of medication. The most important study addressing the useful-
ness of SSRIs for impulsive aggressiveness was completed by Coccaro and
Kavoussi. Coccaro and Kavoussi (1997) studied the effects of fluoxetine
in nondepressed subjects with personality disorders characterized by hav-
ing impulsive aggressiveness. Not all of the subjects had diagnoses of
BPD, but all showed evidence of impulsive aggressiveness. Patients were
randomly assigned to receive fluoxetine, 20–60 mg/day, or placebo in a
randomized double-blind, placebo-controlled design. The authors re-
ported a significant reduction in verbal aggression and aggression
against objects for subjects taking fluoxetine compared with those receiv-
ing placebo. The authors found that changes in impulsive aggression
were independent of changes in depression, anxiety, and alcohol use. In
summary, SSRIs appear to have specific effects on impulsive aggressive-
ness and anger. In addition, the improvement in impulsive aggressiveness
looks to be independent of changes in depression and anxiety.
In addition to mood stabilizers and SSRIs, atypical antipsychotics have
been proved effective in reducing anger and impulsive aggressiveness
(Table 7–1). In a meta-analysis of randomized controlled trials of mood
stabilizers, antipsychotics, and antidepressants (Mercer et al. 2005), the
pattern of findings was similar; all three classes of medication were more
effective for anger and impulsivity than for depressive symptoms in pa-
tients with BPD. Agents should then be chosen on the basis of safety in
overdose and side-effect profile. As indicated, Ms. I did not want to ex-
pose herself to the side effects related to the mood stabilizers. She had
previously tolerated an SSRI without side effects. From the emergency de-
partment, she was prescribed a small amount of an SSRI. As part of the
goal for follow-up, we agreed to monitor whether the medication de-
creased her anger, impulsivity, and recurrent suicidal behavior. If Ms. I
failed to benefit from the first SSRI, I likely would have tried a second
SSRI and then switched to mood stabilizers if she continued to have prob-
lems with impulsivity despite receiving adequate doses of SSRIs. To ob-
tain the improvement in impulsivity in some patients, the upper end of
the dose range (e.g., 60 mg of fluoxetine) may be necessary. Ms. I was
warned that SSRIs can trigger agitation and antidepressant-emergent sui-
cidality, especially over the first weeks of treatment. The patient was ad-
vised to seek immediate help if these side effects occurred.
TABLE 7–1. Summary of evidence for anger/impulsivity

MEDICATION ANGER/IMPULSIVITY
Mood stabilizers
Carbamazepine ++
Valproate ++
Topiramate ++
Lamotrigine +
Antidepressants
Selective serotonin reuptake inhibitors ++
Tricyclics −
Monoamine oxidase inhibitors +
Dual-action ?
Antipsychotics
Typical and atypical +
Benzodiazepines −
Note. The information in this table should be considered tentative and is based
on the authors’ synthesis of the literature and clinical experience.
++ =clear improvement; + = modest improvement; −= some worsening; ?= benefits
unknown.
Approaches that require further study for anger and impulsivity are
the use of naltrexone in repetitive self-harm behavior (Links et al. 1998)
and psychostimulants in impulsive borderline patients with residual adult
symptoms of attention-deficit/hyperactivity disorder (Soloff 1998).
Affective Dyscontrol
Affective instability of borderline patients includes dysphoric moods and
other indications of dysregulation of mood, such as volatility and highly
variable mood.
Vignette: Affective Dyscontrol

Ms. J, a 28-year-old woman, was seen in an emergency department after
taking a small overdose of over-the-counter sleeping pills. This was pre-
cipitated by an argument with her ex-boyfriend, after which she immedi-
ately came for help. Ms. J endorsed that her moods were “all over the
map” and insisted “I never know what emotion is going to hit me next.
That’s what makes me suicidal.” The patient had no history of manic ep-
isodes, and during her only period of significant depression in her early
20s, she had never filled the prescription for antidepressants that she had
been given. Ms. J strongly endorsed having affective instability and most
other criteria for BPD. Before leaving the emergency department, she in-
sisted on having medications to stop her erratic emotions.
Ms. J was adamant that her erratic emotions were putting her at risk,
and when this observation was validated, she listened more calmly to
treatment options. Research evidence regarding medications for affec-
tive instability is summarized in Table 7–2. Although little research has
targeted this feature of BPD specifically, Rinne et al. (2002) found that
fluvoxamine significantly improved rapid mood shifts in women with
BPD. Mood stabilizers in these patients, as discussed earlier, seem to have
little effect on ultracircadian or rapidly shifting moods. Antipsychotic
medications have a broad-spectrum action that seems to be helpful dur-
ing short-term crisis management of extremely emotional states. Al-
though further research on affective dyscontrol is needed, some patients
with BPD describe a dampening down of their emotions while taking
SSRI medication.
Depressive symptoms are an important part of the affective instability
seen in BPD patients. The next case vignette discusses management of a
patient whose dysphoria was more of a problem than was variability of
mood.
Vignette: Episode of Depression

Ms. K, a 24-year-old college student with BPD, had regularly attended her
weekly outpatient appointments. During one of these appointments, she
presented as highly anxious and desperate for my help. Her grades were
falling, and she was missing most of her classes. She complained of in-
creasing depression and irritability. She insisted that her depression was
affecting her appetite, energy, motivation, and sex drive, but she was most
urgently concerned about her inability to do her schoolwork. Ms. K had
always been reluctant to take medications, but now she was pleading for
something to help her cope.
The efficacy of antidepressants for depression in patients with BPD

has received significant research attention. The oldest class of antidepres-
sants (i.e., tricyclics) has not shown significant effectiveness for depres-
sion (Links et al. 1990; Soloff et al. 1986b) and might even worsen
symptoms in some patients (Soloff et al. 1986a).
Monoamine oxidase inhibitors (MAOIs) have shown limited benefits
in the treatment of BPD. Soloff et al. (1993) did not report much supe-
riority of phenelzine over haloperidol and placebo with regard to depres-
sive symptoms but did find significant effects on anger and hostility
TABLE 7–2. Summary of evidence for affective dyscontrol/

depression
AFFECTIVE DYSCONTROL/
MEDICATION DEPRESSION
Mood stabilizers
Carbamazepine +/−
Valproate +/−
Topiramate +/−
Lamotrigine +
Antidepressants
Selective serotonin reuptake inhibitors ++
Tricyclics +/−
Monoamine oxidase inhibitors +
Dual-action +
Antipsychotics
Benzodiazepines +/−
++ =clear improvement; + =modest improvement; +/−= variable improvement or
worsening.
compared with placebo. MAOIs may be of benefit to patients with atypi-

cal depression, meaning those with reactive moods and atypical symp-
toms such as increased appetite (Cowdry and Gardner 1988; Parsons et
al. 1989).
Symptomatic improvements of depression resulting from SSRI medi-
cation in BPD patients seem to be independent of a comorbid diagnosis
of major depression, the presence of an affective borderline subtype, a
history or family history of mood disorder, or the level of current de-
pressed mood (Soloff 2000).
Although the need to treat a coexisting major depression in patients
with BPD is common, clinicians should approach treatment carefully. The
depression should be documented as persisting for at least 2–3 weeks, and
clinicians should recognize that these symptoms often resolve without
medication. Depression is often a result of being admitted to a hospital
(Gunderson 1984; Siever et al. 1985), but as noted elsewhere in the book,
depression is a predictable result of any intervention from which a border-

line patient feels held. Major depression in the context of BPD may repre-
sent a different syndrome from major depression alone, and evidence
suggests that future episodes of depression are more strongly predicted by
the level of borderline psychopathology than by the history of depressive
illness (Gunderson et al. 2004; Links et al. 1995). To prevent future epi-
sodes of depression in BPD patients, the focus should be on treating the
borderline disorder. In treating a current episode of major depression in
patients with BPD, a less complete and more delayed response to antide-
pressant medication can be expected (Newton-Howes et al. 2006; Soloff
1993). Similarly, depressed patients with BPD may respond to electrocon-
vulsive therapy, but the response tends to be less consistent than for de-
pressed patients without BPD (Feske et al. 2004; Zimmerman et al. 1986).
After these complications have been considered, a trial of antidepres-
sant medication appears to be warranted for those borderline patients
with persistent depressive symptoms. The American Psychiatric Associa-
tion (2001) guidelines suggest that the first line of treatment should be
an SSRI. These agents are relatively safe and have few side effects. There-
fore, Soloff (1998) recommends that a second SSRI be tried if the first
SSRI is ineffective. MAOIs remain an option for select patients who can
comply with a diet regimen and can be vigilant about side effects. Studies
with dual-action agents, such as venlafaxine, have reported effectiveness
in an open trial with BPD patients, but more study is needed (Markovitz
and Wagner 1995). In addition to antidepressant medication, mood sta-
bilizers appear to be effective for depressive symptoms in BPD patients
(Mercer et al. 2005).
As a follow-up to the above vignette, Ms. K was prescribed a low dose
of citalopram; her attendance at school improved, and she successfully
completed her year. If Ms. K’s symptoms had not responded after two tri-
als of SSRIs, lamotrigine would have been offered. Because of lamotri-
gine’s effectiveness for depression in patients with bipolar disorder type
II (Bhagwagar and Goodwin 2005), it may be one of the most promising
agents to try in depressed BPD patients. However, patients must be able
to tolerate the slow titration needed to lessen the risk of skin rash.
Cognitive-Perceptual Dyscontrol
Cognitive-perceptual symptoms that are common in borderline patients
include depersonalization or derealization, illusions, ideas of reference,
and brief paranoid states. Psychopharmacological trials have not yet tar-
geted or assessed the disturbed cognitive schemas that are thought to
control aspects of borderline psychopathology in major cognitive theo-
ries by Young et al. (2003) or Beck et al. (2004) (see Chapter 11).
Vignette: Severe Dissociative Symptoms

Ms. L, a 22-year-old single woman, always appeared for her appointments
in Gothic attire with long sleeves that hid her frequent episodes of self-
harm. For Ms. L, the self-harm behavior was less of a concern than the
frightening feelings that accompanied these episodes; she would feel to-
tally disconnected from her surroundings, would lose all sense of time,
and would become aware again only when she saw the blood dripping
from her new wounds. Ms. L. hoped to find something that would prevent
these terrifying episodes.
Antipsychotic medication might be considered to alleviate Ms. L’s

cognitive-perceptual symptoms (Table 7–3). Several randomized con-
trolled trials of traditional antipsychotic medications in borderline pa-
tients have supported the effectiveness of low-dose antipsychotics for
these symptoms (Cowdry and Gardner 1988; Goldberg et al. 1986; Leone
1982; Montgomery 1987; Montgomery and Montgomery 1982; Serban
and Siegel 1984; Soloff et al. 1986b, 1993). The improvements from an-
tipsychotics also have extended to anxiety symptoms, obsessive-compul-
sive symptoms, affective symptoms, and suicidal behaviors.
Studies have suggested a role for atypical antipsychotics. Several case
series involving borderline patients have been completed. Atypicals stud-
ied in these series were risperidone (Khouzam and Donnelly 1997; Szig-
ethy and Schulz 1997), clozapine (Benedetti et al. 1995; Chengappa et al.
1995, 1999; Frankenburg and Zanarini 1993), and olanzapine (Schulz et
al. 1999), as well as randomized controlled trials of olanzapine and ari-
piprazole (Bogenschutz and Nurnberg 2004; Nickel et al. 2006; Zanarini
and Frankenburg 2001; Zanarini et al. 2004b, 2006). Only two open-label
studies of quetiapine in patients with BPD have been published at the
time of this writing (Adityanjee and Schulz 2002; Perrella et al. 2007).
Overall, atypical antipsychotics seem to help alleviate cognitive-percep-
tual symptoms. However, as with traditional antipsychotics, they have a
broad-spectrum effect, including impulsive aggressiveness as noted ear-
lier. Despite frequent side effects, clozapine seems to help otherwise in-
tractable BPD patients with Axis I psychotic symptoms by also reducing
agitation and other troubling symptoms, including severe self-mutilation.
Low doses of antipsychotics, including atypical agents, have been
shown to be effective as nonspecific tranquilizers that reduce, in the short
term, the severity of a broad range of symptoms found in acutely dis-
tressed borderline patients. Atypical agents may be effective with a lower
risk of neurological side effects, but, unfortunately, endocrine effects—
especially weight gain and/or development of diabetes and dyslipi-
demia—may limit the acceptability of olanzapine as well as risperidone
and quetiapine. Aripiprazole appears to have the advantage of not caus-
TABLE 7–3. Summary of evidence for cognitive-perceptual

dyscontrol
COGNITIVE-PERCEPTUAL
MEDICATION DYSCONTROL
Mood stabilizers
Carbamazepine ?
Valproate ?
Topiramate ?
Lamotrigine ?
Antidepressants
Selective serotonin reuptake inhibitors ?
Tricyclics ?
Monoamine oxidase inhibitors ?
Dual-action ?
Antipsychotics
Typical and atypical ++
Benzodiazepines ?
++ =clear improvement; ?= benefits unknown.
ing significant weight gain (Nickel et al. 2007). Ms. L experimented with
the use of low-dose olanzapine and was able to document the reduction
in her frightening dissociative and self-harm episodes. However, the
weight gain with olanzapine was unacceptable. Ms. L eventually decided
that quetiapine provided a better balance between the benefits and side
effects. As shown in Table 7–3, few other medications have been shown to
be effective for the borderline patient’s cognitive-perceptual symptoms.
Had Ms. L been unable to tolerate quetiapine, the next step would have
been to use a low-dose typical antipsychotic with warnings of the potential
risk for extrapyramidal side effects and tardive dyskinesia.
Anxiety
Severe anxiety is extremely common, is often very disabling, and can pro-
mote high-risk behavior in patients with BPD. The typical presentations
of anxiety can be grouped into two subtypes: 1) somatic anxiety, which is
experienced in the body (e.g., through stimulus-seeking behavior and in
panic-related and physical symptoms); and 2) psychic anxiety, with obses-

sional, avoidant, or phobic manifestations. The two types of anxiety help
guide pharmacotherapies (Soloff 1990); Table 7–4 provides a summary.
Somatic anxiety is associated with impulsivity, antisocial behavior, and
a histrionic cognitive style. When this type of anxiety coexists with behav-
ioral dyscontrol, it may be best treated with SSRIs, mood stabilizers, or
low-dose antipsychotics.
Psychic anxiety is observed in patients with a low tolerance to stimu-
lation and a high anticipation of harm. These patients have significant
histories of prior abuse. This type of anxiety may be treated with a long-
acting benzodiazepine such as clonazepam or other agents affecting the
γ-aminobutyric acid (GABA) system, such as gabapentin (Nemeroff
2003). Alprazolam, however, and perhaps other short-acting benzodiaz-
epines can disinhibit, leading to impulsive and violent actions (Gardner
and Cowdry 1985) and dependency. Patients with BPD are at significant
risk for abuse of and dependency on benzodiazepines, and when this oc-
curs, they have difficulty achieving abstinence (Vorma et al. 2005).
Vignette: Anxiety
Mr. M, a 27-year-old gay man, had BPD, comorbid posttraumatic stress
disorder, and a chronic pain syndrome. His disorders caused a variety of
anxiety symptoms, including panic attacks, social anxiety, obsessive rumi-
nations, and disabling avoidant behavior. His avoidance was so extreme
that he seldom left his bedroom, and most of his social interactions were
through his desktop computer. In an attempt to control his anxiety symp-
toms, he had tried many different medications, but the side effects repeat-
edly outweighed any benefits obtained. Mr. M was desperate to “have a
life” and find something he could tolerate that would permit more social-
ization.
Mr. M was able to tolerate gabapentin. While taking this medication,

he noted a slight but meaningful reduction in his anxiety and pain symp-
toms. If Mr. M’s symptoms had not responded to gabapentin, the next
step would have been a low-dose atypical antipsychotic.
Comorbid anxiety disorders may be more common in BPD patients
than previously recognized (Zanarini et al. 1998), and treatment of these
comorbid conditions should be a priority (Links et al. 1998). For exam-
ple, borderline patients with comorbid posttraumatic stress disorder
require integrated pharmacotherapy for the posttraumatic stress symp-
toms. Case reports have suggested that naltrexone may reduce flashbacks
and self-mutilative behaviors in borderline patients (Schmahl et al.
1999); however, a randomized controlled trial of naloxone found no sig-
nificant benefits for acute dissociative states (Philipsen et al. 2004).
TABLE 7–4. Summary of evidence for anxiety

MEDICATION ANXIETY
Mood stabilizers
Carbamazepine ?
Valproate ?
Topiramate ?
Lamotrigine ?
Antidepressants
Selective serotonin reuptake inhibitors +
Tricyclics +/–
Monoamine oxidase inhibitors + (somatic)
Dual-action ?
Antipsychotics
Benzodiazepines
Long-acting + (psychic)
+= modest improvement; +/− =variable improvement or worsening; ?= benefits
unknown.
Summary
When selecting medications for patients with BPD, several points must be
remembered. First, many of the pharmacological studies have been ham-
pered because of their lack of external validity; most studies were carried
out in volunteers or excluded patients with self-harm, suicidal behavior, and
comorbid substance abuse. As a result, it is hazardous to generalize the find-
ings to typical patients with BPD. Most of the studies were short-lived and
did not inform the use of medications over the longer term. Still lacking are
precise measures of several aspects of borderline psychopathology, such as
affective instability, identity disturbance, and severe dissociative symptoms;
research addressing these outcomes is still in the developmental stages. Fu-
ture studies will need to better inform clinicians about the risks versus ben-
efits and about the effect of medication on functional outcomes.
The current evidence related to pharmacotherapy for BPD is modest
at best (Binks et al. 2006; Nose et al. 2006), and it must be remembered
that no drug is licensed as indicated for BPD (Herpertz et al. 2007). De-
spite the limitations, conclusions can be made about pharmacotherapy
for patients with BPD. Medication is mainly an adjunct to psychothera-
peutic management; however, some early evidence suggests that combin-
ing medication and specific psychotherapy approaches may be needed in
patients with depression and comorbid BPD, particularly to improve
quality of life and functioning in addition to symptoms (Bellino et al.
2006; Kool et al. 2003). Overall, for patients with BPD, the studies suggest
that medications are mildly to moderately effective for anger and impul-
sivity and modestly effective for depression. The effectiveness of medica-
tion for affective instability, cognitive-perceptual features, and anxiety is
less proven. When choosing medication for one of the four symptom tar-
gets, the clinician should try to anticipate the expected outcome of the
intervention before starting treatment. Will the patient be expected to
show less impulsivity, less suicidal behavior, or perhaps better perfor-
mance at work as a result of the medication?
Although the chapter focuses on pharmacotherapy for aspects of
BPD, if the treating clinician is confident that an Axis I disorder is present
and leading to difficulties for the patient with BPD, then the Axis I diag-
nosis should not be dismissed or ignored. Treatment should be selected
based on the best practices for the Axis I disorder.
Finally, the patient should always be an active collaborator in selecting
the medication, as discussed in Chapter 6. The partnership will ensure
that the patient is empowered to choose, use, continue, and discontinue
medication to best meet his or her personal needs and goals for therapy.
References
Adityanjee MD, Schulz SC: Clinical use of quetiapine in disease states
other than schizophrenia. J Clin Psychiatry 63 (suppl 13):32–38, 2002
158(suppl):1–52, 2001
Beck AT, Freeman A, Davis DD, et al: Cognitive Therapy of Personality
Disorders. New York, Guilford, 2004
Bellino S, Zezza M, Rinaldi C, et al: Combined treatment of major depres-
sion in patients with borderline personality disorder: a comparison
with psychotherapy. Can J Psychiatry 51:453–460, 2006
Benedetti F, Sforzini L, Colombo C, et al: Low-dose clozapine in amelio-
rating severe self-mutilation in a patient with borderline personality
disorder. J Personal Disord 9:76–82, 1995
Bhagwagar Z, Goodwin GM: Lamotrigine in the treatment of bipolar dis-

order. Expert Opin Pharmacother 6:1401–1408, 2005
Binks CA, Fenton M, McCarthy L, et al: Pharmacological interventions
for people with borderline personality disorder. Cochrane Database
Syst Rev (1):CD005653, 2006
Bogenschutz MP, Nurnberg HG: Olanzapine versus placebo in the treat-
ment of borderline personality disorder. J Clin Psychiatry 65(1):104–
109, 2004
Chengappa KNR, Baker RW, Sirri C: The successful use of clozapine in
ameliorating severe self-mutilation in a patient with borderline per-
sonality disorder. J Personal Disord 9:76–82, 1995
Chengappa KNR, Ebeling T, Kang JS, et al: Clozapine reduces severe self-
mutilation and aggression in psychotic patients with borderline per-
sonality disorder. J Clin Psychiatry 60:477–484, 1999
Coccaro EF, Kavoussi RJ: Fluoxetine and impulsive aggressive behavior in
personality-disordered subjects. Arch Gen Psychiatry 54:1081–1088,
1997
Coccaro EF, Kavoussi RJ: Biological and pharmacological aspects of bor-
derline personality disorder. Hosp Community Psychiatry 42:1029–
1033, 2001
Cowdry RW, Gardner DL: Pharmacotherapy of borderline personality
disorder: alprazolam, carbamazepine, trifluoperazine, and tranyl-
cypromine. Arch Gen Psychiatry 45:111–119, 1988
De La Fuenta JM, Lotstra F: A trial of carbamazepine in borderline per-
sonality disorder. Eur Neuropsychopharmacol 4:479–486, 1994
Dolan-Sewell RT, Krueger RF, Shea MT: Co-occurrence with syndrome
disorders, in Handbook of Personality Disorders: Theory, Research
84–104
Feske U, Mulsant BH, Pilkonis PA, et al: Clinical outcome of patients with
clinical depression and comorbid borderline personality disorder.
Am J Psychiatry 161:2073–2080, 2004
Frankenburg FR, Zanarini MC: Clozapine treatment of borderline pa-
tients: a preliminary study. Compr Psychiatry 34:402–405, 1993
Gardner DL, Cowdry RW: Alprazolam-induced dyscontrol in borderline
Gardner DL, Cowdry RW: Positive effects of carbamazepine on behav-
ioral dyscontrol in borderline personality disorder. Am J Psychiatry
143:519–522, 1986
Goldberg SC, Schulz SC, Schulz PM, et al: Borderline and schizotypal
personality disorders treated with low-dose thiothixene vs placebo.
Gunderson JG, Morey LC, Stout RL, et al: Major depressive disorder and
borderline personality disorder revisited: longitudinal interactions.
Herpertz SC, Zanarini M, Schulz CS, et al: World Federation of Societies
of Biological Psychiatry (WFSBP) guidelines for biological treatment
of personality disorders. World J Biol Psychiatry 8(4):212–244, 2007
Hollander E: Managing aggressive behavior in patients with obsessive-
compulsive disorder and borderline personality disorder. J Clin Psy-
chiatry 60 (suppl 15):38–44, 1999
Hollander E, Tracy KA, Swann AC, et al: Divalproex in the treatment of
impulsive aggression: efficacy in cluster B personality disorders. Neu-
ropsychopharmacology 28:1186–1197, 2003
Khouzam HR, Donnelly NJ: Remission of self-mutilation in a patient with
borderline personality during risperidone therapy. J Nerv Ment Dis
195:348–349, 1997
Kolla NJ, Eisenberg H, Links PS: Epidemiology, risk factors, and psycho-
pharmacological management of suicidal behavior in borderline
personality disorder. Arch Suicide Res 12(1):1–19, 2008
Kool S, Dekker J, Duijsens J, et al: Efficacy of combined therapy and phar-
macotherapy for depressed patients with and without personality dis-
orders. Harv Rev Psychiatry 11:133–141, 2003
Leone NF: Response of borderline patients to loxapine and chlorpro-
mazine. J Clin Psychiatry 43:148–150, 1982
Links PS, Steiner M, Boiago I, et al: Lithium therapy for borderline pa-
tients: preliminary findings. J Personal Disord 4:173–181, 1990
Links PS, Heslegrave RJ, Mitton JE, et al: Borderline psychopathology
and recurrences of clinical disorders. J Nerv Ment Dis 183:582–586,
1995
Links PS, Heslegrave R, Villella J: Psychopharmacological management
of personality disorders: an outcome focused model, in Biology of
Personality Disorders. Edited by Silk KR. Washington, DC, American
Loew TH, Nickel MK, Muehlbaucher M, et al: Topiramate treatment for
women with borderline personality disorder: a double-blind placebo-
controlled study. J Clin Psychopharmacol 26:61–66, 2006
Markovitz PJ, Wagner SC: Venlafaxine in the treatment of borderline per-
sonality disorder. Psychopharmacol Bull 31:773–777, 1995
Mercer D, Douglass A, Links PS: A meta-analysis of antidepressants and
mood stabilizers for borderline personality disorder: significant ef-
fect on mood and anger symptoms. Poster presented at the 55th an-
nual conference of the Canadian Psychiatric Association, Vancouver,
BC, November 4, 2005
Montgomery SA: The psychopharmacology of borderline personality dis-
order. Acta Psychiatr Belg 87:260–266, 1987
Montgomery SA, Montgomery D: Pharmacological prevention of suicidal

behavior. J Affect Disord 4:291–298, 1982
Nemeroff DB: The role of GABA in the pathophysiology and treatment of
anxiety disorders. Psychopharmacol Bull 37:133–146, 2003
Newton-Howes G, Tyrer P, Johnson T: Personality disorders and the out-
come of depression: meta-analysis of published studies. Br J Psychia-
try 188:13–20, 2006
Nickel MK, Nickel C, Mitterlehner FO, et al: Topiramate treatment of ag-
gression in female borderline personality disorder patients: a dou-
ble-blind placebo-controlled study. J Clin Psychiatry 65:1515–1519,
2004
Nickel MK, Nickel C, Kaplan P, et al: Treatment of aggression with
topiramate in male borderline patients: a double-blind placebo-
controlled study. Biol Psychiatry 57:495–499, 2005
Nickel MK, Muehlbaucher M, Nickel C, et al: Aripiprazole in the treat-
ment of patients with borderline personality disorder: a double-
blind, placebo-controlled study. Am J Psychiatry 163:833–888, 2006
Nickel MK, Lowe TH, Gil FP: Aripiprazole in treatment of borderline pa-
tients, part II: an 18-month follow-up. Psychopharmacology 191:
1023–1026, 2007
Nose M, Cipriani A, Biancosino B, et al: Efficacy of pharmacotherapy
against core traits of borderline personality disorder: meta-analysis of
randomized controlled trials. Int Clin Psychopharmacol 21:345–353,
2006
Paris J: Psychosocial adversity, in Handbook of Personality Disorders:
Theory, Research and Treatment. Edited by Livesley WJ. New York,
Guilford, 2001, pp 231–241
Paris J: Commentary on the American Psychiatric Association guidelines
for the treatment of borderline personality disorder: evidence-based
psychiatry and the quality of evidence. J Personal Disord 16:130–134,
2002
Parsons B, Quitkin FM, McGrath PJ, et al: Phenelzine, imipramine, and
placebo in borderline patients meeting criteria for atypical depres-
sion. Psychopharmacol Bull 25:524–534, 1989
Perrella C, Carrus D, Costa E, et al: Quetiapine for the treatment of bor-
derline personality disorder; an open-label study. Prog Neuropsy-
chopharmacol Biol Psychiatry 31:158–163, 2007
Philipsen A, Schmahl C, Lieb K: Naloxone in the treatment of acute dis-
sociative states in female patients with borderline personality disor-
der. Pharmacopsychiatry 37:196–199, 2004
Pinto OC, Akiskal HS: Lamotrigine as a promising approach to border-
line personality: an open case series without concurrent DSM-IV ma-
jor mood disorder. J Affect Disord 51:333–343, 1998
Rinne T, van den Brink W, Wouters L, et al: SSRI treatment of borderline

personality disorder: a randomized, placebo-controlled clinical trial
for female patients with borderline personality disorder. Am J Psychi-
atry 159:2048–2054, 2002
Schmahl C, Stiglmayr C, Bohme R, et al: Treatment of dissociative symp-
toms in borderline patients with naltrexone [in German]. Nerve-
narzt 70:262–264, 1999
Schulz SC, Camlin KL, Berry SA, et al: Olanzapine safety and efficacy in
patients with borderline personality disorder and comorbid dys-
thymia. Biol Psychiatry 46(10):1429–1435, 1999
Serban G, Siegel S: Response of borderline and schizotypal patients to
small doses of thiothixene and haloperidol. Am J Psychiatry 141:
1455–1458, 1984
Siever LJ, Klar H, Coccaro E: Psychobiologic substrates of personality, in
Biologic Response Styles: Clinical Implications. Edited by Klar H,
Siever LJ. Washington, DC, American Psychiatric Press, 1985, pp 35–
66
Skodol AE, Siever LJ, Livesley WJ, et al: Borderline diagnosis II: biology,
genetics and clinical course. Biol Psychiatry 51:951–963, 2002
Soloff PH: What’s new in personality disorders? an update on pharmaco-
logic treatment. J Personal Disord 4:233–243, 1990
Soloff PH: Pharmacological therapies in borderline personality disorder,
in Borderline Personality Disorder: Etiology and Treatment. Edited
by Paris J. Washington, DC, American Psychiatric Press, 1993, pp
319–348
Soloff PH: Algorithm for pharmacological treatment of personality di-
mensions: symptom-specific treatments for cognitive-perceptual, af-
fective and impulsive-behavioral dysregulation. Bull Menninger Clin
62:195–214, 1998
Soloff PH: Psychopharmacology of borderline personality disorder. Psy-
chiatr Clin North Am 23:169–192, 2000
Soloff P, George A, Nathan R, et al: Paradoxical effects of amitriptyline on
borderline patients. Am J Psychiatry 143:1603–1605, 1986a
Soloff PH, George A, Nathan RS, et al: Progress in pharmacotherapy of
borderline disorders: a double-blind study of amitriptyline, haloperi-
dol, and placebo. Arch Gen Psychiatry 43:691–697, 1986b
Soloff PH, Cornelius J, George A, et al: Efficacy of phenelzine and halo-
peridol in borderline personality disorder. Arch Gen Psychiatry
50:377–385, 1993
Stein DJ, Simeon D, Frenkel M, et al: An open trial of valproate in bor-
derline personality. J Clin Psychiatry 56:506–510, 1995
Szigethy EM, Schulz SC: Risperidone in comorbid borderline personality
and dysthymia. J Clin Psychopharmacol 17:326–327, 1997
Tritt K, Nickel C, Lahmann C, et al: Lamotrigine treatment of aggression

in female borderline-patients: a randomized, double-blind, placebo-
controlled study. J Psychopharmacol 19:287–291, 2005
Vorma H, Naukkarinen HH, Sarna SJ, et al: Predictors of benzodiazepine
discontinuation in subjects manifesting complicated dependence.
Subst Use Misuse 40:499–510, 2005
Young JE, Klosko J, Weishaar ME: Schema Therapy: A Practitioner’s
Guide. New York, Guilford, 2003
Zanarini MC, Frankenburg FR: Olanzapine treatment of female border-
line personality disorder patients: a double blind placebo controlled
trial. J Clin Psychiatry 62:849–854, 2001
Zanarini MC, Frankenburg FR, Dubo ED, et al: Axis I comorbidity of bor-
derline personality disorder. Am J Psychiatry 155:1733–1739, 1998
Zanarini MC, Frankenburg FR, Hennen J, et al: Mental health utilization
by borderline personality disorder patients and Axis II comparison
subjects followed prospectively for six years. J Clin Psychiatry 65:28–
36, 2004a
Zanarini MC, Frankenburg FR, Parachini EA: A preliminary randomized
trial of fluoxetine, olanzapine and the olanzapine-fluoxetine combi-
nation in women with borderline personality disorder. J Clin Psychi-
atry 65:903–907, 2004b
Zanarini M, Schulz S, Detke H, et al: A dose comparison of olanzapine for
the treatment of borderline personality disorder: a 12-week random-
ized, double-blind, placebo-controlled study. Paper presented at
25th Biennial Congress of the Collegium Internationale Neuro-
Psychopharmacologicum, Chicago, IL, July 2006
Zimmerman M, Coryell W, Pfohl B, et al: ECT response in depressed pa-
tients with and without a DSM-III personality disorder. Am J Psychia-
try 143:1030–1032, 1986
Chapter 8
FAMILY INTERVENTIONS AND

THERAPIES
FAMILY INTERVENTIONS of any type are sociotherapeutic (Chapter

5)—that is, they directly affect primarily the borderline patient’s social ad-
aptation. As is described in this chapter, family interventions for patients
with borderline personality disorder (BPD) are often initiated during cri-
ses, require education, and often require directives. Such interventions
usually occur after the diagnosis is established and when the borderline
patient is in more restricted levels of care (IV, hospital care, or III, residen-
tial or partial hospital care). These initial family interventions can make
the interactions less stressful both for borderline family members and for
others in a family. More sustained family interventions are usually needed
to change a family’s ways of communicating or relating. In the process of
making such changes, families can become allies in proactively helping
patients change. These changes will primarily take place when the patient
is in level I (outpatient) or II (intensive outpatient) care.
In this chapter, an overall approach to families is described that inte-
grates contributions from both psychoeducational and psychodynamic
therapies. Indeed, as with group and individual therapies, these ther-
apies are complementary but sequential. This family approach was de-
veloped after recognizing that traditional dynamically based family
therapies usually end quickly—and badly. Borderline offspring can batter
the parents into alienated flight, or the borderline individuals themselves
can leave feeling betrayed and “ganged up on.” Yet the failure to involve
families of borderline patients as supports for their treatment often
makes the patients’ own involvement in therapy either superficial or so
fraught with fears of abandonment (often by ostensibly villainous par-
177
ents) that using therapies to change themselves is impossible—a major

reason for premature dropouts (Gunderson et al. 1989).
Comprising this overall approach to families are the phases of family
interventions discussed in further detail later in this chapter:
• Phase 1: Initial family meetings

• Phase 2: Establishing an alliance
• Phase 3: Psychoeducational family therapy
• Phase 4: Psychodynamic family therapy
History
Efforts to intervene with families of borderline patients were first re-
ported in the 1970s. These seminal reports came from a group of com-
mitted, analytically oriented family therapists. Their approach was based
on psychodynamic and systems theories—theories that are linked in view-
ing psychopathology as resulting from conflictual forces within the des-
ignated patients’ social systems. At its extreme, and consistent with the
influential work by Masterson and Rinsley (1975), this approach meant
that borderline psychopathology could not be expected to be meaning-
fully corrected without changing the borderline person’s primary social
milieu, which for many patients is the family. The initial reports were
based on work done with adolescent samples on specialized, relatively
long-term inpatient units at the National Institute of Mental Health (R.L.
Shapiro et al. 1974; Zinner and Shapiro 1975) and subsequently McLean
Hospital (E. R. Shapiro 1978a, 1978b, 1982). These therapists developed
the theory that pathological forms of parental overinvolvement fostered
the borderline offspring’s dependency and abandonment fears. The
therapists also encouraged hopes that intensive long-term family therapy
could bring about curative changes.
When such theory-based, intensive family therapy was immersed
within containing inpatient services and was closely integrated with other
modalities, it was, in my experience, a powerful approach that could be
very useful. Its confrontational, authoritarian approach was, however, of-
ten resented even by the families who could benefit from it. They were, in
any event, self-selected families who sought out and contracted to under-
take this type of treatment program. It was, moreover, an approach that
was not feasible in most settings and was not desirable to most families if
they believed they could avoid it. Certainly, the approach was never
considered appropriate for fragmented, abused, and nonverbal families
or for those whose interactions with their borderline offspring were
sparse—for example, those living elsewhere.
Family Interventions and Therapies ❘ 179
When studies of families with BPD members moved from the province
of clinical observations to that of empirical studies in the 1980s, radical re-
visions in our understanding of the prototypical family occurred (Gunder-
son and Zanarini 1989; Links 1990). Our early work determined that it was
not true that most of these families were overinvolved and separation-
resistant, as suggested by Masterson’s theory (Masterson 1972; Masterson
and Rinsley 1975; see also Chapter 1) and by the pioneering family ther-
apists noted earlier. Rather, we found that most families of borderline
patients were insufficiently involved with the patients during their early de-
velopment (Frank and Paris 1981; Gunderson et al. 1980; Soloff and Mill-
ward 1983a), and these families either perpetrated or were unavailable to
help with traumatic experiences (Gunderson and Sabo 1993; Links and
van Reekum 1993; Links et al. 1990; Millon 1987; Paris et al. 1994a, 1994b).
Neglect and trauma were prototypical (Gunderson and Zanarini 1989;
Zanarini 1997). Another series of studies showed that borderline patients’
parents themselves had serious psychiatric problems, including substance
abuse, depressions, and even BPD itself (Akiskal et al. 1985; Goldman et al.
1993; Links et al. 1988; Loranger et al. 1982; Pope et al. 1983; Schachnow
et al. 1997; Silverman et al. 1991; Soloff and Millward 1983b; Zanarini et al.
1990). All these studies combined to paint a very bleak and very critical pic-
ture of the health, function, and motivation of borderline patients’ fami-
lies. This perception is reflected in the virtual absence of any new articles
about family therapies during the 1980s or 1990s.
A model for BPD that includes significant heritability and social dys-
function is now superimposed on the conflict model found in the early
psychoanalytic literature. This change has quietly encouraged modifica-
tions in the approach to families (Ruiz-Sancho et al. 2001). While these
changes in understanding the family environments of borderline pa-
tients were occurring, relevant research on treating families with a mem-
ber who has schizophrenia opened up a different treatment perspective.
This research showed that schizophrenic individuals who came from fam-
ilies with high expressed emotion—meaning hostile, critical, and overin-
volved—had far higher relapse rates (50% vs. 14%) over the course of 9–
12 months and that a psychoeducational approach could reduce ex-
pressed emotion (i.e., the putative stressor) and thereby greatly reduce
relapse rates (Goldstein 1995; Leff 1989; McFarlane and Dunne 1991;
McFarlane et al. 1995). Indeed, the effect of these psychoeducational
family interventions on relapse rates exceeded the effects resulting from
the introduction of neuroleptics or from any other type of psychosocial
therapy that has been tested (Gabbard et al. 1997).
The basic principles of the psychoeducational family treatments (Ta-
ble 8–1) used in these studies with families who have schizophrenic off-
TABLE 8–1. Principles of psychoeducation for families

Mental illness is a problem within the person, not a symptom of a
problem family.
Family support is needed for treatment of mental illness.
Psychoeducation requires being informed about therapy, prognosis,
and course.
Psychoeducation can diminish harmful anger and criticism.
Families often do not recognize the cost of the illness: family alienation
and social isolation.
“Bad” parents are uninformed or ill, not malevolent.
Families are burdened; new management strategies can reduce this
burden.
spring are radically different from the principles that guided the earlier
family treatments based on psychodynamic system theory. Yet the ratio-
nale and the efficacy of these psychoeducational family treatments of-
fered a model that could not be ignored by a new generation of clinicians
who now had a more deficit-based construct of BPD.
Therapists and Countertransferences

Families, specifically parents, do not see themselves as patients. They are
consumers who clinicians can assume will be somewhat wary and defen-
sive. Therefore, clinicians should encourage them to carefully appraise
treatment recommendations for their offspring or themselves. Anyone
wanting to help families with a borderline relative needs to respect fam-
ilies’ reservations. It helps to have firsthand knowledge about the hard-
ships of parenting. Clinicians who believe that they themselves could
never have a disturbed offspring or who believe that they could readily
manage the problems that borderline patients present to their families
bring critical, intolerant attitudes that aggravate a family’s guilt, anger,
defensiveness, and isolation. Ideally, clinicians who do family interven-
tions combine compassion for the family’s plight with enough experi-
ence and confidence to inspire a family’s trust.
Basic introductory family interventions (phases 1 and 2) often can be
done by the patient’s primary clinician. More sustained family interven-
tions (phases 3 and 4) profit by having an experienced family therapist
who can comfortably coordinate his or her work with the patient’s ther-
apist. Whether a borderline patient’s individual psychotherapist is suited
for doing family therapy depends on that particular therapist’s aptitude
and the particular meaning of family therapy for the psychotherapeutic

relationship. As long as the borderline patient is practicing self-endan-
gering behaviors or is severely vocationally impaired (i.e., usually during
the first year of treatment), it is helpful for the individual therapist to join
family therapies periodically when they are being provided by another
clinician. As discussed later in this chapter, this helps family therapies
become a stronger container (holding force) than when families fail to
witness firsthand the individual therapist’s respectful involvement with
them.
Getting Started: Overcoming Resistance

Family interventions are indicated whenever a BPD patient has signifi-
cant involvement with, or financial dependence on, his or her family.
Hence adolescents or young adult borderline patients are prime candi-
dates. Unfortunately, it is in the nature of many borderline patients to
emphatically resist involvement of their relatives by clinicians. Such dis-
approval or even prohibitions should not be accepted; this resistance re-
quires a serious and sometimes extended working through (Sidebar 8–
1). Resistance should be seen as symptomatic of the typically borderline
pattern of devaluing prior caregivers out of hurt and out of a wish to in-
voke hope for a more idealized protective and exclusive relationship with
a potential new caregiver. This pattern of being devaluative at intake, doc-
umented by Perry et al. (1990), has created a significant bias in the liter-
ature because most of the research that has characterized families of
borderline patients has relied on families’ accounts during intake evalu-
ations (Gunderson and Lyoo 1997).
Vignette
Ms. N, a 23-year-old woman with BPD, came for consultation at the rec-
ommendation of her psychopharmacologist. When I greeted her in the
waiting room, she introduced me to her mother, who was sitting quietly a
few seats away. During the course of the consultation, it became clear that
Ms. N had remained very dependent on her parents, never having sus-
tained a job (“I get too anxious and walk out”) and having had boyfriends
who at times cohabited with her in her family’s house.
After I invited the mother to join us, I began by reviewing the reasons
for the referral and why the borderline diagnosis had been confirmed.
The mother seemed to be familiar with the diagnosis and readily agreed
that “it describes my daughter perfectly.” She went on to talk about how
resistant Ms. N had proved to a long series of therapies beginning when
the mother had started taking her for help when Ms. N was age 13. I said
that Ms. N would seem to be a good candidate for a dialectical behavior
therapy (DBT) group, and we discussed the feasibility of that. Then I gave
them Borderline Personality Disorder: What You Need to Know (see the Appen-
dix at the end of this book for this and other materials) and invited them
to come back, with the patient’s father if possible, in a week. The mother
quietly, without explanation, said that it wouldn’t be possible. I was a bit
surprised but agreed to meet in 2 weeks. Suddenly, I was aware that this
therapy was unlikely to go smoothly. But why?
In the meeting 2 weeks later, it became clear that the mother had
doubts about the likelihood that any therapy would change her daughter
and that the patient’s father, a silent man who drank heavily, believed that
everything about the mental health field was a waste of money.
In this example, it became apparent that between her mother’s co-

dependency and her father’s hostility, Ms. N would endanger her needed
supports at home by becoming involved in therapy. Parental resistance
therefore needed to be addressed first. A clinician wishing to treat this pa-
tient would not have any success unless this patient could believe that her
involvement would not result in the loss of her needed parental support.
Her treatment would require her parents being allied with her involve-
ment in therapy and with its goals.
The initial phases of the family approach described here (phase 1, ini-
tial meeting, and phase 2, alliance building) are directed at calming the
family’s anxieties and establishing an alliance rather than trying to treat
the family per se. In this process, they are helped. Later phases (phase 3,
psychoeducation, and phase 4, psychodynamic therapy), identifiable as
therapies, are explicitly intended to change the families. Special atten-
tion is given to the potential role of multiple-family groups because they
have definite clinical and cost-benefit advantages over single-family inter-
ventions and because they are the only type of family intervention that
has been examined empirically.
Sidebar 8–1: “You Can’t Talk to My Parents”

Borderline patients frequently impose on their clinicians pro-
hibitions about talking to their parents. Clinicians, eager to
avoid conflict or to build an alliance, can find it tempting to
accept such prohibitions. Yet failure to enlist the support of
a borderline patient’s family is one of the major reasons for
patients subsequently dropping out of psychotherapy
(Gunderson et al. 1989). It helps to take a firm stance that
although the patient’s objections are important, they fly in
the face of the value the clinician places on knowing first-
hand about the important people in the patient’s life. Per-
sistence in examining and reality testing the patient’s fears
(e.g., you’ll betray confidences, you’ll take the family’s side,
you’ll “like them better,” they will get too upset) and in ex-
plaining the expected benefits (e.g., this will help your fam-
ily understand you better; this will help your parents be less
suspicious, more supportive of your therapy) is usually suffi-

cient to gain permission.
When borderline patients are in a hospital setting, the
usual evidence of possible dangerousness to self is almost
always sufficient to provide legal justification for overriding
the patients’ prohibitions. Even if a clinician has questions
about actual suicidal intention, this justification should be
used to seize the opportunity it offers. When borderline pa-
tients are not in a hospital, there are still some occasions
when they need to be told that the clinician cannot treat
them (i.e., they cannot expect to improve) without enlist-
ing their parents’ support. This is the case when either 1) the
families’ responses to the patients are life endangering or
2) the patients’ involvement in therapy is experienced as a
betrayal of their parents.
When the patients’ therapy is paid for by their families,
it is essential that this reality become an acknowledged
part of establishing a realistic frame for the therapy (see
Chapter 10). Patients may need to be reminded that their
reliance on parental support for treatment means that
excluding the parents jeopardizes its feasibility. Parents
may need clinicians to remind them that, much like paying
for a child’s college education, they should help decide
whether their investment in therapy is worthwhile. Thera-
pists have a responsibility to inform parents who pay for a
treatment about issues such as their child’s attendance at,
or motivation for, and apparent benefit from therapy.
What remains confidential is what is being talked about.
Family interventions are often best begun during the crises that lead
to hospitalizations. These crises are times when families often feel most in
need of help—especially during the first few hospitalizations. Efforts to
involve families whose borderline member has already been through
many hospitalizations are less likely to be successful because families have
already established adaptations—and often ones that are not helpful,
such as giving up hope, having exaggerated fears of stigma, or having
convictions about psychiatry’s uselessness.
Phase 1: Initial Family Meetings

Problem identification, psychoeducation, and support are essential first
steps in recruiting the borderline patient’s significant others as allies. To
begin an alliance requires that the clinician convey by word and attitude
that he or she is sympathetic to the problems that significant others are
struggling with and knows they have been doing the best they can. (Ac-
tually, the clinician should assume this last point.) The first meeting starts
with unequivocal identification of the patient as having a serious disor-

der, or illness, and having “special needs” because of this. Not only par-
ents but also spouses or children of married borderline patients can
benefit from the same forms of initial contacts, but thereafter they pose
special problems (Sidebar 8–2).
Sidebar 8–2: Families of Married

Borderline Patients
When the borderline patient is married, a clinician should in-
form the spouse about the borderline illness, in the hope
that supportive allowance will be made for the spouse’s
handicaps. The clinician should simultaneously convey a
need to respect and support that borderline partner’s on-
going strengths. Psychoeducation for the spouse or even
very structured skills enhancement instructions for both
partners (Waldo and Harman 1998) can be helpful.
The clinician-therapist does not want to unnecessarily
consign the marriage to a future in which caretaker-
dependent roles are permanent. On the other hand, when
such roles are already stably complementary, couples
therapy is contraindicated (Paris and Braverman 1995).
Couples therapy, like conjoint family therapy, sessions
should await both members’ being able to listen to what
originally each partner could only say about the other in
private—and being able to listen without getting enraged,
terrified, or despairing (Seeman and Edwards-Evans 1979).
For practical purposes, this means that significant change
must have occurred in the borderline spouse before mari-
tal therapy is likely to be of value.
When the borderline patient is a mother, clinicians
should recognize that the children are at high risk for psy-
chiatric problems, including conduct disorders and atten-
tional dysfunction (Links et al. 1990; Weiss et al. 1996). The
patient (and spouse) should be educated about this risk,
with compassionate efforts to ease undue guilt. Most bor-
derline patients are grateful for evaluations of their children
and for any recommendations for assistance.
Problem Identification
Clinicians should actively ask relatives to identify the problems that the
borderline family member has created for them. In our initial survey of
40 families, the most common problems were (in order): 1) communica-
tion, 2) dealing with the hostile or rageful reactions, and 3) fears about
suicide (Gunderson and Lyoo 1997). Once the problems are identified,
clinicians usually can offer assurances that the burden created by these
problems is familiar to mental health staff and can be significantly re-

duced, to everyone’s benefit.
Psychoeducation
Psychoeducation (see Chapters 3 and 11) involves acquainting relatives
with the borderline diagnosis by going through the diagnostic criteria to-
gether and making sure they are understood. This step is followed by
evaluating how these criteria apply or are reflected in their borderline
relative. When a clinician is asked whether someone has BPD, it is useful
to be able to describe it in a way that is relatively jargon free, allowing lay-
persons to reach their own conclusions about whether the diagnosis fits
(see Chapter 1, section “Misuses of the Borderline Diagnosis”). I make
only modest revisions when talking to parents as opposed to the patients
themselves, as shown here:
People with a borderline personality disorder have grown up feeling that

they were unfairly treated, that they didn’t get the attention or care they
needed. They are angry about that, and as young adults, they set out in
search of someone who can make up to them for what they feel is missing.
Hence, they set in motion intense, exclusive relationships, which then fail
because they place unrealistic expectations on the other person. Upon
failing, they feel rejected or abandoned, and they can either become en-
raged again about its unfairness (as they did when growing up), or they
can feel they are bad and unlovable, in which case they become suicidal or
self-destructive.
Either their anger at being mistreated or their feeling bad and being
self-destructive can cause others—especially parents—to feel guilty and
try to make it up to them; it naturally evokes wishes to protect or rescue.
Such responses from others, especially parents, unfortunately validate
borderline persons’ unrealistically high expectations of having their
needs met, and the cycle is apt to repeat itself.
Central to the psychoeducation process is emphasizing that borderline

patients have deficits or handicaps, which can, albeit slowly, be overcome.
(Some patients do not want to hear the “slowly” aspect of this, whereas others
find it reassuring to know that there is no short-term solution. This second
group often has been confused or disillusioned by hearing unrealistically op-
timistic predictions or by being fundamentally aware that the problems are
long-standing and deep-seated.) Encouraging relatives to read or view edu-
cative materials is often instructive. (See the Appendix at the end of this book
for psychoeducational materials.) Assigned reading (bibliotherapy) as
homework to be discussed at the next meeting is a good idea. Equally impor-
tant is to convey a message of respect: that to you the parents are important
allies, not bad people and not rivals with you for their children’s loyalties.
Support
It is important to note that borderline patients are very difficult people to
form helpful relationships with—for clinicians as well as for their fami-
lies. Equally important is to empathize with how burdensome such pa-
tients are for a family. In a study that anticipated shifts in the borderline
construct toward a model based more on medical deficits, Schulz et al.
(1985) compared the burden of having a BPD family member with that of
having someone with a chronic medical illness. Both conditions involve
the burdens caused by dependency and unemployment; in addition,
BPD creates the burden of the borderline individual’s behavior problems
(e.g., drunkenness, promiscuity) and the burden of families’ feeling
blamed, directly or by inference, by both the patients and the clinicians,
who have typically excluded families. Clinicians can offer families an ex-
tremely important support by diminishing the families’ feeling that they
have been responsible for causing the illness—or their fear that clinicians
hold them responsible (Sidebar 8–3). Still, the psychoeducation used for
other disorders that has given the reassuring impression that family envi-
ronment has nothing to do with their offspring’s illness is misleading.
Sidebar 8–3: Finessing the Guilt Issue

Reiss et al. (1995) describe how families enter the mental
health orbit with defensiveness and fear of blame. The very
fact that they have a mentally ill offspring evokes in parents,
regardless of the nature of their child’s psychiatric problem,
an immediate fear that they must have done something
wrong and defensiveness toward expected accusations.
Certainly that defensiveness is exaggerated when the off-
spring has BPD and thus is particularly angry and devalua-
tive, and the child’s hostilities cannot easily be discounted
because he or she is not psychotic (Baker et al. 1992). The
stress of having an offspring who is clearly devaluative
(Perry et al. 1990) and who often says about the family to
anybody and everybody that “they did it; they’re the ones”
makes parents particularly fearful of what mental health
professionals will say.
To reduce family guilt, the clinician should emphasize
to parents that they have a very disturbed and disturbing
offspring—and that he or she is sympathetic to the prob-
lems that such an offspring creates for them. The clinician
should tell the parents that many factors (inborn, develop-
mental, and familial) put children at risk for the develop-
ment of BPD, that early attachment issues involve
“goodness of fit” between a child’s temperament and that
of the caregivers (Thomas and Chess 1984), and that it
takes complex interactions among multiple factors to de-
velop this disorder. When families make more specific in-

quiries about their own responsibility for the disorder, the
clinician should note that it is a reasonable question but
that unfortunately not enough is known to give a very
meaningful answer. At this point, the clinician must tell pa-
tients directly that because issues of causality are usually so
heavily loaded with feelings of anger and guilt, these issues
are rarely constructive for dealing with the current situation
and recurrent crises. The clinician should actively move
parents away from concerns about their possible causal
role and emphasize that the most constructive issue they
can attend to is how to cope with the ongoing problems
they face with this very troubled offspring. The causal issue
is to some extent irrelevant; they have work to do right now.
This approach also sets a tone for their approach to
their offspring’s blaming: that it is not very relevant. Their son
or daughter has problems and must learn how to cope
better than at present. It is a very proactive, future-oriented
approach.
Phase 2: Establishing an Alliance

Establishing a satisfactory working alliance with a family may require only
a few initial meetings, but often it requires more. One reason involves the
guilt and defensiveness identified earlier. Conjoint meetings with the
borderline offspring may be necessary if the patient is very dependent on
the parents; in that situation, failure to have such meetings aggravates
separation fears. Otherwise, however, conjoint meetings remain rela-
tively contraindicated until both the family and the patient have first in-
dependently established a treatment alliance and have found the
requisite perspective or strength to withstand the predictably powerful
conflicts that occur in meeting together.
Two other issues—resistances that can delay forming an alliance—
involve 1) the diagnosis itself and 2) the prospect of greater involvement
in treatment.
Resistance to the Borderline Diagnosis

Some relatives greet the borderline diagnosis with skepticism, being rel-
atively insistent that the problems are better understood as developmen-
tal (e.g., adolescence, authority issues) or are due to one of the more
obvious (or pharmacologically responsive) symptom disorders (Axis I)—
that is, bipolar disorder, depression, eating disorders, or substance abuse.
Occasionally, a family resists the BPD label because they have seen it ap-
plied to patients whom their offspring or the nursing staff have identified
as offensive. In rare cases, families protest the diagnosis because they
know something is wrong with their offspring’s brain functioning. None

of these causes for diagnostic resistance is inconsistent with the BPD di-
agnosis.
It is best to sidestep open disagreements about diagnostic questions.
Clinicians can note how the BPD diagnosis has “added value”: it offers
families a way to understand why the patients’ developmental problems,
their symptoms, and even their offensive behaviors have proved resistant
to treatment (or are unlikely to remit as rapidly as they hope). The clini-
cian can help by persisting in efforts to address the relatives’ skepticism
about the diagnosis by offering more reading and didactic information
about problematic issues such as rages, all-or-none thinking, and irre-
sponsible role functioning. Patience and education will gradually lower
resistance, enlist family collaboration, and allow families to accept the
borderline diagnosis.
Resistance to Being Involved in Treatment

It is often difficult to involve parents who have been abusive or who are cur-
rently abusing substances. In both of these circumstances, involvement
threatens their wish to preserve denial or to avoid shame. Such parents are
often ill themselves—most often alcoholic—or have otherwise parented so
irresponsibly (e.g., been abusive) that they have good reason, consciously
or unconsciously, to fear vilification. Most of the time, when the parental
marriage is intact, one parent-partner is extremely hostile about mental
health involvement, whereas the other is interested though hesitant. Oc-
casionally, persistence by the interested parent can succeed in getting the
hostile, resistant parent involved, but usually it is a standoff. Obviously,
meetings with the interested parent should be offered, but when that par-
ent gets involved, it can catalyze a separation or divorce within a dysfunc-
tional marriage. Because the available parent rarely wants to jeopardize
the marriage on account of this issue, he or she usually drops out or is too
conflicted to gain much from family interventions. When parental separa-
tions occur, they can mobilize one or both parents to get more involved or
otherwise become more attentive to their borderline offspring’s care.
End of Phase 2: Transition From Treatment to Therapy

When the phase of alliance building is complete, the family is “ready” for
therapy per se (i.e., ready to try to change). Readiness for involvement in
family therapy is noted by three indications: 1) accepting the borderline
diagnosis, or at least the possibility; 2) being reconciled to a long-term
course of illness; and 3) wanting help in the way they relate to the bor-
derline relative. At this point, the family members are ready for interven-
tions that will help them to change their ways of communicating and
relating. Workshops in which 10–20 families meet for a half-day to learn
about the diagnosis, share experiences, and hear some hopeful messages
offer an excellent vehicle for consolidating this phase (Berkowitz and
Gunderson 2002).
Phase 3: Psychoeducational Family Therapy

The goals of the psychoeducational approach in phase 3 are to improve
family communication, diminish hostilities, and decrease burden. Table
8–2 outlines the range of issues with which psychoeducation can help. To
accomplish these goals, families are taught how to create a more cool,
calm, and predictable home environment. The therapist functions more
as a teacher or leader than as an explorer, confidant, or transference ob-
ject. Psychoeducational family therapies can be done with individual fam-
ilies or with a multiple-family group.
Single-Family Interventions
Single-family interventions are usually more feasible for clinicians and
more comfortable for parents. It should be emphasized that, as with the
first two phases, conjoint meetings are not yet recommended. Parent
management training, developed by Patterson (1982; Patterson et al.
1992) for treatment of conduct disorder, offers a valuable model. This
training is based on the idea that parent-child interactions may inadvert-
TABLE 8–2. Issues for psychoeducational family therapy

• Lack of parental consensus building: predictable “good cop, bad
cop” roles
• “Parentifying” the child (Shane and Kovel 1988)
• Misattributions of offspring
Independence (actually still dependent)
(Young and Gunderson 1995)
Demandingness or hostility (actually often fearful) (Fonagy 1991)
Sociability (actually lacks close friends)
(Young and Gunderson 1995)
• Reducing emotionality (coping skills)
“Walk away”
Listen
ently promote angry and self-destructive behaviors by poorly thought-

through reinforcements and ineffectual or inconsistent punishments.
Parents who stay up late to reprimand their borderline offspring for dis-
regarding curfew may in fact be reinforcing this behavior. Parents are
coached about alternative responses (e.g., stop staying awake and repri-
manding; rather, reduce your financial support). Parents then field-test
their new learning with their offspring, returning to review what hap-
pened. Unfortunately, manuals for this coaching model have not yet
been written, and it has not been tested with borderline patients.
Adaptations for a Multiple-Family Group

Psychoeducational multiple-family groups were independently begun at
McLean Hospital in 1994 (Gunderson et al. 1997) and at Westchester Di-
vision of the New York Hospital in 1995 (Hoffman and Hooley 1998).
These initiatives came after McFarlane had reported advantages of mul-
tiple-family groups over single-family interventions in families that had a
member with schizophrenia (McFarlane and Dunne 1991; McFarlane et
al. 1995). Our experience and that of Hoffman (1999) confirmed that
multiple-family groups are more satisfactory to consumers as well as more
cost-effective. The main problem with multiple-family groups is logistical:
they require clinical sites with a sufficient flow of borderline patients with
suitable families to constitute a group. Hence the format is better suited
to hospitals or clinics than to most office practices.
The format used in the McLean outpatient program follows a manual-
guided book (Berkowitz and Gunderson 2002) that initially borrowed
heavily from the one used by McFarlane for families with a member who
has schizophrenia (McFarlane and Dunne 1991; McFarlane et al. 1995).
Undertaking the psychoeducational approach by having multiple families
meet together has advantages. By discussing their problems with similarly
beleaguered—and often similarly isolated—families, participants gain a
strong social support system.
Meetings are 90 minutes every 2 weeks. Four to six families is the ideal
number to allow adequate attention to be paid to everyone. Socializing
times at the start and end of meetings make them informal and reinforce
the idea that there is more to life than problems. The structured agendas
for the meetings help keep the meetings cool and task oriented. The
groups can begin to effect changes after 3 months. Many families can
leave after 1 year, although for some, 18 months seems preferable. Three
stages of the psychoeducational multiple-family group are outlined in the
following subsection. (A variation on this format is offered by Hoffman
[1997; Hoffman and Hooley 1998], who uses larger groups with more fre-
quent sessions and more didactic DBT-based exercises.)
Three Stages of Psychoeducational Family Therapies

These stages (early, middle, late) apply to either single-family or multiple-
family psychoeducational interventions.
EARLY STAGE
The early stage involves a more intensive schedule (weekly, if possible, for
about 2 months) and more active direction and didactics by the family
therapist–teacher. The leader’s active structuring of sessions deliberately
decreases the emotionality of meetings. It is especially important to rec-
ognize how upsetting it can be for some parents to hear about others’
problems. A disproportionate number of the families who enter therapy
are hypersensitive, rather than callous, to being exposed to feelings of
sadness and anger or signs of conflict. Fonagy (1995) and Fonagy et al.
(1995) have suggested that parental difficulty in recognizing, tolerating,
or expressing feeling is one of the situations that can lead to children’s
developing borderline deficits. The ongoing process of didactic psycho-
education exercises also can ease common anxieties about self-disclosure
or help seeking.
Family guidelines are given (Table 8–3), and leaders actively promote
adherence to them, advising families to keep them on the refrigerator
door or under their pillows. In virtually every group session, leaders make
frequent reference to them during discussions. The effect, beyond the lit-
eral application of guidelines, is to cause parents to stop and think before
reacting.
One guideline referred to often during the early phase concerns self-
harm (guideline 7). This guideline can lead to immediate changes in a
family’s response, which will bring its members relief and diminish both
the opportunities and the need for splitting or secondary gain and for
their sequelae: emergency department or hospital services. Another
guideline that may provide immediate benefits involves severely split pa-
rental roles (guideline 10) (Sidebar 8–4).
Sidebar 8–4: “Good Cop/Bad Cop”:

A Parental Problem
It is not unusual for parents to divide the roles that they play,
by common agreement, for their children. The most com-
mon example is that one parent, the “good cop,” provides
an abundance of the caregiving and protection (usually
the mother), whereas the other parent, the “bad cop,” em-
bodies the needed disciplinary and authoritarian roles (usu-
ally the father). Under ordinary circumstances, this works
satisfactorily, but for an offspring with BPD, such a division in-
TABLE 8–3. Family guidelines in relationships with a borderline

member
Goals: go slowly
1. Remember that change is difficult to achieve and fraught with fears. Be
cautious about suggesting that “great progress” has been made and
about giving “you can do it” reassurances. “Progress” evokes fears
of abandonment.
2. Lower your expectations. Set realistic goals that are attainable. Solve
big problems in small steps. Work on one thing at a time. Big goals
or long-term goals lead to discouragement and failure.
Family environment: keep things cool
3. Keep things cool and calm. Appreciation is normal. Tone it down.
Disagreement is normal. Tone it down too.
4. Maintain family routines as much as possible. Stay in touch with family
and friends. There’s more to life than problems, so don’t give up
the good times.
5. Find time to talk. Chats about light or neutral matters are helpful.
Schedule times for this if you need to.
Managing crises: pay attention but stay calm
6. Don’t get defensive in the face of accusations and criticisms. However
unfair, say little and don’t fight. Allow yourself to be hurt. Admit to
whatever is true in the criticisms.
7. Self-destructive acts or threats require attention. Don’t ignore. Don’t
panic. It’s good to know. Do not keep secrets about this area. Talk
about it openly with your family member, and make sure
professionals know.
8. Listen. People need to have their feelings heard. Don’t say “It isn’t so.”
Don’t try to make the feelings go away. Using words to express fear,
loneliness, inadequacy, anger, or needs is good. It’s better to use
words than to act out on feelings.
Addressing problems: collaborate and be consistent
9. When solving a family member’s problems, always:
Involve the family member in identifying what needs to be done.
Ask whether the person can “do” what’s needed in the solution.
Ask whether he or she wants you to “do” what’s needed.
TABLE 8–3. Family guidelines in relationships with a borderline

member (continued)
10. Family members need to act in concert with one another. Parental
inconsistencies fuel severe family conflicts. Develop strategies that
everyone can stick to.
11. If you have concerns about medications or therapist interventions, make
sure that both your family member and his or her therapist or physician
knows. If you have financial responsibility, you have the right to
address your concerns to the therapist or physician.
Limit setting: be direct but careful
12. Set limits by stating the limits of your tolerance. Let your expectations be
known in clear, simple language. Everyone needs to know what is
expected of them.
13. Do not protect family members from the natural consequences of their
actions. Allow them to learn about reality. Bumping into a few walls
is usually necessary.
14. Do not tolerate abusive treatment such as tantrums, threats, hitting, and
spitting. Walk away and return to discuss the issue later.
15. Be cautious about using threats and ultimatums. They are a last resort. Do
not use threats and ultimatums as a means of convincing others to
change. Give them only when you can and will carry through. Let
others—including professionals—help you decide when to give
them.
vites trouble. The nurturing parent’s attitudes help consoli-

date the borderline offspring’s desires into entitlements,
and the disciplinary parent allows the borderline offspring
to translate his or her own disappointments into parental
cruelties.
In such circumstances, parents who recognize their
pattern of splitting can benefit from guideline 10. They are
told to correct splitting conscientiously by 1) always com-
municating with each other and only then deciding on
mutually acceptable responses to problems and 2) having
the bad cop offer a favorable parental response and hav-
ing the good cop offer a response that the offspring does
not want. This directive has immediate benefits. Parents
working collaboratively diminish their alienation from each
other; both feel more supported and confident. The bor-
derline offspring feels less responsible for parental conflict
and more truly held by the family.
During the early phase, a variety of exercises can begin to modify ha-
bitual patterns of interaction with the borderline offspring:
• The book Don’t Shoot the Dog: The New Art of Teaching and Training
(Pryor 1999) can be used to teach parents basic skills in behavioral
conditioning.
• Communications and confrontations exercise. Family members are asked
to role-play the how-to of doing confrontations in prescribed ways.
There are three components: “I feel ...”; “You did ...”; and “I want ...”
Hoffman (1999) used the DBT-based “Dear Man” exercise for similar
purposes.
• Managing criticism exercise. This exercise relates to guideline 6 in Ta-
ble 8–3. It uses some standard behavioral therapy techniques and is re-
inforced by instructive imagery (e.g., using a sponge as a metaphor for
patience and resiliency during confrontations).
• Attributions exercise. This exercise borrows from Fonagy’s (1991,
1995) description of how interactional patterns in childhood have led
to some typical ways in which borderline patients misattribute feelings
or motives to parents and others. Adapted from Fonagy’s concept of
“mentalizing,” this exercise teaches parents to “speak Borderlinese”
(i.e., to develop an awareness of the thoughts, fears, and needs under-
lying their borderline offspring’s words and behaviors). For example,
when people with BPD say they “hate” someone, it usually means that
they feel rejected; when they say they “don’t need anyone,” it means
that they believe being needy is unacceptable. In this manner, parents
learn how to understand and respond more accurately to such typi-
cally all-or-none borderline statements.
MIDDLE STAGE
The frequency of meetings diminishes to every 2 weeks in the middle
stage. At this point, conjoint meetings with the borderline offspring can
begin because the parents should have enough distance, support, humil-
ity, and new understanding not to respond to the inevitable opportunities
to get into heated struggles. Good indicators of the borderline patients’
ability to use the conjoint meetings successfully are that the patients have
an alliance with a primary clinician-therapist or are within a more inten-
sive treatment setting (levels II–IV) where they can process the feelings
evoked by meeting with their parents.
The format in the meetings now deals more exclusively with problem
solving. Families are expected to describe a current situation with the ex-
pectation that other family members and the therapist will offer sug-
gestions about how to respond. The family members of the borderline
patient are aware that they will be encouraged to change patterns of

response to the patient. Again, during this stage, the therapist often
encourages the family to make reference to the guidelines and often un-
derscores the message that change is not easy—for anyone. Within mul-
tiple-family groups, this is the stage when the group “comes together”—
develops a cohesion based, in part, on a good working knowledge of one
another’s shared problems.
LATE STAGE
Existing evidence shows that the family functions better now (Sidebar 8–
5). Conjoint meetings with the borderline member have become more
comfortable; open hostilities are usually bypassed. The problem-solving
format is predictable; family members feel bolder and more confident in
giving feedback to one another. Some have made changes that they take
pride in; others have persistent difficult-to-solve problems that everyone
is familiar with. It helps when the borderline offspring knows that a par-
ent is trying to change, even if he or she fails.
The therapist’s role now is seldom directive; rather, the therapist fa-
cilitates efforts to understand or communicate that family members can
increasingly undertake themselves. Gains made by this time (approxi-
mately 1 year) may lead to termination. Sometimes this is a time when
particularly emotional statements about guilt or angry feelings toward
the borderline family member are voiced. Within the multiple-family
group, this usually evokes much support from others.
The psychoeducational therapist encourages families to make ongo-
ing use of the new skills, move on to psychodynamic therapy if it is indi-
cated, return for added psychoeducational meetings as needed, or join
self-help groups and become proactive advocates for other families with
problems similar to their own (Chapter 13).
Sidebar 8–5: Makes Sense, But Does it Work?

Preliminary Findings of the
Psychoeducational Multiple-Family Group
The goals of psychoeducational therapies are to effect
changes in the borderline patient’s family that will second-
arily effect positive changes in the patient’s course of ill-
ness. More specific goals involve improving communica-
tions, diminishing hostilities, and diminishing struggles over
control and independence. One of the central vehicles by
which this is accomplished is making families better in-
formed. Two of the expected benefits from the changes
are decreased family burden and decreased feelings of
alienation.
At present, there has been only one empirical investi-

gation of psychoeducational therapies: a pilot study in
which psychoeducational multiple-family groups were
used (Gunderson et al. 1997). Table 8–4 shows results from
11 families who participated in two psychoeducational
multiple-family groups that were 1 year in duration.
The BPD subjects made changes in desirable ways dur-
ing the year: diminished hospitalizations and self-destruc-
tive acts. Of course, it is not possible to infer that the family
changes were responsible. What is clear is that the con-
sumers felt very pleased with what they learned.
Phase 4: Psychodynamic Family Therapy

Psychodynamic family therapy requires the borderline patient’s active
participation. It should be used selectively, when family readiness has
been established by the family’s completing less-demanding interven-
tions of the types described in the earlier phases. The readiness to com-
municate, the ability to recognize (and not act on) feelings, and the use
of validation are indications. Without such preparation, this phase is con-
traindicated: it often results in alienating either the patient or the family
members to such a degree that treatment of any sort may be infeasible.
TABLE 8–4. Results of pilot study that used psychoeducational

multiple-family groups (N = 11)
AREA ASSESSED DEGREE OF CHANGEa
Communication ++
Hostilities or conflict +
Criticism +
Independence or control ++
Conflict about +
Separation anxiety +
Feeling overcontrolled +
Emotional overinvolvement ±
Knowledge ++
Burden ++
a++= >1
standard deviation (SD) in desired direction; + = 0.5–1 SD in desired
direction; ± = mixed results.
Psychodynamic therapy is not focused on crisis management and on

learning to talk and listen to each other better.
Usually, the duration of psychodynamically based family therapies is
open-ended. The therapy with families having a borderline member in-
volves an effort to 1) enhance family closeness through the expression of
feelings toward one another and 2) enhance understanding of one an-
other through personal self-disclosures and the recognition and accep-
tance of individual differences.
What distinguishes these goals from any other good-quality, standard
psychodynamic family therapy can be found in the particular significance
of certain types of family dynamics in families with a borderline member
(Table 8–5). Earlier accounts of dynamically based therapies for families
with a borderline member—beyond noting inconsistencies and role prob-
lems, with which the educational-behavioral-managerial approach de-
scribed earlier can help—were largely concerned with the recognition of
projections and projective identifications and with repossessing (“owning”)
what rightly belongs to each family member’s own self (E.R. Shapiro 1992).
Marital or Couples Therapy

Initial Meetings
When the borderline patient is a mother, clinicians should recognize that
the children are at high risk for psychiatric problems, including conduct dis-
orders and attentional dysfunction (Links et al. 1990; Weiss et al. 1996). The
patient (and spouse) should be educated about this risk, with compassion-
TABLE 8–5. Issues for psychodynamic family therapy in families

with a borderline member
Hostile or withdrawing responses to separation initiatives
(Masterson 1972)
Marital bonding that promotes distance, projection, and invalidation of
the borderline offspring (Feldman and Guttman 1984; Gunderson et
al. 1980; Shane and Kovel 1988)
Projective identification: projections evoking confirmatory (but
unrepresentative) responses from others (Feldman and Guttman
1984; E.R. Shapiro et al. 1975)
Lack of curiosity about offspring combined with pathological certainty
(E.R. Shapiro 1982)
Note. These issues can be two-way: the borderline offspring can behave toward
the parents in the same ways that parents behave toward the offspring.
ate efforts to ease undue guilt. Most borderline patients are grateful for eval-
uations of their children and for any recommendations for assistance.
When the borderline patient is married, a clinician should inform the
spouse about the borderline illness in the hope that supportive allow-
ances will be made for the borderline spouse’s handicaps. The clinician
should simultaneously convey a need to respect and support that border-
line partner’s ongoing strengths. Psychoeducation for the spouse or even
very structured skills enhancement instructions for both partners can be
helpful (Waldo and Harman 1998).
Initial meetings with both borderline patients and their spouses should
assess suitability for couples therapy. The clinician-therapist does not want
to consign the marriage unnecessarily to a future in which caregiver-depen-
dent roles are permanent. However, when such roles are already stably com-
plementary, couples therapy is probably contraindicated (Paris and
Braverman 1995). Couples therapy, like conjoint family therapy, sessions
should not begin until both members are able to listen to what each partner
originally could say about the other only in private and able to listen without
getting enraged, terrified, or despairing (Seeman and Edwards-Evans
1979). For practical purposes, this means that significant change must occur
in the borderline spouse before couples therapy is likely to be of value.
The American Psychiatric Association (2001) practice guidelines pro-
vide the following goals for couples therapy: stabilizing and strengthen-
ing the couple’s relationship, clarifying nonviable relationships, and
educating the spouse of the borderline patient about BPD and its inter-
personal aspects. Couples therapy also can affect the BPD patient’s inter-
personal functioning and may enrich the person’s individual work.
Possible adverse effects of couples therapy that must be considered are
aggravating spousal violence, committing the borderline spouse to the
“sick role,” or resulting in the therapist aligning with one spouse against
the other. To decide whether couples therapy is appropriate for a border-
line patient, the primary clinician can try to fit the couple into the follow-
ing typology (adapted from Links and Stockwell 2001).
Common Types of Couples

ACTING-OUT COUPLES
Acting-out couples present with a history of acting-out conflicts; impul-
sive, self-destructive behaviors; and sometimes spousal violence.
Vignette
At the request of my close colleague, Ms. O was seen with her partner of
several months. Although Ms. O had a long history of impulsivity and fre-
quent self-harm, which she relayed in detail during the joint session, my
colleague thought that she had been slightly more stable since her part-
nering with Mr. P. Ms. O was angry with her partner because he would not
make a commitment to their relationship, and he seemed more dedi-
cated to his buddies than to her. She was enraged because he had never
introduced her to his parents and never had the money or time for a va-
cation. During the onslaught, Mr. P sat quietly in his chair, and his silence
allowed Ms. O to rage on.
Toward the end of our session as I attempted to wrap up, Ms. O added,
“I’m afraid of him.” When asked about her fear, Ms. O explained that Mr.
P had pushed her and threatened her on several different occasions.
Given my colleague’s perceptions, I was totally perplexed about what to
do with this couple.
With further individual assessment of the spouses, Ms. O reported an
escalating pattern of interpersonal violence, and her self-harm and im-
pulsivity had continued but was more hidden. During the individual ses-
sions, we directed Ms. O and Mr. P to develop a safety plan for themselves.
Ms. O informed herself of safe homes in her area, and we warned Mr. P
that their escalating conflicts could lead to criminal charges. We pointed
out to the couple that couples therapy was contraindicated; individual
therapy to lower the level of impulsivity was recommended; and if the pat-
tern of escalating violence continued, separation was advised.
Couples with high levels of impulsivity are not good candidates for
couples therapy. Highly impulsive individuals with BPD tend to partner
with impulsive, possibly abusive spouses; for example, the borderline
woman who cohabits with an antisocial man. These couples tend not to
be able to sustain marital relationships. However, when the impulsivity is
moderate or infrequent, a marital relationship can help temper the im-
pulsivity (Quinton et al. 1984). Paris and Braverman (1995) noted that
older caregiving husbands could attenuate borderline psychopathology
in their young wives. But when impulsivity has serious consequences,
such as spousal abuse, or is affecting most aspects of the borderline pa-
tient’s functioning, individual rather than couples therapy is indicated.
MUTUAL PROJECTIVE–TYPE COUPLES

Mutual projective–type couples involve the partnering of a borderline
woman with marked identity disturbance with a man who has narcissistic
and borderline features (Links and Stockwell 2001, 2002). They enter re-
lationships that allow them to develop some stability of identity as a result
of their enmeshed attachment. Paradoxically, although this couple may
experience crises, they will have a strong commitment and need for con-
tinuation of the relationship. Such partnerships have been characterized
as a “closed system” (Lloyd and Paulson 1972) having “pathological ho-
meostasis” (Akhtar 1995). Couples with these features are the best candi-
dates for ongoing couples therapy.
Vignette
Mrs. Q, who was receiving treatment for depression and BPD, had a no-
torious reputation as a demanding, entitled patient. In keeping with her
usual presentation, she insisted that her husband accept marital therapy.
The marriage was tumultuous and characterized by many brief separa-
tions. Arguments were frequent, and Mrs. Q would dissolve the conflict by
spending a few days with her elder sister. Two things strengthened the
couple’s commitment to marital therapy: her sister was putting limits on
Mrs. Q’s visits, and their teenage daughter was showing increasing signs of
distress. She was often angry, labile, and prone to feeling victimized. Mr. Q
was aloof and emotionally avoidant but easily recognizable as being tre-
mendously insecure. However, they both came to therapy concerned that
the turmoil was damaging their daughter.
Couples therapy was begun. They were seen weekly until Mr. and Mrs. Q
were able to achieve some relationship stability. Once the couple was more
stable, therapy focused on the interactions that provoked their insecurities
and on their interpersonal skills. For their daughter’s benefit, they pur-
posely worked to lessen the turmoil at home and developed better skills to
work out conflicts. However, once their daughter moved away to college, the
marriage ended with mutual awareness that true intimacy was not possible.
CAREGIVING SPOUSE–TYPE COUPLES

Although caregiving spouse–type relationships typically will be intense,
often a healthy spouse can serve as a receptacle for the emotional out-
pourings. The healthy spouse will tolerate the confusion and anger if the
borderline partner meets some of his or her needs. Typically, this pattern
is seen when a borderline woman partners with a relatively healthy man
who is conscientious and obsessional.
Vignette
Mr. and Mrs. R came for therapy after Mr. R’s release from the hospital fol-
lowing major surgery. Mr. R was exhausted and felt burnt out. Mrs. R was
highly anxious given her husband’s condition. She was being followed up
for recurrent depression and BPD. Despite high levels of tension in the
home, the couple was able to parent six children, co-own a large automo-
tive dealership, and maintain a prominent role in the community. They
were strongly committed to their family, and Mr. R was looking for help to
be able to deal with his wife’s moods.
Couples in which the spouse with BPD is married to someone who has
relative psychological health are best served by a psychoeducational
model. This allows the healthier spouse to stabilize and maintain the re-
lationship. Psychoeducational approaches for family members of border-
line patients, as discussed earlier in the chapter, will be very useful for the
healthy spouse and will potentially prevent the spouse from burning out
before the spouse with BPD is able to change.
Therapeutic Processes
Couples therapy is appropriate for BPD patients who are involved in trou-
bled relationships in which both partners expect to continue into the fu-
ture. Couples with a partner who has BPD often experience a repetitive
cycle of going from crisis, to a sense of security and comfort, to a new crisis.
As noted repeatedly in this book, when borderline patients experience
threats of abandonment or rejection, this generates great anxiety and can
lead to suicidal behavior. However, such crises will quickly dissolve when a
threatened attachment relationship is reestablished or a new holding envi-
ronment is established. This is the first step in couples therapy (Links and
Stockwell 2001, 2002). Such stabilization can lead to dramatic improve-
ment in a short time. Next, the therapist has to strengthen this security and
increase the couple’s feeling of safety. Once the couple is more secure and
engaged in therapy, the therapy explores the precipitants and conse-
quences of their insecurity and how it likely characterizes many of their pre-
vious relationships. Finally, the therapist addresses the skills deficits that
these couples manifest on the basis of the individual and couple assessment.
Summary
The introduction of structured psychoeducational approaches to fami-
lies with a borderline member has been welcomed. Such an approach ac-
tively allies families with treatment goals, builds skills, and, if done well,
improves communication and reduces hostilities with the borderline
family member. The format presented in this chapter seems to work well,
but other formats are feasible and can be expected to evolve along with
the growth of clinical experience and of scientifically based knowledge
about the pathogenic—or ameliorative—role of the family. What is
clearly evident already is that such interventions require only modest
training, are readily exportable, and are very cost-effective. The role of
traditional expressive psychodynamic therapies may still be important,
but these therapies should be initiated selectively, often only after fami-
lies have already benefited from more educational approaches.
References
Akhtar S: Quest for Answers: A Primer for Understanding and Treating
Severe Personality Disorders. Northvale, NJ, Jason Aronson, 1995
Akiskal HS, Chen SE, Davis GC, et al: Borderline: an adjective in search of
a noun. J Clin Psychiatry 46:41–48, 1985
American Psychiatric Association: Practice guidelines for the treatment
of patients with borderline personality disorder. Am J Psychiatry
158(suppl):1–52, 2001
Baker L, Silk KR, Westen D, et al: Malevolence, splitting, and parental rat-
ings by borderlines. J Nerv Ment Dis 180:258–264, 1992
Berkowitz CB, Gunderson JG: Multifamily psychoeducational treatment
of borderline personality disorder, in The Multifamily Group. Edited
by McFarlane WR. London, Oxford University Press, 2002, pp 593–
613
Feldman RB, Guttman HA: Families of borderline patients: literal-
minded parents, borderline parents, and parental protectiveness.
Am J Psychiatry 141:1392–1396, 1984
erations in the treatment of a borderline patient. Int J Psychoanal
72(Pt 4):639–656, 1991
Fonagy P: Playing with reality: the development of psychic reality and its
malfunction in borderline personalities. Int J Psychoanal 76(Pt
1):39–44, 1995
Fonagy P, Steele M, Steele H, et al: Attachment, the reflective self, and
borderline states: the predictive specificity of the Adult Attachment
Interview and pathological emotional development, in Attachment
Theory: Social, Developmental, and Clinical Perspective. Edited by
Goldberg S, Muir R, Kerr J. Hillsdale, NJ, Analytic Press, 1995, pp
223–278
Frank H, Paris J: Recollections of family experience in borderline pa-
tients. Arch Gen Psychiatry 38:1031–1034, 1981
Gabbard GO, Lazar SG, Hornberger J, et al: The economic impact of psy-
chotherapy: a review. Am J Psychiatry 154:147–155, 1997
Goldman SJ, D’Angelo EJ, DeMaso DR: Psychopathology in the families
of children and adolescents with borderline personality disorder. Am
J Psychiatry 150:1832–1835, 1993
Goldstein MJ: Psychoeducation and relapse prevention. Int Clin Psycho-
pharmacol 9:59–69, 1995
Gunderson JG, Lyoo IK: Family problems and relationships for adults
with borderline personality disorder. Harv Rev Psychiatry 4:272–278,
1997
atry 150:19–27, 1993
Gunderson JG, Zanarini MC: Pathogenesis of borderline personality dis-
order, in American Psychiatric Press Review of Psychiatry, Vol 8. Ed-
ited by Tasman A, Hales RE, Frances AJ. Washington, DC, American
Gunderson JG, Kerr J, Englund DW: The families of borderlines: a com-
parative study. Arch Gen Psychiatry 37:27–33, 1980
Gunderson JG, Frank AF, Ronningstam EF, et al: Early discontinuance of
borderline patients from psychotherapy. J Nerv Ment Dis 177:38–42,
1989
Gunderson JG, Berkowitz C, Ruiz-Sancho A: Families of borderline pa-

tients: a psychoeducational approach. Bull Menninger Clin 61:446–
457, 1997
Hoffman PD: A family partnership. J Calif Alliance Ment Ill 8:52–53, 1997
Hoffman P: Family intervention: DBT family skills training. Presented at
the Sixth International Congress on Disorders of Personality, Geneva,
Switzerland, September 1999
Hoffman PD, Hooley JM: Expressed emotion and the treatment of bor-
derline personality disorder. In Session 4(3):39–54, 1998
Leff J: Controversial issues and growing points in research on relatives’
expressed emotions. Int J Soc Psychiatry 35:133–145, 1989
Links PS (ed): Family Environment and Borderline Personality Disorder.
Links PS, Stockwell M: Is couple therapy indicated for borderline person-
ality disorder? Am J Psychother 55:491–506, 2001
Links PS, Stockwell M: The role of couple therapy in the treatment of nar-
cissistic personality disorder. Am J Psychother 56:522–538, 2002
Links PS, van Reekum R: Childhood sexual abuse, parental impairment
and the development of borderline personality disorder. Can J Psy-
chiatry 38:472–474, 1993
Links PS, Steiner M, Huxley G: The occurrence of borderline personality
disorder in the families of borderline patients. J Personal Disord 2:
14–20, 1988
Links P, Boiago I, Huxley G: Sexual abuse and biparental failure as
etiologic models in BPD, in Family Environment and Borderline Per-
sonality Disorder. Edited by Links PS. Washington, DC, American
Lloyd RA, Paulson I: Projective identification in marital relationships as a
resistance in psychotherapy. Arch Gen Psychiatry 27:410–414, 1972
Loranger AW, Oldham JM, Tulis EH: Familial transmission of DSM-III
borderline personality disorder. Arch Gen Psychiatry 39:795–799,
1982
Masterson JF, Rinsley DB: The borderline syndrome: the role of the
mother in the genesis and psychic structure of the borderline per-
sonality. Int J Psychoanal 56:163–177, 1975
McFarlane WR, Dunne E: Family psychoeducation and multi-family
groups in the treatment of schizophrenia. Dir Psychiatry 11(20):2–7,
1991
McFarlane WR, Lukens E, Link B, et al: Multiple-family groups and psy-
choeducation in the treatment of schizophrenia. Arch Gen Psychia-
try 52:679–687, 1995
Millon T: On the genesis and prevalence of the borderline personality
disorder: a social learning thesis. J Personal Disord 1:354–372, 1987
Paris J, Braverman S: Successful and unsuccessful marriages in border-

line patients. J Am Acad Psychoanal 23:153–166, 1995
Paris J, Zweig-Frank H, Guzder J: Psychological risk factors for borderline
personality disorder in female patients. Compr Psychiatry 35:301–
305, 1994a
Paris J, Zweig-Frank H, Guzder J: Risk factors for borderline personality
in male outpatients. J Nerv Ment Dis 182:375–380, 1994b
Patterson GR: Coercive Family Process. Eugene, OR, Castalia Publishing,
1982
Patterson GR, Reid JB, Dishion TJ: Antisocial Boys. Eugene, OR, Castalia
Publishing, 1992
Perry JC, Herman JL, van der Kolk BA, et al: Psychotherapy and psycho-
logical trauma in borderline personality disorder. Psychiatr Ann 20:
33–43, 1990
Pope HG Jr, Jonas JM, Hudson JI, et al: The validity of DSM-III borderline
personality disorder: a phenomenologic, family history, treatment re-
sponse, and long-term follow-up study. Arch Gen Psychiatry 40:23–
30, 1983
Pryor K: Don’t Shoot the Dog: The New Art of Teaching and Training, Re-
vised Edition. New York, Bantam Doubleday Dell, 1999
Quinton D, Rutter M, Liddle C: Institutional rearing, parenting difficul-
ties and marital support. Psychol Med 14:107–124, 1984
Reiss D, Hetherington EM, Plomin R, et al: Genetic questions for envi-
ronmental studies: differential parenting and psychopathology in ad-
olescence. Arch Gen Psychiatry 52:925–936, 1995
Ruiz-Sancho AM, Smith GW, Gunderson JG: Psychoeducational ap-
proaches, in Handbook of Personality Disorders: Theory, Research,
460–474
Schachnow J, Clarkin J, DePalma CS, et al: Biparental psychopathology
and borderline personality disorder. Psychiatry 60:171–181, 1997
Schulz PM, Schulz SCH, Hamer R, et al: The impact of borderline and
schizotypal personality disorders on patients and their families. Hosp
Seeman MV, Edwards-Evans B: Marital therapy with borderline patients:
is it beneficial? J Clin Psychiatry 40:308–312, 1979
Shane M, Kovel V: Family therapy in severe personality disorders. Inter-
national Journal of Family Psychiatry 9:241–258, 1988
Shapiro ER: The psychodynamics and developmental psychology of the
borderline patient: a review of the literature. Am J Psychiatry 135:
1305–1315, 1978a
Shapiro ER: Research on family dynamics: clinical implications for the
family of the borderline adolescent. Adolesc Psychiatry 6:360–376,
1978b
Shapiro ER: On curiosity: intrapsychic and interpersonal boundary for-

mation in family life. International Journal of Family Psychiatry 3:69–
89, 1982
Shapiro ER: Family dynamics and borderline personality disorder, in
Handbook of Borderline Disorders. Edited by Silver D, Rosenbluth
M. Madison, CT, International Universities Press, 1992, pp 471–493
Shapiro ER, Zinner J, Shapiro RL, et al: The influence of family experi-
ence on borderline personality development. International Review
of Psychoanalysis 2:399–411, 1975
Shapiro RL, Zinner J, Shapiro ER: Concurrent family treatment of nar-
cissistic disorders in adolescents. Int J Psychoanal Psychother 4:379–
396, 1974
Silverman JM, Pinkham L, Horvath TB, et al: Affective and impulsive per-
sonality traits in relatives of borderline patients. Am J Psychiatry 148:
1378–1385, 1991
Soloff PH, Millward JW: Developmental histories of borderline patients.
Compr Psychiatry 24:574–588, 1983a
Soloff PH, Millward JW: Psychiatric disorders in the families of border-
line patients. Arch Gen Psychiatry 40:37–44, 1983b
Thomas A, Chess S: Genesis and evolution of behavioral disorders: from
infancy to early adult life. Am J Psychiatry 141:1–9, 1984
Waldo M, Harman MJ: Borderline personality disorder and relationship
enhancement marital therapy, in The Disordered Couple. Edited by
Carlson J, Sperry L. Bristol, PA, Brunner/Mazel, 1998, pp 285–297
Weiss M, Zelkowitz P, Feldman RB: Psychopathology in offspring of moth-
ers with borderline personality disorder: a pilot study. Can J Psychia-
try 41:285–290, 1996
Young DW, Gunderson JG: Family images of borderline adolescents. Psy-
chiatry 58:164–172, 1995
Zanarini MC (ed): Evolving perspectives on the etiology of borderline
personality disorder, in Role of Sexual Abuse in the Etiology of Bor-
derline Personality Disorder. Washington, DC, American Psychiatric
Press, 1997, pp 1–14
Zanarini MC, Gunderson JG, Marino MF, et al: Psychiatric disorders in
the families of borderline outpatients, in Family Environment and
Borderline Personality Disorder. Edited by Links PS. Washington,
DC, American Psychiatric Press, 1990, pp 67–84
Zinner J, Shapiro ER: Splitting in families of borderline adolescents, in
Borderline States in Psychiatry. Edited by Mack J. New York, Grune &
Stratton, 1975, pp 103–122
Chapter 9
GROUP THERAPY
THE ROLE OF GROUP THERAPIES in the management of patients with

borderline personality disorder (BPD) is undergoing considerable revi-
sion. Previously, group therapies targeted the borderline patients’ dis-
turbed interpersonal relationships. These groups use an open-ended and
unstructured approach, and they usually include patients with a mixture
of diagnoses. More recently, group therapies have targeted the deficits
that characterize BPD and have emphasized skills training as their major
function. The current literature is about group therapies that are time-
limited, carefully structured, and typically uniform with regard to the in-
clusion of only patients with BPD.
Despite these developments, group therapies are still most commonly
started during level III (residential, partial or day hospital) or during
level II (intensive outpatient) programs. Clinicians, particularly those
working in urban centers, often have access to skills training groups that
might be helpful for their borderline patients. In this chapter, I describe
the group content, structures, and process that are recommended for pa-
tients with BPD. Most comments are applicable to group therapies in
general; however, because of the recent focus on skills training groups,
particular reference is made to the application of these groups to patients
with BPD. To help engage the resistant patient, an approach to use with
BPD patients is presented. The chapter concludes by anticipating some
of the more common problems that can arise when patients with BPD
participate in group therapies.
Indications
For a variety of reasons, group therapy may be particularly indicated for
patients with BPD. Group therapies offer borderline patients opportuni-
ties to observe their maladaptive interpersonal interactions and to learn
207
new ones; they may be more accepting of peer feedback than feedback
from the therapist alone; peer pressure can help set limits on impulsive
behavior patterns; and group therapy can help dilute the transference re-
lationship and promote identification with other patients (Munroe-Blum
1992). Even patients who are resistant to attending group therapy often
will find it helpful.
Despite the possible value of group therapy, this therapy does not al-
ways work. In our experience, extremely narcissistic patients who are un-
able to share the spotlight with others and are constantly seeking the time
and attention of the group leaders can have a negative influence. These
patients will be ostracized because of their dismissive and arrogant atti-
tudes to others in the group. Also, patients with comorbid psychotic ill-
nesses such as schizoaffective disorder and major depressive disorder,
with psychotic features, should be referred to such groups only if they
have had an extended period of stability without psychotic symptoms.
Even then, the intense feelings typical of involvement with borderline
group members may exacerbate their psychotic features. Patients with ac-
tive substance abuse problems or severe eating disorders are probably
best serviced by referral to specialized programs until these issues have
been stable for a few months. Although many of these specialized pro-
grams use group therapy and teach relevant skills, substance abuse and
eating disorders have implications for the person’s psychological and
physical health that need to be addressed as priorities. Patients who are at
risk for interpersonal violence when in crisis or are highly intimidating to
others are not good group candidates because those group members who
are survivors of prior trauma will feel too unsafe to be engaged.
Many men with BPD benefit from group therapies, although Black et
al. (2004) suggested that two or more men should be required. Other-
wise, the single male participant can become the recipient of all the neg-
ative projections for all other men in the world.
Engaging Patients and the Primary Clinician’s Role

To have the borderline patient engage in group therapy can be difficult.
To get borderline patients engaged, the clinician should take time to ex-
plain why he or she is urging this. The reasons for the borderline patient
to engage in group therapy overlap considerably with Yalom’s (1995) ge-
neric description of what can make group therapy useful for anyone. For
example, we would say to prospects:
You should know that you can learn things in groups that you can’t learn
in individual psychotherapy. Specifically, you can learn that others have
similar problems and learn how they cope with them. You can also learn
Group Therapy ❘ 209
how you unknowingly impede making the close relationships that you
want, and you can work in the group on changing those patterns. More-
over, you can learn some things faster in groups than you can in individual
therapy. For example, you can learn to listen when people express feel-
ings you usually can’t stand, and you can learn to understand why people
have those feelings.
Patients are usually receptive to the idea that it is easier to understand

or recognize a problem when it is observed in others and that this aware-
ness may help them understand or recognize their own problems.
Participation in a group makes it more likely that what borderline pa-
tients learn in psychotherapy will be more effective. Individual therapists
can easily underestimate the stressful effects of things they say or do.
Thus, as noted in the earlier discussion of split treatment (Chapter 4),
groups offer a useful way to process such stress. Conversely, the group will
undoubtedly prove stressful at times, and when that occurs, individual
therapists can help patients learn from those experiences.
The group leader and primary clinician or individual therapist must
be free to talk to each other, although borderline patients may try to ob-
struct this interaction. Indeed, this dialogue should be assumed. The pri-
mary clinician’s being familiar with the content of the concurrent group
intervention will facilitate the patient’s work both inside and outside the
group. The primary clinician has the important role of helping patients
generalize their skills to settings outside of the group and promoting the
commitment to practicing what they learn.
Skills Training Groups

In recent years, four skills training group therapies have been developed
that target specific deficits that characterize BPD (Table 9–1). The best
known of these—dialectical behavior therapy (DBT)—is a more compre-
hensive treatment for BPD that is elaborated on in Chapter 11. Here we fo-
cus on the skills group component. The Systems Training for Emotional
Predictability and Problem Solving (STEPPS) is also thought to address
core behavioral and emotional BPD issues, but it is designed to be coordi-
nated with usual community treatments, and it does not have an individual
therapy component (Black et al. 2004). In Toronto, Ontario, we developed
a group intervention for BPD (and other patients) that focuses on recur-
rent suicidal behavior called the Psychosocial/Psychoeducational Group
Intervention for People With Recurrent Suicide Attempts (PISA; Berg-
mans and Links 2002). A less comprehensive and more time-limited skills
training group, the Acceptance-Based Emotion Regulation Group Inter-
vention (ABERG), has been developed by Gratz and Gunderson (2006).
210
TABLE 9–1. Skills training modules: similarities and differences
CORE ELEMENTS DBT PISA STEPPS ABERG
❘
Psychoeducation Education Awareness of illness Psychoeducation about

• About diagnoses • Learn about BPD emotions (week 1)
• Providing a language • Maladaptive cognitive
filters, behaviors, and
feelings
Emotion awareness Mindfulness Emotional literacy Emotional management Awareness, understanding,

• Focusing on the moment • Identification of skills training and acceptance of emotion
• Awareness with judgment emotions • Distancing (weeks 2–6)
Distress tolerance • Scale of intensity • Communicating Willingness to experience
• Crisis survival strategies • Behavior as a choice emotion negative emotions
• Radical acceptance of • Challenging distortions (weeks 7, 8)
reality (“yes” rather than and negative thoughts
• Distracting Modulate intensity and
“yes, but”) duration of emotional
• Managing problems
Emotion regulation responses
• Observe and identify Ability to engage in goal-
emotional states directed behavior and
• Validate and accept one’s inhibit impulsive behavior
emotions (weeks 9, 10)
• Decrease vulnerability to
negative emotions
• Increase experience of
positive emotions
Group Therapy
TABLE 9–1. Skills training modules: similarities and differences (continued)
CORE ELEMENTS DBT PISA STEPPS ABERG
Interpersonal Interpersonal effectiveness Interpersonal Behavior management
relationships • Assertiveness training relationships; borrows skills
• Cognitive restructuring from Linehan • Goal setting
• Balancing objectives with Personal safety • Leisure
maintaining relationships • Able to identify what • Physical health
and self-esteem is safe • Abuse avoidance
• Crisis survival strategies includes constructing
• Network of support emotional intensity
continuum
• Interpersonal
relationships includes
problem solving, setting
boundaries
Problem solving Problem solving Included abovea

Teaching and using
problem solving
Goal setting Included abovea Identify and clarify valued

directions (weeks 11–14)
Note. ABERG=Acceptance-Based Emotion Regulation Group Intervention; BPD=borderline personality disorder; DBT= dialectical behavior
therapy; PISA= Psychosocial/Psychoeducational Group Intervention for People With Recurrent Suicide Attempts; STEPPS= Systems Training
❘
for Emotional Predictability and Problem Solving.
211
aIn Core elements, see Interpersonal relationships for STEPPS.
In each of these group therapies, borderline patients are considered

to have deficits in regulating their emotions and/or behaviors. These
group therapists approach patients in a collaborative and didactic man-
ner and offer skills to help regulate emotions and behavioral responses,
to cope with crises, and to improve interpersonal functioning. Specifi-
cally, as summarized in Table 9–1, three of the four groups start with a
module dedicated to educating participants about the features of the di-
agnosis. All the group interventions focus on teaching participants skills
related to emotional awareness, expression, and acceptance. This con-
tent reflects the important shift in therapies. Patients are encouraged to
learn how to process emotions rather than to control emotions. Much of
this work is drawn from aspects of emotion-focused psychotherapy
(Greenberg 2001). The groups typically include a module on skills to in-
crease interpersonal effectiveness—an important part of which is learn-
ing to identify what “safe” means and how this safety can be obtained in
the person’s external or internal worlds. Several of the groups specifically
teach problem-solving skills and encourage participants to develop a per-
sonal method of scaling their distress to better communicate their needs
to their support network. The primary clinician should become familiar
with patients’ personal scaling methods if these are an outcome of group
involvement. These scales must become a living document that will be re-
vised and reworked as patients better understand themselves.
DBT Skills Groups

DBT was the original example of skills training group therapy. In this
model, suicidal and self-harm behaviors are understood as dysfunctional
attempts to solve the problem of the patient’s dysregulated emotions.
The suicidal behaviors exist to help regulate emotional arousal and/or to
elicit more nurturing responses from those in the person’s environment
(see Chapters 1 and 4). The core elements of the skills training groups in
DBT are indicated in Table 9–1. Primarily on the basis of a single study
(Shearin and Linehan 1994), Linehan has considered DBT groups with-
out the other program elements to be of little benefit and has discour-
aged the use of individual elements of DBT treatment on their own. Still,
such groups are not uncommon and, in our clinical experience, are fre-
quently very useful.
STEPPS
Reflecting agreement with the premise of DBT, Blum and colleagues
(2002) at the University of Iowa likewise understood BPD to represent
core difficulties in emotional and behavioral regulation. However, these
investigators wanted to develop an intervention program that would take
advantage of participants’ existing social and professional support sys-

tems. Their group therapy program set out to train the participants in
specific emotion and behavior management skills and then have the pa-
tient’s support system act as a “reinforcement team” for the participant’s
newly learned skills. The resultant complete program involves a 20-week
basic skills group called STEPPS and a 1-year twice-monthly extension
group program called STAIRWAYS (Setting goals, Trusting and taking
risk, Anger management, Impulsivity control, Relationship behavior,
Writing a script, Assertiveness training, Your choices, Stepping out).
The STEPPS program involves three components: 1) the awareness of
illness component teaches about the thoughts, feelings, and behaviors
that define BPD; 2) emotion management skills training covers five basic
skills to help the participant predict, anticipate, and manage emotional
intensity episodes; and 3) behavior management skills training involves a
variety of skills such as goal setting, sleep hygiene approaches, and inter-
personal relationship skills (Black et al. 2004). The STEPPS program en-
courages participants to rely on their supports, both professional and
nonprofessional, to reinforce their new skills. Blum et al. (2002) reported
early promising results of their program: participants showed a signifi-
cant decrease in symptoms associated with BPD at the completion of the
20-week group. The intervention seemed most helpful for reducing neg-
ative behaviors and affect (Blum et al. 2002). A randomized controlled
trial of the STEPPS program is under way that will help assess the pro-
gram’s effectiveness.
PISA
PISA was developed to serve an inner-city population who often are un-
derhoused, underemployed, and undereducated; stresses client valida-
tion and participation; and is intended to be an adjunct to individual
psychotherapy and to include patients receiving pharmacotherapy. The
group intervention is meant to work in tandem with the participants’ in-
dividual therapy, and therapists are able to attend evening sessions to fa-
miliarize themselves with the group content. Its goals are to decrease the
duration, intensity, and frequency of suicidal behaviors by developing an
awareness of and language for the emotional experiences occurring dur-
ing a crisis. This focus is based on the theory that recurrently suicidal pa-
tients have deficits in their capacity to identify and describe emotions
(i.e., alexithymia; Taylor 2000). This deficit can be established by scoring
higher than 51 on the Toronto Alexithymia Scale (Taylor 2000). The
group involves educational modules, skills related to developing emo-
tional literacy, relationship management skills, and problem-solving
skills. Preliminary data indicated decreases in the Toronto Alexithymia
Scale score over the 20-week intervention, which supports the proposed
mechanism of action (Links et al. 2004).
ABERG
Like STEPPS and PISA, ABERG therapy was developed to be an “eco-
nomically and clinically feasible” intervention in settings where more in-
tensive therapies are not available or an adjunct to individual outpatient
therapy. Gratz and Gunderson (2006) conceptualized the intervention as
helping patients with BPD increase their acceptance of emotions (espe-
cially negative emotions), control their impulsive behaviors, and use
strategies to modulate their emotional responses. The intervention uses
acceptance-based strategies and DBT strategies and also includes sessions
on identifying and choosing actions consistent with the person’s valued
directions. The last weeks of the groups work on identifying and clarify-
ing valued directions. This content highlights the difficulty patients with
BPD have in living their lives in keeping with their personal values. In a
randomized controlled trial, Gratz and Gunderson (2006) found signifi-
cant effect of the group intervention on self-harm, emotional dysregula-
tion, experiential avoidance, and BPD-specific symptoms compared with
treatment as usual. In the study, 42% of participants showed a clinically
significant reduction of self-harm of 75% or greater. The authors need to
carry out larger-scale randomized controlled trials and study the mainte-
nance of gains over longer follow-up periods.
Psychodynamic Group Therapies

Interpersonal Group Psychotherapy
Psychodynamically oriented long-term group therapy is probably the earli-
est form of group therapy for patients with BPD (Roller and Nelson 1999).
Typically, these groups are open-ended and consist of patients with a variety
of diagnoses but include one to four patients with BPD. The group is seen
as providing these patients with a safe interpersonal holding environment
that allows a lessening of pathological defenses such as splitting and projec-
tive identification and a gradual maturational process (Leszcz 1989; Ver-
heul and Herbrink 2007). A recent review by Verheul and Herbrink (2007)
concluded that research supports the effectiveness of interpersonal group
psychotherapy for patients with personality disorders, noting that it was an
effective follow-up after day hospital treatment. They cautioned against
prescribing psychodynamic outpatient group therapy alone as a treatment
for severe personality disorders. Specifically, interpersonal group psycho-
therapy should be used with caution for BPD patients with limited frustra-
tion and anxiety tolerance, with significant impulsivity and poor reality
testing, or with limited ability to mentalize. Sidebar 9–1 discusses some ben-
efits and limitations of interpersonal group psychotherapy.
Sidebar 9–1: Research on Interpersonal Group

Therapy With BPD
Marziali and Munroe-Blum (1994, 1995; Munroe-Blum and
Marziali 1995) carried out one of the only randomized con-
trolled trials of interpersonal group psychotherapy. The in-
terpersonal group psychotherapy was based on a model
of therapy called relationship management (Dawson 1988;
Dawson and MacMillan 1993) and emphasized empathy
and deemphasized interpretation. The 38 borderline outpa-
tients randomly assigned to interpersonal group psycho-
therapy were compared with 41 others assigned to weekly
individual psychodynamic therapy that, insofar as it was
modeled after that described by Kernberg (1968, 1975,
1986; see Chapter 12), emphasized interpretations. The
therapists’ training and experience were equal.
Both groups showed clinically significant but similar lev-
els of improvement at 1 year, and both groups sustained
those improvements during follow-up. Improvements were
most dramatic in behaviors (e.g., hospitalization, suicidal
acts, impulsivity), to a lesser degree for depression, and
least for social adjustment. The investigators did not believe
that the patients’ overall “character pathology” was much
affected. Perhaps the most instructive lesson from this study
was that few of the interpersonal group psychotherapy
subjects actually received the intended treatment. This
study highlights, at least, that borderline patients typically
resist group involvement.
Mentalization-Based Group Therapy

Another form of group therapy combines a psychoanalytic theory with a
skill-building approach. Here the group focuses on learning the psycho-
educational skill of mentalization (Tobias et al. 2006). The intent of men-
talization group therapy is to enhance the overall treatment offered and
to give patients a common language to use with their clinicians. The di-
dactic element of the group involves teaching: the concept of mentaliza-
tion, how psychiatric illness and impaired mentalization interact with
each other, and how treatment heightens the capacity to mentalize. The
group participants use experiential exercises to practice the skill of men-
talizing; for example, patients report a recent significant interaction with
“just the facts” and no mention of their emotional state. The group mem-
bers attempt to infer the person’s mental state from these facts, and the
person who provided the example responds to the various inferences.

Tobias and colleagues offer the 50-minute group sessions weekly for 6–8
weeks, and up to 24 patients can attend each session. At this time, there is
no formal evaluation of this psychoeducational program; however, given
the interest in mentalization-based treatment, further reports are antici-
pated (see Chapters 3, 5, 11, and 12).
Group Structure
Size
Four members is a useful minimal group size to keep in mind for getting
started. Having four members translates into the likelihood of having
three in attendance, and although starting with only three attendees may
be hard to justify in cost-benefit terms, it has advantages over a long delay
while waiting for enough members who are ready to join. Borderline pa-
tients are always ambivalent about joining groups, and delay can easily
flip the balance toward not starting. Six to eight members is optimal. A
group larger than 10–12 does not permit enough individual activity to
keep everyone engaged.
Length
Meetings are once or twice weekly and last for 1–1.5 hours. Twice a week
offers the group more of a holding function, but this frequency is rarely
feasible because of the ambivalence that typifies most borderline pa-
tients’ motivation for group. Groups should meet in early evenings to di-
minish conflicts with vocational activities.
Duration
Recent evidence suggests that patients with BPD can make substantial
changes in their self-harm behavior (Gratz and Gunderson 2006; Wein-
berg et al. 2006) in a short time. Most of the skills training groups dis-
cussed earlier are no more than 20 weeks in duration, but DBT typically
lasts for 12 months; however, the efficacy of DBT programs of shorter du-
ration is being evaluated. In the PISA program, many patients have ap-
plied to repeat the group therapy because they feel more time is needed
to reach their goals from the group. The STEPPS program allows patients
to repeat the basic program (20 weeks) and also offers patients an ex-
tended program (for 1 year) that reinforces the skills learned and teaches
additional skills such as assertiveness training and goal setting. These
time frames differ from those that are usually beneficial for a more inter-
personally focused group therapy (interpersonal group psychotherapy).
Given that the interpersonal group psychotherapy primary goals are so-
cial and interpersonal, the benefits can be expected to require a mini-

mum of 4 months but will usually accrue for 1.5–2 years (see Chapter 3).
After 2 years, the rate of added benefit often diminishes.
Because of the socializing function that interpersonal group psycho-
therapies can serve for borderline patients, it is best not to limit continued
involvement, even if it seems to involve little new learning. The group ther-
apy aftercare given at Ulleval Hospital in Norway had a mean duration of 1
year, but as experience with this modality grew, this duration seemed
“rather short” (Wilberg et al. 1998, p. 218). The duration suggested here
(1.5–2 years) extends that suggested by Marziali and Munroe-Blum (1994).
At the start of their project, they, like Piper and Josie (1998), believed that
having a pre-established time limit (they thought 30 sessions) for the inter-
personal group psychotherapy would “accelerate the achievement of im-
portant changes” (p. 689). However, after their project, they concluded that
a longer period in interpersonal group psychotherapy (at least a year)
might offer further advantages—especially for a subgroup they termed
pseudocompetent patients, meaning the borderline patients who were intellec-
tually well defended. In summary, the clinician should anticipate the dura-
tion of group therapy to be 2–6 months for skills training groups, 12 months
for DBT, and 18–24 months for interpersonal group psychotherapy.
Leadership
The leaders of groups of patients with BPD should have some general and
some specific attributes. First and foremost, the leader must be comfort-
able and curious about working with these patients. Requiring a clinician to
lead these groups when the clinician is not interested in or comfortable
leading groups for patients with BPD is usually detrimental for everyone.
The group leader should have at least a master’s degree and should be well
trained in the particular approach undertaken. As with other modalities
(e.g., pharmacotherapy, individual therapies, family therapy), prior clinical
experience and especially experience with case management are valuable
assets. The risks of impulsive, self-destructive, or inappropriate (boundary-
violating) relationships are significant even with experience and supervi-
sion. Having experience in conducting groups in either inpatient or resi-
dential settings is the safest way for clinicians to learn about limit setting,
safety assessments, communication with other team members, and other
administrative roles needed to work comfortably as an outpatient group
therapist with borderline patients. Clinicians who have acquired these skills
can combine them comfortably with the usual supportive, interpretive, and
other facilitative functions needed for good group therapy leadership.
(Such learning is also needed to be able to function as a borderline pa-
tient’s primary clinician or therapist; see Chapters 5 and 10.)
To conduct a skills training group, the leader should be comfortable

with the role as educator and be very familiar with the content and format
of the group. Black et al. (2004) warned that in a skills training group, the
focus is on fostering skills rather than “putting out fires.” This same issue
applies to all BPD group therapies. A common problem that must be
managed is a participant’s wish to recall in painful detail previous suicidal
and self-harm attempts. Some participants do this, even though they have
been warned that the purpose of the group is to learn from previous cri-
ses rather than to be absorbed in the events. This content only serves to
emotionally escalate all the group members, and the leader will have to
refocus the participants on what can be learned from the example.
Having a cotherapy team lead groups has several benefits. Here, too,
the principle of split treatment (Chapter 4) is at work: the patient who is
angry at one therapist can preserve a “good” image for the other, which
makes flight less likely. Cotherapists help to check each other’s counter-
transference reactions, maintain continuity of treatment during absences,
and otherwise decrease the burden of leading groups (Greenbaum and
Pinney 1982). For example, if a group participant is in immediate crisis
and needing intervention, then one cotherapist can remove the person
from the group and help find appropriate assistance with the current cri-
sis. As with other modalities, supervision for group therapists is recom-
mended. As discussed in Chapter 4, supervision provides support to the
leaders, provides opportunities to express negative reactions and emo-
tions, and helps the therapy remain adherent to its original model.
Common Problems
Group participation can lead to some challenging problems for patients
with BPD and their therapists. The primary clinician needs to be aware of
these. The following examples describe three common problems (i.e.,
disruptive behaviors, outside-of-group contacts, and silence) and guiding
principles to assist the clinician in their management.
Disruptive Group Behaviors

The patient complains in individual therapy that the group is driving her
crazy and making her worse. Principle: Always keep an eye on the health of the
group. Individual therapy is not the goal or purpose of the group.
Vignette
Ms. S joined a skills training group because she wanted strategies to cope
with her high levels of impulsivity. She immediately struggled with all the
elements of learning within a group. She could hardly sit still during ses-
sions, she randomly spoke up and interrupted other members, and she
could not attend to any of the learning materials. Ms. S, other group
members, and the group leaders were all totally exasperated by these be-
haviors. Although her primary clinician initially lobbied hard for her
continued membership in this group, the group leaders and primary cli-
nician eventually concluded that the health of the group took priority.
Ms. S accepted that she needed one-to-one coaching on the skills and gra-
ciously bid the group farewell.
There must be the opportunity for open communication between the

primary clinician and the group leaders. A group leader and an individual
therapist (or primary clinician) do not always need to talk to each other
about a shared patient, but this option must be open. Actual communica-
tion is needed when the patient has attendance problems, experiences
safety risks, or endangers the usefulness of either therapy (e.g., drug use, an
affair with another group member). Occasionally, a borderline patient is in
group therapy with no other therapy, but this option is not recommended
and has inherent problems for patients who practice dangerous self-destruc-
tive behaviors. The group therapist may be forced to take on a responsibility
for that patient’s safety that will be very disruptive to the group’s function.
Usually, borderline patients in a group will benefit from having a primary
clinician who serves as the therapist. Here it is critical that the group leader
and the primary clinician or therapist respect the value of each other’s mo-
dality so that when either party is told by the patient how stupid or insensi-
tive the care in that other modality has been, the therapists are neither
sympathetic nor protective (see section “Splits, Splitting, and the Virtues of
Split Treatments” in Chapter 4). Actually, one of the potential misuses of
communication between a therapist and a group leader would be to report
to each other when a borderline patient says anything critical or devaluative
about the other. Such reactions may be experienced as defensive, validating
the borderline patients’ fears that their anger is dangerous and perhaps also
undermining the need for the separateness of the two modalities.
Outside-of-Group Contacts
The patient reports in individual therapy of having outside-of-group con-
tact with a co-patient. She has not yet disclosed this contact to her group.
Principle: Having outside-of-group contacts must be discussed within the group.
The participants must feel safe within the group, and rules regarding
confidentiality, outside-of-group contact, and appropriate behavior must
be reinforced by both the group and the individual therapists.
Vignette
Ms. T told her individual therapists between group sessions that her co-
patient, the quiet and shy Mr. U, showed up unexpectedly at her apart-
ment door intoxicated. Although Mr. U’s purpose in visiting was to “invite
her to church on Sunday,” Ms. T felt quite threatened and uncomfort-
able. To remove him from her front door, she agreed to have a coffee with
him. But later, she could only extricate herself from the encounter by tak-
ing herself to the hospital emergency department, pleading with Mr. U
that she was in crisis. Her therapist insisted that she needed to discuss this
interaction at her next group session. In the subsequent group sessions,
the outside-of-group contact was discussed openly, and the need to re-
member that everyone has the right to be safe was addressed. Many rele-
vant skills were examined related to the encounter, including how Mr. U
could be more interpersonally effective with a woman whom he liked.
Of course, the situations of outside-of-group contact can be even more

complex, as the following example illustrates:
Vignette
The following dialogue transpired in the last few minutes of the thir-
teenth group meeting (13 out of the 20 scheduled sessions):
Ms. V: “I know because of the rules… I have to mention this. Mr. W
and I have been meeting for coffee after group.”
Mr. W: Nods in agreement but is silent.
(Group is silent)
Ms. V: “We have been talking together between meetings and trying to
help each other with our skills.”
(Silence)
Ms. V: “Well, we haven’t just been talking together.. . we have been
sleeping together.”
(Longer silence)
First cotherapist: “I appreciate you sharing this with the group and
keeping to our rules.”
(Silence)
Second cotherapist: “We are just about out of time for today, but
I trust you feel that this is a safe decision for you both. Certainly, I have
noted in the group that you and Mr. W are always in agreement. Maybe we
should give the group time to discuss what’s been raised next week.”
In the postgroup discussion, the cotherapist team was at loggerheads
with each other. One wanted to kick the pair out of the group, whereas
the other thought that they should be commended for keeping to the
rules. Cooler heads finally prevailed, and the cotherapists agreed to stick
to their rules: try to ensure that the group members were safe; that con-
fidentiality was maintained; and that outside-of-group contact, even sex-
ual, be discussed with the group members. Leaving their long debriefing
session, the cotherapists felt united and resolved to control what they
could reasonably control.
The propensity for BPD patients to want to meet outside the group ses-
sions invariably creates tension for the groups and individual therapists.
In the worst case, borderline patients who meet outside the group de-
velop covert alliances that then exclude others and create an intimidating
team within sessions. In the best case, such outside-of-group socialization
addresses the facts that BPD patients usually have few friends and that the
experience of psychiatric care, especially hospitalization, has added to
their alienation from community-based peers. Because of the potential
value of social networking as an outgrowth of group participation and the
potential harm from a control struggle, it is usually unwise for therapists
explicitly to discourage outside contact. This principle is counter to the
expectations for group therapies that are trying to teach interpersonal ef-
fectiveness; but the principle needs to make accommodations for some of
the unique aspects of borderline patients. Instead, it is best to identify ex-
plicitly how such outside contacts can create group problems (e.g., “It is
very hard to say anything in the group that [you know] someone doesn’t
want to hear if you risk a friendship because of it”). It is realistic and suf-
ficient to expect that group members will identify and be prepared to dis-
cuss any significant contacts they have outside groups.
Silence
The patient reports to her therapist that she has only uttered her name
since starting group. Principle: The primary clinician needs to remember the
goals for advising the patient to join the group and to determine if the resistance is
related to group therapy or the individual therapy.
Vignette
Ms. X continued to attend her group sessions as advised by her individual
therapist, but she had hardly spoken a word over the 3 months. Her indi-
vidual therapist was frustrated with Ms. X’s participation; however, she
had expected that Ms. X would find group therapy very intimidating. The
opportunity for dialogue with the group therapists was already in place,
and the individual therapist made an effort to catch them before Ms. X’s
next group. The group therapists were very encouraged by Ms. X’s in-
volvement with the group. Although she had been almost mute, Ms. X was
very nonverbally involved in the group and often provided the group
lengthy written feedback on her personal diary cards that were completed
between groups.
Ms. X was very quietly participating in the group therapy, and no re-
medial action was necessary. However, a patient can be silent for many dif-
ferent reasons, including active resistance to the initial referral for group
therapy. If this is believed to be the reason, then the primary clinician
should address the issue as a negative transference reaction that must be
dealt with in the individual sessions.
Summary
Group therapies for patients with BPD have moved toward psychoeduca-
tional and skills training groups; however, more traditional interpersonal
group psychotherapy can still be an important adjunct during level I pro-
grams. Skills training groups have shown promise to foster improvement
in the participants’ behavior problems over relatively short interventions.
The primary clinician should encourage and sometimes insist on the pa-
tient’s participation in group therapy, but the clinician needs to be famil-
iar with the group modality to assist in the generalizing of benefits to
outside the group and anticipate certain problems that may arise when
patients with BPD are group participants.
References
Bergmans Y, Links PS: A description of a psychosocial/psychoeduca-
tional intervention for recurrent suicide attempters. Crisis 23:156–
160, 2002
Black DW, Blum N, Pfohl B, et al: The STEPPS Group Treatment Pro-
gram for outpatients with borderline personality disorder. Journal of
Contemporary Psychotherapy 34:193–210, 2004
Blum N, Pfohl B, St. John D, et al: STEPPS: a cognitive-behavioral sys-
tems-based group treatment for outpatients with borderline person-
ality disorder—a preliminary report. Compr Psychiatry 43:301–310,
2002
Dawson D: Therapy of the borderline client, relationship management.
Can J Psychiatry 33:370–374, 1988
Dawson D, MacMillan HL: Relationship Management and the Borderline
Patient. New York, Brunner/Mazel, 1993
Gratz KL, Gunderson JG: Preliminary data on an acceptance-based emo-
tion regulation group intervention for deliberate self-harm among
women with borderline personality disorder. Behav Ther 37:25–35,
2006
Greenbaum DN, Pinney EL: Some comments on the role of cotherapists
in group psychotherapy with borderline patients. Group 6:41–47,
1982
Greenberg LS: Emotion-Focused Therapy: Coaching Clients to Work
Through Their Feelings. Washington, DC, American Psychological
Association, 2001
Kernberg O: Borderline Conditions and Pathological Narcissism. New
Kernberg O: Severe Personality Disorders: Psychotherapeutic Strategies.
New Haven, CT, Yale University Press, 1986
Leszcz M: Group psychotherapy of the characterologically difficult pa-

tient. Int J Group Psychother 39:311–335, 1989
Links PS, Bergmans Y, Barr A: Affective deficits in recurrent suicidal be-
haviour. Symposium presented at the 157th annual meeting of the
American Psychiatric Association, New York, May 3, 2004
Marziali E, Munroe-Blum H: Interpersonal Group Psychotherapy for
Borderline Personality Disorder. New York, Basic Books, 1994
Marziali E, Munroe-Blum H: An interpersonal approach to group psy-
chotherapy with borderline personality disorder. J Personal Disord
9:179–189, 1995
Munroe-Blum H: Group treatment of borderline personality disorder, in
Borderline Personality Disorder: Clinical and Empirical Perspec-
tives. Edited by Clarkin JF, Marziali E, Munroe-Blum H. New York,
Guilford, 1992, pp 288–299
Munroe-Blum H, Marziali E: A controlled trial of short-term group treat-
ment for borderline personality disorder. J Personal Disord 9:190–
198, 1995
Piper WE, Josie JS: Group treatment of personality disorders: the power
of the group in the intensive treatment of personality disorders. Psy-
chotherapy in Practice 4(4):19–34, 1998
Roller B, Nelson V: Group psychotherapy treatment of borderline per-
sonalities. Int J Group Psychother 49:369–385, 1999
Shearin EN, Linehan MM: Dialectical behavioral therapy for borderline
personality disorder: theoretical and empirical foundations. Acta
Psychiatr Scand Suppl 379:61–68, 1994
Taylor GJ: Recent developments in alexithymia theory and research. Can
J Psychiatry 45:134–142, 2000
Tobias G, Haslam-Hopwood G, Allen JG, et al: Enhancing mentalization
through psychoeducation, in Handbook of Mentalization-Based
Treatment. Edited by Allen JG, Fonagy P. Chichester, England, Wiley,
2006, pp 249–267
Verheul R, Herbrink M: The efficacy of various modalities of psychother-
apy for personality disorders: a systematic review of the evidence and
clinical recommendations. Int Rev Psychiatry 19:25–38, 2007
Weinberg I, Gunderson JG, Hennen J, et al: Manual assisted cognitive
treatment for deliberate self-harm in borderline personality disorder
patients. J Personal Disord 20:482–492, 2006
Wilberg T, Friis S, Karterud S, et al: Outpatient group psychotherapy: a
valuable continuation treatment for patients with borderline person-
ality disorder treated in a day hospital? A 3-year follow-up study. Nord
J Psychiatry 52:213–221, 1998
Yalom ID: The Theory and Practice of Group Psychotherapy, 4th Edition.
New York, Basic Books, 1995
Chapter 10
INDIVIDUAL PSYCHOTHERAPIES
Getting Started
Introduction: Prerequisites
Without question, individual psychotherapies have been the cornerstone of
treatments for borderline personality disorder (BPD). One study showed
that more than 90% of BPD patients who received any treatment had indi-
vidual psychotherapy (Bender et al. 2001). In that study, the mean length of
time in BPD patients’ prior psychotherapy was 51 months. Although some
of what these patients called psychotherapy doubtless included what in this
book is referred to as primary clinician functions or case management (see
Chapter 4), Bender and colleagues’ finding is all the more remarkable for
having occurred within a managed care environment, where such lengths
of treatment are discouraged.
Psychotherapy, as used here, refers to a modality that is not primarily
designed to relieve symptoms or diminish self-destructive or otherwise
maladaptive behaviors. Psychotherapy is designed to help patients psy-
chologically change for the better—to alter maladaptive psychological ca-
pacities or to develop new psychological ones. As such, therapies differ
from treatments (such as medication, diet, or hospitalization), which pa-
tients can passively receive or resist but do not require their active collab-
oration. Psychotherapies require shared goals and at least intermittent
collaboration. These requirements that constitute readiness for psycho-
therapy (Table 10–1) often develop out of discussions with someone who
had a case manager or primary clinician role. Establishing sustained col-
laboration (a working alliance) is itself an achievement for many border-
line patients (see Chapter 3). Many problems in psychotherapies with
borderline patients derive from having begun without having first estab-
225
TABLE 10–1. Readiness for psychotherapy

PATIENT THERAPIST
Sees problem in self Trained to competence in skills

Seeks change in self that facilitate change
Patient (or others) can assume Agrees with patient’s goals
primary responsibility for safety Can help patients contain
emotions and stay safe
lished shared goals and collaborative intentions. It is not that borderline

patients who are not ready for individual therapy are untreatable; they
need other, less-demanding forms of treatment.
Individual psychotherapies may begin while borderline patients are
in hospital, residential, or intensive outpatient programs. However, in the
context of these higher levels of care (Chapters 4 and 5), the role of in-
dividual psychotherapy is often adjunctive and is secondary to case man-
agement goals, which include symptom relief and behavioral control.
These settings provide a useful context for developing the contractual and
role forms of alliance (as described in Chapter 3).
In reference to the five therapeutic functions described in Chapter 3
of this book, all psychotherapies involve nonspecific forms of support
(e.g., concerned attention and empathy). The different types of psycho-
therapy vary in their level of structure, and borderline patients require
more structure than do other patient types. Psychotherapists need to be
particularly attentive to developing agreed-on roles and goals, bound-
aries, and the like. The therapeutic processes of involvement and valida-
tion (as described in Chapter 3) distinguish individual psychotherapies
from other modalities. The sustained involvement with a trustworthy and
caring other offers a powerful corrective attachment experience (dis-
cussed in Chapters 11 and 12). The validation process is particularly crit-
ical to the subject of this chapter—helping borderline patients become
engaged—although validation as a technique that helps them own (i.e.,
experience as part of who they are) feelings and motives is also a process
in all types of effective longer-term psychotherapy.
Little is known about what characteristics distinguish borderline pa-
tients who are suitable candidates for psychotherapy and what subtypes of
BPD are best suited for the different forms of psychotherapy currently
available. Almost all psychotherapies require a verbal and conceptual
ability to manage abstractions, metaphors, irony, and irreverence. In a ret-
rospective review of 299 patients with borderline personality organization
(see Chapter 1), of whom 206 had DSM-III (American Psychiatric Associ-
Individual Psychotherapies ❘ 227
ation 1980) BPD, who had all received intensive (three or more sessions a
week), psychoanalytically oriented therapies, Stone (personal communi-
cation, April 1999) considered 132 of the patients ideally amenable to this
modality. A disproportionate number of the less amenable were among
the 206 patients who met criteria for BPD. Stone would add serious sub-
stance abuse to Kernberg’s list of poor amenability factors, and like me,
he would emphasize the issue of motivation cited in Table 10–1.
Getting Started
The Problem of Dropouts
A series of studies initiated in the 1980s documented a very high dropout
rate from individual psychotherapies by borderline patients. Skodol et al.
(1983) found that 67% of the borderline patients dropped out of indi-
vidual psychotherapy in 3 months. In the Treatment of Depression Study,
40% of the Cluster B patients dropped out within 16 weeks (Shea et al.
1990).
Several other studies done in McLean Hospital’s research program
have underscored this problem. In a study of 60 borderline patients who
were beginning individual psychotherapies at McLean Hospital (Gun-
derson et al. 1989), we found that 42% dropped out within 6 months.
The most common reasons were 1) too much frustration, 2) lack of fam-
ily support, and 3) logistics (travel, time, costs). As Yeomans et al. (1993)
noted, the rate of dropouts in our study was lower than can reasonably be
expected in outpatient settings because some of our samples were hospi-
talized for all or most of the initial 6 months. It is of note that our health-
ier BPD patients were more likely to stay in psychotherapy if it was started
in outpatient settings (not overly controlled), whereas the more severe
BPD patients were more likely to remain in psychotherapies that were
started during inpatient stays. In another study (Waldinger and Gunder-
son 1984), we surveyed senior and expert therapists who had contributed
to this literature. We discovered a similar pattern. This survey indicated
that in office practice, even the experts have problems keeping border-
line patients engaged in psychotherapies: of 790 borderline patients,
54% continued psychotherapy beyond 6 months, and only one-third
(33%) went on to complete their therapy satisfactorily. These studies
have made it clear that engaging borderline patients in individual psy-
chotherapy is a difficult task and that whatever role individual psycho-
therapies might be able to play, it is likely to be unfulfilled because about
half of the patients will leave before its benefits can be expected. A clear
implication is that before initiating psychotherapy, both a patient and a
therapist should carefully consider their readiness (see Table 10–1).
Dropouts from manual-guided transference-focused psychotherapy

(TFP), a type of therapy developed by Kernberg, have received the most
attention. Yeomans et al. (1993) found that 9 of the first 14 therapy trials
ended prematurely (within 12 months). Those patients who dropped out
scored particularly high on narcissistic themes (Horner and Diamond
1996) and on impulsivity (Yeomans et al. 1994). Yeomans et al. (1994) re-
ported that more skills in engaging patients can be developed and that
this results in better retention of patients. This was confirmed in a sub-
sequent TFP trial in which only 30% dropped out (Smith et al. 1995).
Kernberg (1982) had earlier warned against the use of TFP for those bor-
derline patients with serious antisocial traits, those for whom secondary
gain was very extreme, or when either extreme situational instability or
impulsivity would preoccupy the treatment content. He cited capacity for
introspection and psychological-mindedness as assets.
The advantages of cognitive-behavioral therapy in retaining border-
line patients were reported in two recent studies in which dynamic ther-
apies had considerably higher rates of dropout (Giesen-Bloo et al. 2006;
Linehan et al. 2006). The considerable within-session structure (e.g.,
role-playing, directives, education, homework, mental exercises) typi-
cally offered in cognitive therapies makes them antiregressive and anxi-
ety relieving. In contrast, dynamic therapies, which traditionally depend
heavily on patient initiative, are typically more anxiety provoking. They
offer less structure within sessions and impose implicit expectations of a
close emotional engagement with the therapist. As is described in Chap-
ter 12, for dynamic therapies to be effective with borderline patients, they
cannot rely on the patient’s initiatives, and therapists need to be particu-
larly active and interactive.
Generally, manual-guided or empirically evaluated treatments show
fewer dropouts than in naturalistic studies. Early trials with dialectical
behavior therapy (DBT; Linehan 1993) and psychodynamic therapy
(Stevenson and Meares 1992) had only a 16%–17% dropout rate in 1 year.
Although subsequent DBT trials have had higher dropout rates (e.g.,
25%; Linehan et al. 2006), there is little doubt that the selection of pa-
tients for research trials provides safeguards against poor candidates.
Contracting Roles
As described in Chapter 3, the earliest (i.e., contractual) form of alliance
involves an agreement between the patient and the clinician about goals
and each person’s respective roles (Table 10–2). Although establishing
goals for therapy has been discussed in Chapter 3, the contracting about
roles discussed here includes agreeing about practical issues such as fees,
scheduling, and frequency of visits; each of these topics is discussed sepa-
TABLE 10–2. Contracting

Practicalities
Fee, schedule for payments
Schedule of visits
Attendance, missed appointments
Vacations
Crisis management
Goals
Insight: to understand yourself
Change: to modify maladaptive attitudes and behaviors or to resolve
conflicts or to develop new ways to attain satisfaction
Relational: to establish new capacities for sharing, attachment, empathy,
trust
Mental: improved capacities to think or to conceptualize self and others
Roles and responsibilities
Therapist Patient
Listen, observe Discuss self
Guide, direct Discuss therapy
Respond informatively
Recognize limits of responsibilities
rately in the next section, “Structuring the Therapeutic Frame (External

Boundaries).”
Most of the empirically validated psychotherapies involve establishing
a contractual alliance as an essential first step (Clarkin and Levy 2006a;
Linehan 1993; Young et al. 2003). Of these, Kernberg’s TFP (see Chapter
12) has the most emphasis on and formalizes the process of creating a
contract (Clarkin and Levy 2006a; Clarkin et al. 1999; Selzer et al. 1987;
Yeomans et al. 1993). Akhtar (1992), like Kernberg et al. (1989), uses
contracting to create an agreed-on frame, which he refers to when prob-
lems are encountered; thus, the frame does not seem arbitrary, reactive,
or punitive. Linehan (1993) also gives great significance to establishing a
contract for borderline patients before starting DBT—for example, es-
tablishing clear goals (target behaviors) for change and making a specific
commitment to those changes, as well as a commitment to attend regu-
larly (see Chapter 11). These clinicians may use multiple sessions, almost
always a minimum of two, to reach an agreement about roles and goals

before therapy begins.
The TFP model emphasizes that the contract should be explicit about
the limits of the therapist’s role and responsibilities; for example, “it does
not fall within the role of the therapist to get involved in the actions of the
patient’s life through phone calls, emergency room visits, etc.” (Yeomans
et al. 1993, p. 256). This limit stems from a conviction that a therapist’s in-
volvement with the patient’s life outside sessions is frequently the cause of
treatment failures. Yeomans et al. (1993, 1994) noted that the contract
begins with a statement by the therapist about the (minimal) conditions
under which therapy can be conducted and that this statement is fol-
lowed by a dialogue that invites the patient to respond. During the dia-
logue, again at the therapist’s initiative, problems are anticipated (e.g.,
coming to sessions intoxicated, intercession crises, or the patient’s not
wanting to leave) that are based on the patient’s history.
When a new patient is known to have posed serious behavior problems
in prior therapies, Kernberg et al. (1989) recommended that “precondi-
tions and structure for their management must be established” (p. 29)
before treatment can take place. He cited as an example a patient who re-
peatedly refused to leave a former therapist’s office and then had spent
much of each day in that therapist’s waiting room hoping to talk to her be-
tween appointments. Kernberg recommended telling such a patient,
“‘You will leave my office and waiting room at the end of each session. Do
you understand why I am saying this [presumably it has been explained],
and is this something you feel you can do?’” Should the patient voice un-
certainty, the therapist is advised to continue, “‘If you do not leave, I will
call for help in removing you. If I have to do that three times, the treatment
will end. I shall inform your parents about this so that they, like you, will
know in advance that this treatment could turn out to be brief’” (p. 29).
Although the idea of a contract is generally helpful, it can have signif-
icant downsides. The first is that requiring too explicit or rigorous con-
formity to a therapy’s boundaries may prompt many patients to be
excluded or to drop out. Many borderline patients are neither reliable
nor foresighted enough to broker a meaningful contract. After a border-
line patient feels attached to a therapist or after work within the therapy
has highlighted a behavioral pattern’s maladaptiveness, patients may
then agree to curtail behaviors that they could not or would not agree to
curtail at the onset of a therapy. In my experience, “contracting” that is
limited to an agreement about practical issues, usually behavioral or in-
terpersonal, and a few simple statements about the therapist’s limitations
(of omniscience or omnipotence) are usually sufficient. I then cite issues
that arose during the evaluation that seemed to have troubled the patient
and that I foresee as amenable to change. I underscore, as does DBT

(Chapter 11), that change is expected—that change is the explicit mea-
sure by which I judge, and encourage patients to judge, whether therapy
is a worthwhile investment of our time and their money.
Therapists can pay too much attention to their own boundaries and
thereby create a second problem for new patients. Specifically, therapists
may think it is necessary to tell a prospective or new borderline patient
whether or when they will be available or how they will respond to issues
such as lateness, rages, and unpaid bills. When a patient has had specific be-
haviors that potentially endanger therapy or safety, they should of course
be addressed, but I, as in schema-focused therapy (SFT; Chapter 11), delib-
erately try to set a less defensive and adversarial tone at the start. The initial
assumption about the failure of a prior therapy should be that the failure
involved misunderstanding by the therapist as well as misconduct by the pa-
tient. With respect to Kernberg’s patient’s prior refusal to leave, I’d start by
noting that such a habit would pose a difficult problem for any new therapy
too. If asked, “What will you do?,” I would respond by saying, “Do? I don’t
know. I know that I, as well as other patients who come here, would feel un-
comfortable by your lingering. You probably already know that, don’t you?
Would you wish for that? If you knew your presence was disturbing, would
you feel like staying anyway? Why would you want to do that?” In this way,
I would expect to be able to avoid setting limits (see Chapter 4).
Contracts offer advantages when patients are being recruited into a
randomized controlled psychotherapy study because they exclude the pa-
tients who are most unlikely to succeed. Randomized controlled trials are
rarely conducted with samples that are broadly representative. Only
about 10% of the patients who meet diagnostic criteria for any diagnosis
will pass all other inclusion criteria and then stay compliant with psycho-
therapy research protocols long enough to provide outcome data (Gun-
derson and Gabbard 1999). This complication was particularly evident in
the studies of mentalization-based therapy (Chapters 5 and 12) and stud-
ies of interpersonal group therapy (Chapter 9). These findings are a ma-
jor reason that the scientific community is now giving more attention to
effectiveness research, in which patients are recruited within natural treat-
ment settings, and exclusion criteria are limited.
Structuring the Therapeutic Frame

(External Boundaries)
It is the therapist’s responsibility to establish a framework for a therapy
(Spruiell 1983). These frames are the obvious signs that the therapist is a
professional at work, work that involves discipline, expectations, and re-
straints. Therefore, these frames are the skeletal representation of the

therapist’s boundaries, boundaries that have obvious interpersonal coun-
terparts (as discussed in Chapter 4 and elsewhere in this chapter). Com-
ponents of the external boundaries are billing, frequency of sessions,
scheduling, and the seating arrangement.
Bills
Some clinicians in private practice see borderline patients only if they re-
ceive a high fee, justifying this requirement on the basis of the extra dif-
ficulties they expect. High fees may be justified for these patients, but
such fees should be based on expertise. The therapist who is apprehen-
sive about the difficulties expected from borderline patients can justify
the high fees only if he or she uses the money to pay for supervision. Oth-
erwise, these patients should be referred.
Getting the bill paid consistently is often a problem with borderline
patients. It can be hard to distinguish tardy or missing payments that are
based on a patient’s general lack of organization and conscientiousness
from those that are based on anger, denial, or feelings of entitlement.
Skill building through education, reminders, and planning may be help-
ful when payment problems are due to a general lack of responsible func-
tioning. Problems that are an expression of acting out angry or entitled
feelings about the therapy (or life) require interpretation and potentially
limit setting. It is often a useful option to have patients who are delin-
quent about their bill pay at each session so that the issue is very difficult
to overlook. The following vignette examines payment problems.
Vignette
Ms. Y had been in therapy for two sessions per week at a reduced fee for
several years. She rarely paid the bill on time or fully, and eventually her
failure to do this led to suspension of the therapy. As a condition for my
seeing her once weekly, she contracted to pay it off gradually and to bring
her payment to each session. She did this quite satisfactorily for her first
few sessions. I treated this as an expected and unremarkable event. At the
end of the third session, she neglected to pay her bill, and I then took this
up actively on her fourth visit. She was duly apologetic and assured me
that it wouldn’t be an issue. At the end of that session, she inquired about
the amount and then asked for a pen with which to make out her check.
I was aware that by helping her pay the bill, I was extending her session,
but not to do so might have excused her negligence. I told her that I was
sorry that this couldn’t be discussed, but I was sure she could find a pen.
Perhaps this response was acceptable for that session, but I never found a
satisfactory or sustained way to explore her financial irresponsibility.
Eventually, this patient’s unpaid bills elsewhere in her life led to her relo-
cation and termination of therapy.
It is tempting to bypass borderline patients’ involvement with their

bill payments when the payment can more reliably be procured by billing
others (insurers or family). Most borderline patients prefer the clinician
to do this, and although I try to keep the patient as the responsible inter-
mediary, I confess that sometimes the hassle is not worth the benefits.
Even when billing is taken care of by others, I underscore to the patient
that the issue of being financially responsible inevitably must return to
him or her—that to become the respectable person the patient wants to
be will require the “not yet adult” patient to assume such burdens.
The issue of whether to bill for missed sessions can be of particular trou-
ble to borderline patients (or to their families). Therapists differ in their
standards about this, but borderline patients periodically test the clini-
cian’s willingness to make exceptions for them. Therefore, therapists usu-
ally will find it valuable to advise patients of their policy and be prepared to
maintain it. I think it is useful for most borderline patients to agree that
they are expected to pay for any missed session short of emergencies
(which do not include hangovers, headaches, and unexpected guests).
Frequency
Figure 10–1 and Table 10–3 show the relation of the frequency of visits to
therapeutic goals. If the therapeutic relationship is to be sufficient for the
correction of unstable introjects or remission of a pattern of insecure at-
tachments, two or more psychotherapy sessions a week are probably re-
quired, although this conclusion is untested. Both in Kernberg’s TFP
(Kernberg et al. 1989) and in Young et al.’s (2003) SFT, twice-weekly ther-
apies are considered necessary to be capable of effecting structural
change (although TFP’s emphasis is more on the requirements for trans-
ference analysis, and SFT’s emphasis is more on the requirements for the
relationship to be corrective). In my experience, three times a week is
usually preferable, if possible, for these options.
Therapies with a mandate to help patients understand themselves al-
most always require more than once-weekly sessions. The exception to
this general rule is that dynamic therapy once a week is feasible when
therapists see a patient who is “held” by residential or intensive outpatient
services (levels of care II–IV) and later “held” by other outpatient modal-
ities (e.g., family, skills group). If, however, the therapist sees the patient
once weekly in the absence of other modalities, the therapist must pro-
vide the holding functions: must get involved in crisis management,
emergency telephone calls, medications, and other issues of the patient’s
current reality—activities by the therapist that carry great meaning to the
patient but are inadequately examined. This type of “therapy” requires
enough directives, advice, limits, and the like that the therapist’s activities
FIGURE 10–1. Relation of frequency of sessions and treatment

goals.
involve what is better labeled case management, and the therapist’s role is
what has been identified here (Chapter 4) as that of the primary clinician.
It is misleading to think of such case management activities as psychother-
apeutic, although they can, nonetheless, be very helpful. When a clini-
cian’s role is dominated by these activities, once-weekly sessions are
usually sufficient.
For a twice-weekly outpatient psychodynamic therapy to succeed in
the absence of a second modality, the patient will need to have reasonably
good impulse control and low liability risks. Twice-weekly sessions are suf-
ficient for the holding function, but this frequency often lends itself to
more supportive, current-events–focused therapy when borderline pa-
tients lack significant other supports. Three times a week is more desir-
able with competent therapists for development of themes and for
focusing on the therapeutic relationship. The corrective benefit of devel-
oping a trusting and secure attachment is more likely than with twice-
weekly therapies, no matter how skillfully delivered. Still, in addition to
the problem of financial feasibility, such intensive therapies should be
undertaken only when therapists are appropriately ready (see later in this
chapter, section “Therapists”). For patients, being ready for such therapy
means having adequate social supports and impulse control.
Scheduling
Therapists generally try to do their best to accommodate the scheduling
that patients request. Two caveats here relate to BPD. One is that some-
TABLE 10–3. Relation between goals and frequency of sessions
FREQUENCY
(PER WEEK) GOAL
1 Management and support; this can be an anchor (a

stabilizing influence) that helps the patient to learn and
grow from life experiences.
2 Management and “therapy”: this can foster change via
insight, using either dynamic or cognitive strategies.
3 Optimal for therapies when examination or corrective
experience of the therapy relationship is central.
≥4 The patient’s life is likely to revolve around therapy until
growth occurs; frequency can be useful for patients who
lack internal structure but carries significant potential
for being harmful.
times borderline patients’ ambivalence about assuming a responsible role

will lead them to schedule appointments at a therapist’s convenience, at
the expense of work, school, or other “get a life” activities. Therapists must
recognize this tendency by patients and try to schedule appointments at
times that will support such activities. A second caveat is that therapists
may adopt too rigid a stance about appointments with borderline patients
to avoid getting caught in control struggles or being manipulated. Both
are possible, but it is better to discover and discuss such occurrences—
then after due processing, set a limit if necessary (see Chapter 4)—rather
than to assume the worst.
Gutheil and Gabbard (1993) point out another issue: the “special-
ness” that patients can attribute to being the last patient in a therapist’s
day. Although the authors point out how this can be part of a slippery
slope leading to boundary violations, what may be more generally useful
regarding borderline patients is that when a very intense transference de-
velops, with marginal or erratic reality testing, it is best to schedule such
patients during “high-traffic” times when other patients, other staff, and
other interests by a therapist are evident.
Seating
This discussion assumes that therapy appointments will take place within
a therapist’s office. The general principles about seating are to let the pa-
tient decide where he or she would like to sit but to retain a role as ad-
viser. Of the three standard arrangements—across, convergent, and
parallel—borderline patients may initially prefer to sit across to protect

their distance. That is fine, but I encourage patients to see the convergent
arrangement as a step forward in trust. (In this arrangement, the chairs
are side by side, angled somewhat toward each other, so that when each
person looks ahead, his or her vision converges separately from the other
person’s.) Psychoanalysis, in which the parallel arrangement exists, is the
only therapy in which patients are disinvited to look at the therapist. Be-
cause of fearfulness (projections), few borderline patients would want or
accept this “blank screen” opportunity, and for those counterphobic bor-
derline patients who would, it should be discouraged.
Fonagy (1995) referred to a borderline patient who sat on the floor of
his office, paced back and forth, or lay on his couch facing him. His ac-
ceptance of these arrangements was no doubt encouraged by his training
in child analysis. I view such tolerance with adult patients as dangerously
regressive. It underestimates a borderline patient’s expectable awareness
of the behavior’s inappropriateness and undermines the task orientation
that therapists need to sustain. It is not that a limit needs to be set; it is that
the meaning of any grossly inappropriate behavior, including testing, de-
fiance, or disrespect, needs to be the subject of an exploration, actively
initiated by the therapist. I prefer to pursue such exploration, even if a pa-
tient is resistant, because of my concern for how the behavior (such as sit-
ting on the floor) distracts from the agreed-on goals of the therapy (i.e.,
is therapy interfering). Limits are rarely needed (see Chapters 3 and 12).
Therapists
Qualifications
This book is a testimonial to the complicating consequences of greater
knowledge about treating BPD. Specifically, we know that the therapeutic
tasks and modalities required for patients at the variety of treatment set-
tings and within the current wide range of phases of their improvement
require clinical staff who have specific experience, training, and personal
qualities. Even within the relatively narrower group of therapeutic tasks
that are needed for borderline patients who are ready to undergo indi-
vidual therapy, the factors of experience, training, and personal qualities
still need to be considered (see Fine 1989).
Regrettably, therapists vary considerably in their skill with borderline
patients. Some psychiatrists, many social workers, and most nurses recog-
nize that they “aren’t good for borderline patients” and would happily
avoid them (B. Pfohl, K. Silk, C. Robins, M. Zimmerman, and J. Gunder-
son, “Attitudes Towards Borderline PD: A Survey of 752 Clinicians,” un-
published data, May 1999). However, many mental health professionals
believe they are capable with borderline patients but are still not in fact
good for them. This overestimation of oneself is usually based on naiveté
about oneself or about borderline patients, but it is sometimes based on
the appeal (Main 1957) such patients can have for prospective therapists:
the prospect of being very helpful to someone for whom life has been un-
fair and whom others have reportedly failed stimulates heroic fantasies.
Being blind to one’s limits also can be propelled by the very practical
pressures to fill one’s time, whether in private practice or in a clinic. Pfohl
et al.’s study also indicated that mental health professionals with more ad-
ministrative experience in hospital or residential programs had less po-
larized ideas about the borderline patient’s likely responsiveness to
psychotherapies and that psychologists proved distinctly, and quite uni-
formly, more optimistic.
Borderline patients can easily get into psychotherapies with clinicians
whose training or experience is clearly inappropriate for the patients’
therapeutic goals (Figure 10–2). One common example involves clini-
cians with experience and training only in short-term or nonintensive be-
havioral therapies who, often in response to borderline patients’ requests,
escalate the frequency of visits to three or even more times a week. Such in-
tensity invites a regressive dependency, which then cannot be adequately
used for personal growth because of the therapist’s lack of training either
in transference management or in making this availability contingent on
the patient’s improvement. Rather, borderline patients’ dependent hopes
for direction, protection, or nurturance are likely to be enacted in a rela-
tionship with such therapists, who interpret these hopes as needs. Border-
line patients welcome such therapies, but they may conclude that being
sick is the only way their dependent longings will be fulfilled.
Vignette
Ms. Z, a 31-year-old woman, had talent as a writer and was the mother of a
3-year-old daughter. She entered therapy after the second of two hospital-
izations for suicidality, during which it was found that she had been drink-
ing very heavily since being deserted by her husband. She was assigned to
begin therapy with a young female psychiatrist who had graduated from a
dynamically oriented training program, in which she had been recog-
nized for her conscientiousness and supportive attitudes and had special-
ized in treatment of substance abuse. Within a few days after Ms. Z’s
hospitalization, during meetings with her new therapist, Ms. Z had no ev-
idence of depression; indeed, she seemed outgoing and energetic.
Because of Ms. Z’s intellectual curiosity and her wish to get over her
“habit” quickly (and because finances were not an issue), on discharge
from the hospital the therapist agreed to meet with her twice weekly.
Within the first week, Ms. Z called twice with concerns about her depres-
sion (i.e., her sense of “badness”) returning. On the second telephone
FIGURE 10–2. Relation of therapist type and treatment goals.
call, this problem gave way to talking about Ms. Z’s daughter’s cough. In
response to Ms. Z’s request, the therapist gave advice about medication,
spelled out the names, and, in an attempt to be helpful, advised the pa-
tient about a pharmacy close to where she lived.
Within a few weeks, a pattern began whereby Ms. Z came three times
a week and called two or three evenings, often regarding the care of her
child. The therapist, becoming resentful about the growing demands on
her time, hesitantly suggested that she would charge for the telephone
time. When the patient got angry (“I thought you really cared,” “If you
want to quit, just say so,” etc.), the therapist responded by dropping the is-
sue. The therapist then felt even more resentful and sought supervision.
This vignette illustrates how neither psychiatric training nor a gener-

ally good set of skills in a related area can offer assurances to patients
about selecting a qualified therapist. To become good with borderline pa-
tients requires experience and training; specifically, otherwise good ther-
apists still need to have supervised experience with these patients. To
develop the needed skills usually requires 2–3 years of fairly extensive,
preferably multifaceted, contacts such as those derived from inpatient or
residential settings. Psychiatrists who manage medications for many bor-
derline outpatients usually learn the basics; such experiences make one
comfortable with the management issues.
If an otherwise good therapist wants to develop psychotherapeutic
skills for borderline patients but has not had case management experi-
ence, he or she will need good supervision while treating several patients
over the course of the first year or two of treatment (when management is-
sues are prominent). Only psychiatrists or psychologists who have already
TABLE 10–4. Reasons for a therapist’s taking on a borderline

patient
A therapist’s readiness to take on a borderline patient should involve
the following considerations:
This patient is interesting (challenging, touching, confusing, attractive,
needy, smart, etc.).
I believe the patient can change (for the better).
This is essential, not adjunctive, therapy—a serious responsibility for
which I have time and energy. I’m prepared to persevere despite the
expectation of being burdened, inconvenienced, criticized, hurt, and
possibly failing.
The patient is suffering (hostilities, acting-out, and symptoms are there
for a reason).
I believe that I can help, but that cannot be assured.
I know that if treatment doesn’t seem helpful, I will seek consultation,
and I can discontinue. (The therapy is not a lifetime contract for the
patient or for me.)
My life is good. (Life outside my role as a therapist is reasonably
fulfilling.)
seen borderline patients nonintensively, with good supervision, are good

candidates to do intensive therapies with properly selected borderline pa-
tients. Here too (see Chapter 4), supervision or consultation is desirable.
Even with intensive training or good supervision, high-frequency ses-
sions with patients should be undertaken with caution because personal
qualities are also important. It is not merely a matching issue: therapists
who do well with one patient will do well with most. Therapists should not
take on borderline patients without consideration of the issues about the
patients and themselves listed in Table 10–4.
The vignette about Ms. Z also illustrates the pressures that borderline
patients bring to bear on inexperienced, naive, or untrained therapists.
Borderline patients welcome anyone’s attention and support, and they
are willing to give up ego or social functions in the service of feeling taken
care of (the issue of regression is discussed in Chapter 2). They cannot be
relied on to make an informed selection of a therapist that is based on
who can help them to change. As described in Sidebar 10–1, such obser-
vations raise the question of whether therapists treating BPD patients
should be required to have special training and to be certified for com-
petence.
Sidebar 10–1: Quality Assurance: Should

Therapists Be Credentialed to Treat BPD?
A very important lesson that has emerged from the efforts
to study the efficacy of therapies with borderline patients is
the difficulties that investigators encounter in training moti-
vated trainees to achieve competence. Both Kernberg
and Linehan have had many otherwise capable students
who, after receiving their training, were still unable to con-
duct the therapy with high levels of competence. In the
study in which DBT failed to surpass Kernberg’s TFP (Clarkin
et al. 2007), DBT advocates raised questions as to whether
this result occurred because the DBT was not high enough
quality. Similarly, efforts to train therapists to do Kernberg’s
manual-guided TFP (Clarkin and Levy 2006a; Kernberg et
al. 1989) also have been frustrated by the discovery that
most therapists fail to adhere to the prescribed practices
(Yeomans et al. 1994). Indeed, in the study in which TFP
emerged as less effective than SFT, TFP advocates could
cite the failure of TFP therapists to be trained to adherence
(see Chapter 11). The fact was that 85% of the SFT therapists
were competent and adherent, whereas fewer than 50%
of the TFP therapists were (Giesen-Bloo et al. 2006). Better
results were found for manual-assisted cognitive treatment
when it was provided by competent and motivated ther-
apists than when therapists were not selected for either
competence or motivation (see Chapter 11). These diffi-
culties within research settings confirm that many therapists
are just not very good with borderline patients. Problems
with competence (and adherence) abound. Although this
problem may be true for other diagnostic groups, it seems
to be truer for borderline patients.
The implications of these findings for interpreting re-
search and for clinical practice are quite profound. One
implication is that competence and adherence are impor-
tant and that in the absence of either, it is hazardous to ex-
trapolate lessons from these studies for one’s own clinical
practice.
The experience of both Linehan and Kernberg is that
self-selected would-be practitioners of their quite different
treatments often are not well suited to manage borderline
patients. In a study at the Karolinska Institute in Sweden (J.F.
Clarkin, personal communication, November 1999), it be-
came clear that training therapists to become competent
and adherent to any manualized therapy model is very
time-consuming and fraught with difficulties. There are
some obvious reasons for this problem. To do well, one must
be empathic and flexible and composed in the face of
predictably intense feelings and demands. As noted earlier
in this chapter, to attain such composure, in my experi-

ence, therapists usually need to have had significant expo-
sure within residential or hospital settings where structures,
supports, and collaboration are built in.
In 1998, the European branch of the International Soci-
ety for the Study of Personality Disorders approved a plan
to develop quality assurance guidelines for personality dis-
order assessment and interventions that require that thera-
pists be audited and monitored to ensure high standards.
The introduction of quality assurance considerations is diffi-
cult. Table 10–5 outlines the sort of credentialing require-
ments that a new therapist might expect. On an informal
basis, quality assurance is conducted by clinicians being
supervised and being required to present cases to others
for feedback. Supervision provides both support and con-
tinuing education and is sufficient for most purposes. Audit-
ing and monitoring of therapists would require pressures
from professional organizations, governmental agencies,
or third-party payers. At present, Linehan is the leading ad-
vocate for credentialing therapists. In this regard, she is set-
ting a standard that other therapies may learn to aspire to.
Important exceptions to these cautions about initiating individual psy-

chotherapies involve clinicians in training. Many are sufficiently appre-
hensive (or offended) by borderline patients that they will avoid working
with them. However, trainees who want to become good psychotherapists
and who have good supervision available should embrace the opportu-
nity to work with borderline patients: these patients offer exceptional
learning experiences, and, surprisingly, they often will get better. Thera-
pists who do not find the issues that surround the treatment of borderline
patients (action, dependency, anger) interesting or who do not actually
like such patients will be unlikely to do well with them.
Qualities
Some qualities cannot be taught, or untaught. This explains why thera-
pists with very different training and theories can become excellent ther-
apists (Sidebar 10–2). One study (Rosenkrantz and Morrison 1992)
concluded that therapists who are high on “anaclitic, depressive and fu-
sion tendencies” do poorly, whereas those who are “high boundary” ther-
apists function well. In my experience, therapists who do well are usually
reliable, somewhat adventurous, action oriented, self-confident, and
good-humored. Linehan might add “irreverent” to this characterization.
This translates into being active and responsive. Positively unworkable
are therapists who are effete, depressive, anxious, genteel, or controlling.
TABLE 10–5. Proposed credentialing requirements for primary

clinicians or therapists treating borderline patients
Experience
One or more years of experience with borderline patients within
residential or inpatient units
An estimated 100 hours of one-to-one contact in either
psychopharmacological or nonintensive psychotherapies
Training
Didactic training in the theory and technique of the modality
Adequate hours of training by a certified trainer in the modality to be
practiced to achieve competence
Opportunity and commitment to ongoing supervision
Mental health
Good affect tolerance, empathic ability, self-sufficiency
Sidebar 10–2: Listening to Kernberg or Linehan:

Can Charisma Cure BPD?
Listening to Kernberg, clinicians aspiring to treat BPD could
feel convinced that the ability to do so would depend on
their comfort with aggression (their own and others’) and
their ability to identify and interpret unrecognized motiva-
tions or conflicts and might, in addition, be increased by
completing psychoanalytic training.
Listening to Linehan, clinicians with the same hopes
could conclude that to attain the same goal would de-
pend on their identifying problems the patient wants to
change and having mastered the comparatively clear
DBT theory and skills; mastery in this instance would require
attendance at several intensive weeks of workshops and
would, in addition, be assisted by more extensive training in
cognitive-behavioral therapy.
Still, maybe the success of both Kernberg and Linehan
with borderline patients is not attributable to either their
theories or their training. Perhaps the secret to their suc-
cesses lies in what has been disparagingly called the non-
specific components of what they offer. Both Kernberg
and Linehan are charismatic. They are authoritative: they
embody confidence, clarity, forcefulness, and certainty.
Swenson (1989) noted that both meet “the patient’s emo-
tional intensity or lability head on with a steady emotional
intensity of their own. Both therapists give the patient the
feeling that they are present, engaged, and indestructi-

ble” (p. 32). No doubt patients are transported by the same
impulse to accept what these therapists say that has led
many professionals to try to do as they say. Moreover, both
Kernberg and Linehan seem undaunted by controversy,
perhaps even enjoying debate and challenge. Both wel-
come opportunities to demarcate their positions and to
clarify how their views may be distinguished from the views
of others. No doubt patients are impressed by these thera-
pists’ efforts to make themselves clear and by their atten-
tion to fine distinctions. And both are prepared to share
their viewpoints, whether sought or not. Patients of either
Kernberg or Linehan, I believe, feel confident that their
opinions, judgments, or decisions will be heard and re-
sponded to—what Swenson summarized as feeling “emo-
tionally held” (p. 32).
Attentive, challenging, and responsive: it is my impres-
sion that these same qualities distinguish others who are ex-
cellent therapists with borderline patients. Certainly, these
qualities seem valuable in getting borderline patients en-
gaged, but the longer-term processes by which sustained
remissions of BPD occur may be where differences in theory
and training can be critical.
Many borderline patients have strong preferences for one gender or

the other, but their reasons bear exploration. It may be that they associate
femininity with qualities of closeness and empathy—admittedly more
common in females, but hardly unheard of in males, especially those who
have chosen the helping professions. Patients may associate masculinity
with protectiveness and directiveness—again, qualities more common in
males but hardly confined to them. As an approach to finding a therapist,
gender is far less useful than the desired personal qualities.
So how does a borderline patient select a good therapist? It is not easy
(Table 10–6). Patients should be cautious because many therapists have
neither the training required nor the familiarity and comfort with their
issues to be qualified to treat them. As discussed earlier, certain therapist
qualities also seem almost universally desirable for borderline patients.
For example, therapists need to like working with issues of action, anger,
and dependency.
Engagement
In the first phase of treatment with a borderline patient, the therapist’s
goal should be primarily to engage the patient. How a prospective ther-
apist manages the issues discussed earlier about contracting and estab-
TABLE 10–6. How a borderline patient should choose a therapist

The patient should ask himself or herself the following questions about
the therapist:
Do I want to be involved with this person? (Do I think I could like him
or her?)
Does the therapist seem to want to get involved with me? (Does he or
she seem interested in me?)
Can I learn something from the therapist? (Does he or she seem
confident, knowledgeable, ready to convey what he or she knows?)
Does the therapist have training, skills, and experience?
Is the therapist sufficiently reliable, conscientious, and durable? (Does
he or she seem fragile, unpredictable, or restless?)
lishing a framework, as well as the therapist’s personal qualities, are

important determinants of whether a patient becomes engaged. Still, en-
gagement per se involves moving the alliance from the contractual type
described earlier to the relational type (see Chapter 3). The earliest in-
dications of patient engagement involve the perception that the therapist
is likable or wants to be helpful (Alexander et al. 1993). Engagement is
unlikely when the therapist is described as “no personality,” “blah,” or
“nothing to say.” Research indicates that even by 6 weeks of therapy, pa-
tients should indicate an overall positive relationship with a new therapist
and a hopefulness about benefiting from the therapy (Gunderson et al.
1997; Horwitz et al. 1996). Therapies that fail to get this type of start may
warrant consultation.
The three components of helping the patient becoming engaged are
1. Invoking the patient’s attachment to the therapist

2. Invoking the patient’s hopes for change
3. Invoking the patient’s interest in self-disclosure and self-examination
(i.e., the learning process)
As noted in Chapter 6, a therapist who administers medications often

can jump-start a borderline patient’s hopes for change and confidence in
the therapist’s intention to be helpful. During these initial few months, it is
very useful for the therapist to be quite active in structuring the sessions,
encouraging the patient with tasks such as writing an autobiography be-
tween sessions, and giving the patient encouragement to think about what
they have discussed. Psychodynamic therapists who might otherwise ex-
pect the patient to take the lead can learn from cognitive-behavioral ther-
apists, who have found that a directive, businesslike approach that
purposely does not evoke intense transference is useful in alliance build-
ing (see Chapter 11). The therapist should convey an interest in the psy-
chotherapeutic task and implicitly—sometimes explicitly—offer hope that
the patient is capable of change and capable of having a more satisfactory
future. With these supports, a borderline patient begins to develop both a
realistic hope that change can occur, albeit slowly, and an appreciation of
the therapist’s commitment to him or her as well as to the task.
The basic axiom of dynamic therapies, to let positive transference alone
but to be active about early signs of negative transference, applies to bor-
derline patients. Still, when borderline patients begin treatment with ex-
treme idealization and optimism about therapy or the therapist, these
tendencies should not be mistaken for a sign either that they are commit-
ted to the treatment tasks or that their optimism is connected to expecta-
tion of personal change. Indeed, whenever such idealization and optimism
become evident, I conscientiously and good-naturedly demur—in order to
diminish the risk that the inevitable disillusionment will be too bitter. This
approach is similar in style to what Linehan calls “irreverence” and heeds
Dawson and MacMillan’s (1993) warning about being “too therapeutic”
(see Chapter 4). For patients who begin their treatment with skepticism
and devaluation, this approach should be actively explored. Such attitudes
can be self-fulfilling and often represent a defense against their hopes that
got them to see you. By 6 months, patients should have acquired some hope
that “therapy might help.” This hope can derive primarily from the experi-
ence of the therapist’s involvement or from the therapist’s convictions, or,
often enough, from the actual evidence of change (see Chapter 3).
Borderline patients are very sensitive to whether a prospective thera-
pist seems interested in them. For most borderline patients, lack of inter-
est translates into feeling rejected and unwanted or into feeling “I’m bad”
(Young’s Punitive Parent mode as conceptualized in SFT; this therapy and
its modes are fully described in Chapter 11, section “Schema-Focused
Therapy”). The more fearful about lack of interest the patient is, the
more sensitive to signals of inattention he or she will be. For the border-
line patient, signs of a therapist’s lack of interest are worse even than signs
of being misunderstood. At signs of inattention, some borderline pa-
tients will become silent and withdrawn. Others will become irritated and
say “Pay attention” or “You’re not listening, are you?” or demand “What
are you thinking about?” (Young’s Angry Child mode). These protests are
clear and meaningful requests that therapists should attend to by becom-
ing more active and interactive. Often honest self-disclosure is useful.
Therapists whom borderline patients want to become engaged with are
those whose interest in patients is evidenced by being reactive and inter-

active: patients call them “all there,” a “real person.” These qualities are
interpreted by borderline patients as a therapist’s “likability” and “help-
fulness”—that is, those qualities that Alexander et al. (1993) reported
best determine whether patients will want to become engaged.
Likability is most closely related to a therapist’s level of activity and in-
terest—usually evident in the small, off-the-record exchanges at the start
or end of sessions, during which patients and therapists often exchange
comments on things such as weather, clothing, transportation, or the
news. Common values should become apparent through such informal
comments, as well as in relation to the patient’s description of his or her
life. It is useful for therapists to disclose common values, attitudes, and
the like in the service of developing the relational alliance (Chapter 3).
In the early 1990s (Jordan et al. 1991), a psychotherapeutic approach
based on women’s psychology that encouraged many therapists, especially
females, to bring an openly caring (empathic, validating, nonconfronta-
tional) approach to borderline patients was similar with its emphasis on
validation to what the developers of DBT, SFT, and mentalization-based
therapy were independently recognizing. These responses evoke what
Young referred to as the Abandoned Child mode. This approach helps pa-
tients become attached; and it has diminished, I believe, the usual fre-
quency of early dropouts. The clinical theoretical basis for this approach
derived from the psychoanalytic theory by Kohut (1971) and its applica-
tion to BPD by Adler (see Chapter 12). What the empathic, validating ap-
proach does is ignite an idealizing transference, which will encourage
engagement with therapists. Whether the therapy facilitates change prob-
ably depends on a therapist’s repertoire of other forms of intervention,
which are described in Chapters 11 and 12.
Several characteristics of borderline patients, unrelated to a thera-
pist’s interventions or style, can affect whether they become engaged in
therapy. Borderline patients who start within longer-term baseline hospi-
talizations or who have had more prior psychotherapy are more likely to
become engaged in further psychotherapy (Gunderson et al. 1989). Psy-
chotherapists may think this observation means that prior treatment ex-
periences make borderline patients more tolerant of the inherent limits
of therapy, but it is more likely that the engagement occurs because pa-
tients have become dependent on psychotherapist support. The overrid-
ing fact is that the borderline patients who get the most psychotherapy are
those who improve the least. In any event, the more general principle of
split treatment (Chapter 4)—that is, the use of a second collaborating and
complementary modality to help contain splits, projections, and flight—
can greatly enhance the likelihood of engagement in psychotherapy.
Generic Qualities of Effective Psychotherapies

The current array of four “comprehensive” empirically validated thera-
pies for borderline patients is unexpected, exciting, and, for ordinary cli-
nicians, confusing. Behind the sometimes heated debates about the
relative merits of these approaches, there is a more reassuring apprecia-
tion that the therapies must share underlying processes to account for
their similar profile of benefits. This commonality has already prompted
preliminary consideration (Clarkin and Levy 2006a, 2006b). Sum-
marized in Table 10–7, reviews of the extensive literature on individual
psychotherapies for BPD identify generic characteristics that appear to
distinguish effective therapies (Gunderson and Links 2007; Waldinger
1987). Notably, the need for a stable framework and high levels of ther-
apist activity that are essential in psychotherapy also characterize case
management activities (Chapter 4), psychopharmacology (Chapters 6
and 7), and family (Chapter 8) and group (Chapter 9) therapies. Making
self-destructive behaviors dystonic is underscored in case management
(Chapter 4), hospital care (Chapter 5), and family interventions (Chap-
ter 8). These generic qualities provide a background for the descriptions
of seemingly distinct principles and practices of the individual psycho-
therapies found in the next two chapters.
The emergence of empirically validated treatments has introduced
specific forms of training and measures of a therapist’s qualifications to
provide that therapy. Table 10–8 shows the variations in therapist training
and whether adherence or competence is measurable. This table also
shows that the intensity and feasibility of receiving the ongoing supervi-
sion that is needed to stay “on model” vary considerably. It is clear that
DBT has the best-established training. It is also clear that mentalization-
based therapy lags behind the other models in specifying the training
and supervisory requirements and in developing adherence and compe-
tence assessments.
According to current randomized controlled trials (reviewed in Chap-
ters 11 and 12), DBT has the strongest scientific support and the best-
established and most-available methods for training therapists. It clearly
has more sites where it is taught and practiced than any of the other em-
pirically validated treatments. Evidence indicates that TFP might be as
effective as DBT (Clarkin et al. 2007), but the training of competent ther-
apists may require more time, supervisory requirements are more de-
manding, and it may be more difficult for TFP therapists to sustain
adherence (see Sidebar 10–1). TFP has the advantage of a large cadre of
dynamically oriented therapists who are familiar with and devoted to its
theoretical base. On the other hand, this same community is traditionally
TABLE 10–7. Distinguishing characteristics of effective

psychotherapies
Therapists must provide a stable treatment framework. This starts by
establishing agreed-on roles and goals. Therapists should stress that
psychotherapy is a collaborative enterprise.
Therapists must identify adverse effects of self-destructive behaviors.
They should establish from the outset that safety is an important issue,
but it must be made clear to the patient that therapists are neither
omnipotent nor omniscient.
Highly active and involved therapists are essential to elicit a sense of
agency and alliance.
Therapists should establish a connection between actions and feelings
with detailed chain analyses.
Therapists must manage and pay careful attention to countertransfer-
ence. Ongoing discussion of the therapy with colleagues is valuable
(even for experienced therapists) and is built into those therapies that
have been empirically validated (see Table 10–8). Such discussions
enhance the therapist’s ability to contain and clarify the borderline
patient’s projections (Gabbard and Wilkinson 1994) and contain
countertransference anger and resentment, which can safeguard
against suicidality (Maltsberger and Buie 1974). In the absence of
ongoing discussions with other clinicians, therapists should set a low
threshold for seeking consultation.
Source. Gunderson and Links 2007.
resistant to the discipline and constraints required for adherence. SFT has
a theory that combines both behavioral and psychoanalytic concepts. It
may have advantages over TFP because of therapists’ ability to learn and
adhere to it. Still, SFT has not been widely practiced, and it may require a
commitment similar to what Linehan has invested if it will generate a
cadre of devoted teachers and practitioners. A fourth therapy, mentaliza-
tion-based therapy, also bridges cognitive and psychoanalytic concepts. Its
value as an outpatient therapy is currently being tested against a treat-
ment-as-usual condition. Its base in Fonagy’s early developmental obser-
vations is very appealing, but mentalization-based therapy remains
seriously behind the other three therapies in not having established mea-
sures for competence and adherence and in not having been tested out-
side Bateman’s oversight or in comparison to any of the other validated
treatments.
Individual Psychotherapies
TABLE 10–8. Therapist training for borderline personality disorder: empirically validated treatments
TRAINING COMPETENCE ADHERENCE
THERAPY WORKSHOPS ONGOING SUPERVISION SUPERVISOR MEASURE MEASURE
Dialectical behavior therapy 10 days Weekly group Peers Yes Yes

Mentalization-based therapy 1–3 days Weekly individual Developer No No
Transference-focused 28 hours Weekly individual and group Certified trainers Yes Yes
psychotherapy every 4 months
Schema-focused therapy [?] Weekly individual and group Certified trainers Yes Yes
every 4 months
❘ 249
Summary
In this chapter, I have discussed the conditions that can determine
whether psychotherapy with a borderline patient should be initiated and,
if so, what the conditions are that will allow it to succeed. I hope that read-
ers will recognize that, although exceptions to every rule exist, it makes
no sense to ignore probabilities. Clearly, not all borderline patients are
candidates for psychotherapy. Capability and motivation need to be as-
sessed. How the therapy framework is established is of critical impor-
tance, and therapists should have a good understanding of issues of
scheduling, billing, and agreed-on goals to give the venture the best
chance of success. Of particular importance is to recognize how the
framework should be fitted to the patient’s needs and to the capabilities
of both patient and therapist. It is also clear that not everyone can treat
borderline patients well; therapists should consider their capabilities.
Finally, intensive schedules of psychotherapy should be offered only by
qualified professionals.
References
Akhtar S: Broken Structures: Severe Personality Disorders and Their
Treatment. Northvale, NJ, Jason Aronson, 1992
Alexander LB, Barber JP, Luborsky L, et al: On what bases do patients
choose their therapists? J Psychother Pract Res 2:135–146, 1993
Mental Disorders, 3rd Edition. Washington, DC, American Psychiat-
Clarkin JF, Levy KN: Psychotherapy for patients with borderline person-
ality disorder: focusing on the mechanisms of change. J Clin Psychol
62:405–410, 2006a
Clarkin JF, Levy KN: Special issue: putative mechanisms of action in the
psychotherapy treatment of borderline personality disorder. J Clin
Psychol 62:405–516, 2006b
Clarkin JF, Yeomans FE, Kernberg OF: Psychotherapy for Borderline Per-
sonality. New York, Wiley, 1999
Clarkin JF, Levy KN, Lenzenweger MF, et al: Evaluating three treatments
for borderline personality disorder: a multiwave study. Am J Psychia-
try 164:922–928, 2007
Dawson D, MacMillan HL: Relationship Management and the Borderline
Patient. New York, Brunner/Mazel, 1993
Fine R (ed): Current and Historical Perspectives on the Borderline Pa-
tient. New York, Brunner/Mazel, 1989
malfunction in borderline personalities. Int J Psychoanal 76(Pt
1):39–44, 1995
Borderline Patients. Washington, DC, American Psychiatric Press,
1994
Giesen-Bloo J, van Dyck R, Spinhoven P, et al: Outpatient psychotherapy
for borderline personality disorder: randomized trial of schema-
focused therapy vs. transference-focused therapy. Arch Gen Psychia-
try 6:649–658, 2006
47:679–704, 1999
Gunderson JG, Links P: Borderline personality disorder, in Gabbard’s
Treatments of Psychiatric Disorders, 4th Edition. Gabbard GO, Edi-
tor-in-Chief. Washington, DC, American Psychiatric Publishing,
2007, pp 805–821
1989
Gunderson JG, Najavits LM, Leonhard C, et al: Ontogeny of the thera-
peutic alliance in borderline patients. Psychother Res 7:301–309,
1997
Gutheil TG, Gabbard GO: The concept of boundaries in clinical practice:
theoretical and risk-management dimensions. Am J Psychiatry 150:
188–196, 1993
Horner MS, Diamond D: Object relations development and psychother-
apy dropout in borderline outpatients. Psychoanalytic Psychology
13:205–223, 1996
Horwitz L, Gabbard GO, Allen JG, et al: Borderline Personality Disorder:
Tailoring the Psychotherapy to the Patient. Washington, DC, Ameri-
can Psychiatric Press, 1996
Kernberg OF: The psychotherapeutic treatment of borderline personali-
ties, in Psychiatry 1982: The American Psychiatric Association Annual
Review, Vol 1. Edited by Grinspoon L. Washington, DC, American
Kernberg O, Selzer M, Koeningsberg HW, et al: Psychodynamic Psycho-
Kohut H: The Analysis of the Self. New York, International Universities
Press, 1971
Linehan MM, Comtois KA, Murray AM, et al: Two-year randomized con-
trolled trial and follow-up of dialectical behavior therapy vs. therapy
by experts for suicidal behaviors and borderline personality disorder.
Main T: The ailment. Br J Med Psychol 30:129–145, 1957
Maltsberger JT, Buie DH: Countertransference hate in the treatment of
suicidal patients. Arch Gen Psychiatry 30:625–633, 1974
Rosenkrantz J, Morrison TL: Psychotherapist personality characteristics
and the perception of self and patients in the treatment of border-
line personality disorder. J Clin Psychol 48:544–553, 1992
Selzer MA, Koenigsberg HW, Kernberg OF: The initial contract in the
treatment of borderline patients. Am J Psychiatry 144:927–930, 1987
Shea MT, Pilkonis PA, Beckham E, et al: Personality disorders and treat-
ment outcome in the NIMH Treatment of Depression Collaborative
Research Program. Am J Psychiatry 147:711–718, 1990
Skodol AE, Buckley P, Charles E: Is there a characteristic pattern to the
treatment history of clinic outpatients with borderline personality?
J Nerv Ment Dis 171:405–410, 1983
Smith TE, Koenigsberg HW, Yeomans FE, et al: Predictors of dropout in
psychodynamic psychotherapy of borderline personality disorder.
J Psychother Pract Res 4:205–213, 1995
Spruiell V: The rules and frames of the psychoanalytic situation. Psycho-
anal Q 52:1–33, 1983
1992
Swenson C: Kernberg and Linehan: two approaches to the borderline pa-
tient. J Personal Disord 3:26–35, 1989
Waldinger RJ: Intensive psychodynamic therapy with borderline patients:
an overview. Am J Psychiatry 144:267–274, 1987
Yeomans F, Selzer M, Clarkin J: Studying the treatment contract in inten-
sive psychotherapy with borderline patients. Psychiatry 56:254–263,
1993
Yeomans FE, Gutfreund J, Selzer MA, et al: Factors related to drop-outs by
borderline patients: treatment contract and therapeutic alliance.
J Psychother Pract Res 3:16–24, 1994
Chapter 11
COGNITIVE-BEHAVIORAL
THERAPIES
Dialectical Behavior Therapy and
Cognitive Therapies
Overview
A cognitive-behavioral concept of personality disorders involves pervasive
and inflexible patterns of thought (cognitions), feelings (emotions), and
behavior that are self-perpetuating (i.e., governed by the principles of op-
erant conditioning) and self-reinforcing (i.e., governed by classical condi-
tioning). It does not involve intrapsychic structures, an unconscious, or
paradigmatic self-other units, as does a psychodynamic conceptualization
of personality disorders. This chapter begins with a review of the clinical
applications of cognitive-behavioral therapies for borderline patients that
might help clinicians who are not trained in those concepts.
There has been a notable surge of interest in cognitive-behavioral ap-
proaches to borderline personality disorder (BPD) in recent years, gener-
ated in part by larger trends toward empiricism (cognitive-behavioral
therapies are less inferential and more easily assessed than are dynamic
therapies) and the pressures of a managed care environment to define dis-
crete goals and discrete time frames. Moreover, the pioneering contribu-
tion of a manual-guided BPD-specific behavioral treatment (i.e., dialectical
behavior therapy [DBT]; discussed later in this chapter) has dramatically
energized a whole new generation of cognitive-behavioral therapists.
Other cognitive-behavioral therapies have now been empirically validated,
and we should anticipate continued expansion of this approach.
253
The thesis that borderline patients require specialized modifications

of any standardized or usual way of providing traditional institutional or
outpatient therapies is a repeated theme in this book. Some related cau-
tions about this theme as related to cognitive-behavioral therapy are dis-
cussed in this section. For example, a traditional assumption of cognitive-
behavioral forms of intervention, much like the assumption behind med-
ication prescription, is that after an explicit, rational agreement about
targets for change has been arrived at, patients will pursue these goals to
the best of their abilities. As detailed by J. E. Young (1990), cognitive-
behavioral therapies traditionally have assumed that patients have access
to feelings, thoughts, and discrete problems and a willingness to do
homework. This approach no doubt arose because cognitive-behavioral
therapies developed as quasi-experimental services within academic psy-
chology—not within hospital settings, where more severe, unselected,
and unmotivated patients were found. Traditional cognitive-behavioral
training has been for short-term interventions and has not provided
guides for managing the intense interpersonal relationship (including
the transference) or the hostile misattributions (e.g., malevolent control
or abandonment) that characterize borderline patients.
An assumption by cognitive-behavioral therapists that patients will
form rational working alliances to treat discrete problems can be harmful
without modifications that accommodate the special problems inherent in
borderline psychopathology (see Chapter 3). Cognitive-behavioral thera-
pies that fail to make these special provisions for BPD are doomed to fail,
either from dropouts (Coker et al. 1993) or from adverse reactions (Fried-
man and Chernen 1994). Beck and Freeman (1990) and J. E. Young
(1990) took a step toward avoiding the pitfall of covert resistance to change
by emphasizing that before work on change is possible, the primary task is
to establish a working alliance—a collaborative empirical approach.
As noted in Chapters 3, 4, and 6, this approach mirrors the emphasis
given to the need to build an alliance within other types of therapy for
borderline patients. Beck and Freeman (1990) cited mistrust, rejection
sensitivity, and the struggles over control as typical resistances to estab-
lishing such an alliance. They noted, however, that the rather active task-
oriented and businesslike manner that characterizes most cognitive-
behavioral therapists can dispel some of the relational problems, such as
transference elaborations, that are found in dynamic therapies. More-
over, the actively collaborative empirical approach that characterizes cog-
nitive-behavioral theories is a useful model to guide the interventions of
anyone who wants to work effectively with borderline patients.
Cognitive-behavioral therapists often are taught to appreciate the sec-
ondary gains (positive reinforcers) that can make patients’ dysfunction
Cognitive-Behavioral Therapies ❘ 255
desirable and thus resistant to change. Yet these reinforcers may not be
evident with borderline patients who resist therapist communications
with other informants. Even if the cognitive-behavioral therapist suspects
that these reinforcers exist, he or she may find the subject difficult to ad-
dress when the borderline patient carries a psychiatric diagnosis as an ex-
planation for his or her disability. Moreover, although borderline
patients might agree to work on making seemingly desirable changes
(e.g., to stop purging or to attend classes), progress will be impeded if the
cognitive-behavioral therapist does not understand the meanings (in
cognitive-behavioral terminology, underlying assumptions or learned associ-
ations) attached to these changes—meanings that often make changes
feel dangerous or undesirable to borderline patients.
Basic Operant Conditioning Applications for

All Treatment Settings
All clinicians who work with borderline patients need to understand and
apply basic operant and classical conditioning principles. This includes
the recognition that much of what seems transparently maladaptive (e.g.,
withdrawal, rage, or self-destructive behavior) actually serves adaptive
functions by virtue of the responses (reinforcing) that it evokes. Robbins
(1988) was the first analyst to recognize this “functional” analytic under-
standing of self-destructive behaviors. For example, people with BPD act
rather than talk about feelings because they have learned that actions
evoke reinforcing responses such as attention, whereas talking about feel-
ings evokes negative responses such as anger.
Therefore, clinicians need to convey by word and, even more impor-
tantly, by behavior, that the patients’ behavioral adaptations are not useful
in the present stage of their lives. This is called contingency clarification. A
verbal example of contingency clarification is pointing out how the exces-
sive demands for attention or reassurance that typify most relationships of
borderline patients actually alienate the very people whose love and care
they most hope to gain. Furthermore, clinicians must always be conscious
of how their reactions and behaviors reinforce or diminish patients’
pathological behaviors. For example, a patient may demand attention (a
dynamic interpretation) by staying beyond a session’s contracted time. He
or she should be responded to both with appreciation for the historically
adaptive function of such behaviors and with negative consequences—for
example, disapproval (an aversive interpersonal response intended to ex-
tinguish the behavior) or, if need be, shortening the next appointment
(an action to extinguish the behavior). If it is an option, I leave the office
myself. (As an addendum to this example, the very effective aversive re-
sponse I make is to reluctantly begin working on other tasks when a pa-

tient is silent or otherwise disregarding the tasks of therapy.)
Here are two particularly important examples of how a consciousness
about reinforcements should affect a clinician’s reactions to borderline pa-
tients: 1) seeing them less frequently when they are more dysfunctional
(e.g., in the hospital) and 2) consistently noting that suicidal acts or threats
are reasons to doubt a therapy’s value, not reasons for its continuation or in-
tensification. In a larger sense, families, clinical settings, and psychothera-
pists should all consciously make efforts to offer less attention to negative
(maladaptive) behaviors and more attention to positive (adaptive) behav-
iors. Thus, patients who are getting better in therapy can expect to have it
continued and patients who are failing to improve should get less of the cur-
rent types of therapy than they have been receiving. One of the important
guidelines for families is that they should always react to (never ignore) self-
destructiveness or other crises, but they should not overreact (see Chapter
8). This principle exists basically to prevent patients’ sense of being ignored
or neglected but not to gratify (reinforce) the patients’ expectation that cri-
ses or self-destructiveness can successfully be used to obtain power or sadis-
tic control or to revive affection (all dynamic interpretations).
Of critical importance is that when responses by therapists or others
that have reinforced negative behaviors are discontinued, the discontinu-
ation must be sustained because they will initially evoke what behaviorists
know as an extinction burst—that is, the unwanted behavior will increase be-
fore it decreases. A common example is that when a therapist changes his
or her policy about accepting telephone calls between therapy sessions, a
burst of telephone calls will occur before the new policy is accepted. An-
other example: when a parent who has been an enabler (e.g., buying cig-
arettes for his or her child) says that he or she is going to stop, a redoubled
burst of pleas, demands, and dramatic “withdrawal symptoms” is a predict-
able consequence. The occurrence of such extinction bursts is evidence
that the recently initiated change is potentially powerful—if sustained.
Dialectical Behavior Therapy

Since 1987, the single most remarkable entry in therapeutic strategies for
borderline patients is a package of cognitive-behavioral strategies called di-
alectical behavior therapy (Linehan 1987a, 1987b). Basic DBT is a form of an
intensive outpatient program (level II; see Chapter 5) that has been the
subject of numerous articles and two seminal books by its developer, Mar-
sha Linehan, a behavioral psychologist. Workshops involving 10 days of
training in DBT are oversubscribed in the United States, Europe, Australia,
and Asia. Some state departments of mental health and some managed
care companies have endorsed DBT as a treatment of BPD, and some of

these payers even protest payment for practitioners who offer non-DBT
therapies. Clinical programs devoted to providing DBT are springing up
throughout the United States and Europe. Major clinical trials that test its
application in a variety of settings have been conducted here and abroad.
Swenson (2000) has suggested that the widespread dissemination of
DBT reflects the hunger of the mental health community for a treatment
that is clear, understandable, and seemingly learnable. This coincides with
the hunger of funding agencies to invest in therapies with scientifically sup-
ported effectiveness. Without question, DBT has inspired enthusiasm for
treating BPD in patients who have often demoralized prior generations of
clinicians. Still, at the heart of why DBT has become so widespread is the sim-
ple fact that patients like it and find it helpful. Not all, but most, borderline
patients endorse it because it is a practical, respectful, and hopeful therapy.
Theory
Linehan (1993a) proposed that the core psychopathology of borderline
patients involves a biologically based failure of emotional regulation that has
interacted with what she perceives as a socially pervasive invalidating envi-
ronment. The biological side of Linehan’s biosocial theory echoes theories
that have been prominent in the psychiatric literature since the mid-1970s
(Akiskal 1981; Klein 1977; Stone 1980) and subsequently given scientific
substantiation by Siever and Davis (1991) as a type of psychobiological dis-
position to BPD (see Chapter 1). Linehan differs, however, by positing
that the emotional problems are not anchored in or reflective of what psy-
chiatrists call mood disorders. The social side of her theory picks up on a
theme that is part of the clinical and research literature about BPD fami-
lies. Invalidation reflects 1) the emphasis given to the marked discrepan-
cies between the borderline patients’ perceptions of themselves and their
parents’ perceptions and 2) the lack of communication about these dif-
ferences (Feldman and Guttman 1984; Gunderson and Lyoo 1997; Gun-
derson et al. 1980; Shapiro 1982; D.W. Young and Gunderson 1995).
The emphasis in Linehan’s biosocial theory on emotional dysregula-
tion as the core BPD psychopathology is consistent with the DBT focus on
maladaptive behavioral symptoms (e.g., impulsive and inappropriate ex-
pressions of emotions). This focus accords with the use of DBT for ac-
tively self-injurious patients whose maladaptive behaviors are believed to
function as escapes from or expressions of negative emotions.
Basic Services (Individual Plus Group)

Although DBT is a set of behaviorally based treatment principles of direct
relevance to BPD that can be implemented via a range of formats (e.g.,
groups, milieus, families), the basic DBT package described in this sec-
tion is the one that has had empirical support. The package has three
components: 1) once-weekly psychotherapy with a trained primary clini-
cian or therapist whose work is coordinated with 2) a weekly 2.5-hour
skills training group led by trained coleaders; both of these services are
backed up by 3) telephone consultations with the primary therapist or, if
the therapist is unavailable, by arranged coverage. The point of tele-
phone contact is to prevent emergencies by providing skills coaching and/
or relationship repair. Only as a last resort does the therapist use tele-
phone contact for assessing and managing emergencies. Of note is that a
fourth nonoptional component of DBT does not include patients. A
weekly consultation meeting of the three members of the team is held to
ensure adherence to the procedures and to diminish countertransfer-
ence problems. Thus, each patient is receiving 3.5 or more hours of di-
rect contact (6 hours of therapist time) and an additional 3 hours of
indirect therapist time each week. As has been noted elsewhere (Chapter
4), split treatment (use of multiple therapists) is deeply embedded in the
structure of DBT (i.e., two modalities and coleaders of the skills group
with required coordination of the team). These components offer a very
structured, coordinated type of intensive outpatient program (see Chap-
ter 5) that is intended to be comprehensive. Patients treated within Line-
han’s published research protocols are actively discouraged from using
or relying on other therapies, such as hospitals or medications.
The group therapy component consists of a weekly 2.5-hour social
skills training. The course is composed of four social skills modules: mind-
fulness, emotion regulation, distress tolerance, and interpersonal ef-
fectiveness. A clearly written accompanying manual guides therapist
interventions for each of these modules and provides text (with home-
work) for patients. Much of what DBT groups include has been offered by
preexisting cognitive-behavioral therapies (e.g., contingency manage-
ment, exposure, training in assertiveness or social skills, cognitive schemas
as triggers). In DBT, these components are packaged in an integrated,
learnable way, and they connect nicely with the agenda in the individual
therapy. Most borderline patients find the structured educative format
more useful and less stressful than the more expressive goals of interper-
sonal groups (see Chapter 9).
The individual psychotherapy component of DBT addresses a hierar-
chy of target problems, giving priority to self-destructive behaviors, “ther-
apy-interfering” behaviors, and problems of daily living (see Table 11–1
for first-stage targets).
In accordance with DBT theory that invalidating environments are
pathogenic, the therapist is actively supportive and specifically emphasizes
TABLE 11–1. Hierarchy of first-stage targets in dialectical

behavior therapy
1. Decreasing suicidal behaviors
2. Decreasing therapy-interfering behaviors
Nonattendance
Noncollaboration
Noncompliance
3. Increasing quality-of-life-improving behaviors
Behavioral skills
Mindfulness skills
Interpersonal effectiveness
Emotion regulation
Distress tolerance
Self-management
Source. Adapted from Linehan 1993a.
validation of the patient’s feelings, much as advocated by Adler (1986)

and Stevenson and Meares (1992), who deploy Kohut’s self psychology
(see Chapter 12). The contract involving contingencies (see Chapter 10)
recognizes the therapeutic necessity for protecting the limits and bound-
aries of the therapist, echoing a theme within the more general psychiat-
ric literature (Chapters 3 and 4) and psychoanalytic literature (Chapter
12). In other respects, such as the structuring of the format and content of
sessions and the unapologetic use of directive and educative techniques,
this therapy is very different from traditional individual dynamic thera-
pies. The dialectical tension between acceptance strategies such as valida-
tion and change strategies such as contingencies and homework is akin to
the admixture of support and frustration found in dynamic therapies. In a
small process study, the advantage of having both acceptance and change
techniques was confirmed (Shearin and Linehan 1992).
The following vignette illustrates ways in which DBT’s individual ther-
apy can be distinguished from traditional dynamic therapies.
Vignette
Clinician (after greeting patient warmly): Let’s start by looking at your di-
ary card. [Therapist reviews the patient’s daily diary to assess suicidality,
self-care, and periods of misery.]
Patient: It’s been a good week—uh, 2 weeks actually.

Clinician: That’s good. So tell me the reason you weren’t here last
week. [Inquiring about the second priority, therapy-interfering behav-
ior.]
Patient: I overslept—had been unable to get to sleep, so finally I took
a sleeping pill. I didn’t want to, and I felt terrible when I woke up.
Clinician: That’s a tough spot to be in. [Validates patient’s dilemma.]
Did you remember that you had an appointment?
Patient: Not until after I woke up. Then I knew I wouldn’t be able to
get here.
Clinician: How late did you sleep? [Assessing feasibility of her attend-
ing.]
Patient (irritated): I don’t remember.
Therapist: You sound irritated.
Patient (calmer): Well, you make it seem as if I didn’t want to come.
Therapist: Well, I’m sorry you missed. [Placating self-disclosure.] It’s
good that you got the sleep you needed. Now let’s discuss ways to see what
we can do to prevent this from happening again. [Invites collaboration on
problem solving.]
This brief and seemingly unremarkable interaction shows six ways in

which DBT is distinguished from a dynamic therapy:
1. The DBT therapist structures the start of the session by asking to re-
view the patient’s diary card to assess suicidality (priority 1) and ex-
treme emotional experiences. A dynamic therapist more typically
would wait for the patient to identify what she wants to talk about.
2. The DBT therapist then inquires about therapy-interfering behavior
(priority 2): the missed appointment. This might, and should, be pur-
sued in a dynamic therapy too but not as a standardized priority. More-
over, a dynamic therapist would probably make either a more open-
ended inquiry such as “What’s been going on?” or a negative transfer-
ence inquiry such as “Did your absence relate to our past session?”
3. In response to the patient’s saying that she had taken a sleeping pill to
sleep and then had felt “terrible” when she awoke, the DBT therapist
offers an empathic and validating response. Although some dynamic
therapists might do this instinctively, common responses would be to
make further inquiries about the insomnia (e.g., “What was on your
mind?”), the medications (e.g., “Were you reluctant to use the sleep-
ing pills?”), or the feeling bad (e.g., “What did you feel bad about?”).
All of these inquiries would be based on the goal of adding meanings
to the patient’s understanding of these events.
4. When the patient gets irritated, the DBT therapist inquires about the
in-the-moment interaction, with the rationale that the irritation
could, if unattended to, become a problem interfering with therapy
(priority 2). A good dynamic therapist also would inquire about the
irritability but would view it as potential material for a transference-
countertransference analysis.
5. When the DBT therapist discloses that she is sorry the patient missed
the appointment, she reveals that she wants the patient to come and,
by inference, that their sessions have value. The DBT therapist can be
presumed to have assessed the risk that the response would reinforce
the patient’s absences and to have concluded that it would not. A dy-
namic therapist might be hesitant to make the response because it
forecloses an opportunity to explore how the patient expected the
therapist to feel—that is, to examine transference, or at least the ther-
apeutic relationship.
6. The DBT therapist engages the patient in developing alternative ways
of coping with the problem. The target problem is defined behav-
iorally as the missed appointment, and a functional analysis—the
function served by missing, for example, a good night’s sleep—is un-
derscored. This validates the patient’s need for self-care. A dynamic
therapist would be hesitant to adopt a proactive, “what can be done”
approach, fearing it would enact a parental transference, and the
therapist would not define the problem behaviorally but would see it
as an issue of conflict and motivation. Moreover, an analysis of the
missed appointment would be less likely to begin by the therapist’s
identifying that it served useful functions.
Empirical Support
Basic DBT (the combination of weekly individual and weekly group ses-
sions) has consistently established its ability to diminish patients’ deliber-
ate self-harm and suicidal behaviors and both emergency department
and hospital use significantly more than was observed in similar BPD pa-
tients who received treatment as usual (Linehan et al. 1991, 1993). These
results were confirmed in independent randomized controlled trials by
Turner (2000), Koons et al. (2001), and Verheul et al. (2003). The com-
parison treatment in these trials (treatment as usual) largely consisted of
medication management, intermittent counseling, and hospital emer-
gency services.
In recognition of the weakness of this control condition, Linehan et
al. (2006) subsequently completed a study comparing DBT with individ-
ual therapy that was conducted by experienced and enthusiastic nonbe-
havioral therapists. Results of that study confirmed the superiority of
DBT in the same outcome domains.
Within the health care system, basic DBT represents a particularly co-
herent and empirically validated form of level II, intensive outpatient
care. As noted in Chapter 5, this level of care offers the minimal and
therefore optimal level of care for actively self-destructive borderline pa-
tients. Among DBT’s contributions are documenting the advantages of
mandatory split treatment (see Chapters 3, 4, and 5), mandating ongoing
supervision, and establishing from the onset that the therapy is for the
purpose of changing (although, dialectically, the need to change is coun-
terbalanced by acceptance of the reasons that patients resist change). No
one before Linehan has been as direct and insistent in offering this chal-
lenge and this hope. It follows that goals need to be established and that
the failure to make progress reflects poorly on the treatment.
Therapists
DBT also has established standards for competence by therapists. The
training considered to be necessary to provide DBT (two 5-day work-
shops) with intervening homework can be obtained through Linehan
herself and several other authorized trainers. Still, Linehan (rightly,
I think) does not feel confident that many therapists, even on comple-
tion of these workshops, can administer high-quality DBT, and she hopes
to develop true credentialing criteria (see Chapter 10, Sidebar 10–1). In
this respect, Linehan would set a new and higher standard. One study in-
dicated that 109 clinicians with diverse experience and training and roles
trained by a state department of mental health could acquire reasonable
intellectual mastering of DBT (Hawkins and Sinha 1998). I suspect that cli-
nicians who have trouble being active, directive problem solvers and
those who are deeply wedded to psychodynamic explorations find it
hardest to adhere to or become competent in DBT. On the contrary, it is
my impression that well-trained cognitive-behavioral therapists who are
experienced in working with BPD patients and temperamentally com-
fortable being active and directive can learn to administer the individual
therapy component of the DBT manual capably, even without the inten-
sive workshops.
Expanding Applications
DBT’s usefulness has been examined in broadening patient groups and
in different treatment settings (Chapman 2006; Koerner and Dimeff
2000; Lynch et al. 2003). DBT has been shown to be effective for BPD pa-
tients with some types of substance use disorder (i.e., both opioid depen-
dence [Linehan et al. 2002] and alcohol abuse [McMain 2004]).
Interestingly, the benefits have not been shown for borderline patients
with other forms of substance use disorder (McMain 2004; Verheul et al.
2003). Whether borderline or not, patients with binge-eating disorder
have shown benefits from modified forms of DBT; in one trial, only 20
sessions were used (Telch et al. 2001), and in another, only individual
DBT was used (Safer et al. 2001). Modifications of the basic DBT services
not yet subjected to randomized controlled trials are also finding appli-
cations for families (Hoffman 1999; Hoffman and Hooley 1998) and ad-
olescents (Miller et al. 1997) and for both inpatient (level IV) (Bohus et
al. 2000; Springer and Silk 1996) and partial hospital (level III) services
(Simpson et al. 1998). All these applications and others that are in
progress require modifications of the basic DBT package.
The need remains to establish whether, if properly applied in “real-
world settings,” DBT can help reduce suicidal behavior and improve
functioning (i.e., to establish whether it can be a clinically practical and
cost-effective treatment outside academic clinics). To address this issue,
McMain and colleagues at the University of Toronto have undertaken a
large randomized controlled trial involving the cost-effectiveness of DBT
in real-world settings by comparing DBT with consensus-based best prac-
tices therapy (McMain 2004). For this study, McMain and colleagues
adapted the American Psychiatric Association’s (2001) practice guide-
line for the treatment of patients with BPD (Oldham 2005) for a general
hospital psychiatric outpatient setting. This therapy, called general psychi-
atric management, consists of three components: 1) dynamically informed
psychotherapy, 2) case management, and 3) structured algorithmic med-
ication management. This therapy was thought to reflect high-standard
yet “typical” outpatient care provided by a multidisciplinary team associ-
ated with a general hospital psychiatric program. The Toronto study will
test whether DBT’s considerable direct service costs are offset by reduc-
tions in other health service costs. This will be important data to guide
mental health service planners.
Separating Individual and Group Components

One study that used only the individual therapy component showed that
it led to more reductions in parasuicide, impulsivity, anger, and global
mental health than did a so-called client-central individual therapy
(Turner 2000). Although this conclusion needs further testing, the re-
sults when the skills training group was offered alone seem less encour-
aging.
Many BPD patients will not persevere when a DBT skills group is of-
fered alone (Linehan 1993b) unless the primary clinician-therapist
makes their attendance mandatory (just as they will leave other groups;
see Chapter 9). Therapists often need to take the position that the skills
learned in DBT group are necessary. Beyond therapists’ being insistent
and supportive, it is important that they be familiar with and reinforce
the messages taught in DBT.
In an unpublished study in which half of a sample of borderline pa-

tients who were receiving ongoing, non-DBT types of individual therapy
were randomly assigned to receive the skills training group component of
DBT, the skills group seemed to add little (study cited in Linehan and
Heard 1993). Even though DBT skills groups without DBT individual
therapists would be expected to diminish the group’s potency, I share
Swenson’s (1989) surprise at its complete ineffectiveness. I have been
repeatedly impressed by the effectiveness of the skills training group
when combined with individual therapists who are not DBT trained.
Moreover, other types of groups based on social learning theory have
been reported by others (Goodpastor et al. 1983) to be effective for bor-
derline patients. I suspect that the reported failure of the skills group
without a DBT individual therapist was because the mandates that partic-
ipants do assignments, complete homework, and make changes were not
actively reinforced by an individual therapist. An actively supportive indi-
vidual therapist may be essential for holding patient projections (see
Chapter 4, section “Splits, Splitting, and the Virtues of Split Treatments”)
and thereby greatly reducing dropouts.
Patients can comfortably use DBT while participating in other thera-
pies, such as the medication, family, and rehabilitation modalities. Swen-
son (1989), however, warns that combining the DBT skills group with
dynamically based psychotherapy, especially intensive therapies (more
than twice weekly), should be undertaken with caution. The supportive,
validating aspect of the DBT approach—like that encouraged by the Stone
Center’s “relational therapy” (see Jordan et al. 1991)—will clearly invite an
idealized split when a patient contrasts it with the less-structured, nondi-
rective, and sometimes frustrating approach of transference-based thera-
pies. Another type of split also may occur. The expectation in DBT of
learning and then applying new skills in living may become discounted
when a patient contrasts this with a dynamic therapist, who can become
idealized by virtue of noncontingent listening without explicit expecta-
tions of behavior change. Whereas a dynamic therapist sees self-destructive
acts as generated by interpersonal stressors, the DBT therapist sees them as
responses to intense emotions (Kehrer and Linehan 1996). Such messages
are not incompatible, but without thoughtful consideration, such differ-
ences can seed harmful splitting (as noted in Chapter 4). The success of
this or any combination of therapies depends on whether the therapists
have goals in common, mutual respect, and frequent communication.
Vignette
A patient called her psychodynamic therapist on a Friday evening because
she didn’t “know whether to continue living is worthwhile.” Her therapist
listened empathically, but after noting the patient’s improved mood and
her turn to more cheerful and prosaic topics, he then indicated that he
needed to go. The patient said, “Oh, I’m sorry. I should have guessed
you’d be busy.” About 15 minutes later, she called her DBT therapist with
the same concerns, and this therapist coached her to use self-soothing
skills.
The following week, the patient and her two therapists met to clarify
their roles. The dynamic therapist proposed that because the patient
seemed to feel better after being listened to by him, she would benefit
from trying to understand why this was so. The patient agreed that being
listened to helped, but she added that being coached by the DBT-trained
therapist was actually more useful. The DBT therapist believed that the
improved coping skills that derived from DBT offered more hope for
change. The patient agreed with the DBT therapist, and they all agreed
that the DBT therapist would have the role of responding to the patient’s
safety concerns in the future.
This vignette illustrates the very distinctive theories and responses of

a DBT therapist and a psychodynamic therapist. The DBT therapist be-
lieved that by virtue of the specific coaching DBT can offer, he should as-
sume the primary clinician’s role. Discussion with the patient supported
that. The dynamic therapist, however, continued to believe that the pa-
tient’s relief came from caring attention and had little to do with skills the
DBT therapist was coaching. Although the prospect of fewer telephone
calls was welcome to the dynamic therapist, he believed that the new plan
would 1) make it less likely that the patient would become aware of what
motivated her calls for help; 2) make the patient less insightful about her
need for a dependent, secure attachment; and 3) perhaps make it less
likely that this form of corrective relationship would evolve. The DBT
therapist believed that the patient would learn to generalize the use of
new coping skills and that that would allow her to need others’ help less
frequently.
DBT and dynamic therapies can be complementary, but that relation-
ship seems to rely on time-consuming efforts at collaboration. Basic DBT
(group and individual components combined) appears to be highly useful
in preparing patients for subsequent psychodynamic therapies by increas-
ing patients’ ability to recognize and tolerate emotions without action, a
possibility recognized by Linehan (1997). DBT may make possible the
emotional expression, interpersonal confrontations, and transference
analysis that intensive dynamic therapies traditionally involve. Also, when
the patient has mastered the social skills modules in the group component
of DBT, the advanced forms of individual DBT currently being developed
may offer further opportunities for personal growth; the designated tar-
gets are the effects of trauma and improving the quality of life.
Limitations
DBT’s endorsement should be qualified by respectful recognition of its
limitations. First, there is a risk that DBT will become prematurely reified
(it has had only modest revision since its inception) and that its advocates
could become a cult organized around special knowledge as opposed to
an incremental advance within a wider mental health culture. A prece-
dent for this can be found within psychoanalysis.
Second, it is important to remember that BPD patients can get better
from other approaches, as dramatically illustrated in more recent longi-
tudinal studies. “Remission” frequently occurred without DBT or any
other coherent BPD-specific treatment (Skodol et al. 2005; Zanarini et al.
2003) and even after short interventions or situational change (Gunder-
son et al. 2003). Improvement of BPD was also evident in the trial in
which DBT had about the same results as did two other forms of manu-
alized therapy (Clarkin et al. 2007; see Chapter 12).
Third, DBT does not work for all BPD patients. Bohus et al. (2000)
showed that a significant fraction of inpatient BPD patients did not get
better. Patient suitability depends on acceptance of the BPD diagnosis
and on wanting to change the emotional, behavioral, interpersonal, and
self-awareness issues to which DBT is specifically addressed. Like most
other psychotherapies, DBT also requires that patients have the intellec-
tual ability to grasp and remember the concepts—a prerequisite that
would be troublesome for patients with learning difficulties. Suitability
for the group component also requires that patients be able to tolerate
sharing attention and listening to others speak of their problems without
getting so disturbed that they leave or become disruptive.
Overview
Because the still growing body of empirical support for DBT already ele-
vates it above other treatment approaches, borderline patients should be
advised to seek this treatment if it is available. DBT has rapidly become
widely used and is sometimes advocated as a standard of care by Linehan
and increasingly (as noted earlier) by managed care companies and state
departments of mental health. As noted in the preceding discussion,
such a policy seems dangerously premature. Finding DBT’s appropriate
place within the larger framework of other mental health services will re-
quire time and more data. Still, DBT’s place in the therapeutic armamen-
tarium for BPD is secure and can be expected to grow. In the process of its
growth, its place within a multimodel context will be clarified, and refer-
rals to it will become more discerning. Linehan and other DBT experts
are examining the process by which change occurs. This perspective will
help clarify areas of overlap and distinction between DBT and alternative
cognitive and psychodynamic treatments and theories (e.g., Swenson
1989; Westen 1991).
Cognitive Therapies
Cognitive therapy clinicians postulate that borderline patients have dis-
turbed cognitions that 1) develop early in their lives, 2) have maladaptive
consequences, 3) are self-perpetuating, and 4) are, although difficult to
change, the targets for cognitive therapies. The chief difference from be-
havioral therapies is that behavioral therapies focus on behaviors, whereas
cognitive therapies focus on changing dysfunctional cognitive schemas
about oneself and one’s environment. Although the distinction between
cognitive approaches and behavioral approaches seems to disappear
when applications are described, in principle cognitive therapies rest on
the idea that behavioral (including interpersonal) problems are mediated
by disturbed thinking. For example, a deliberate self-destructive action
would be seen as an outgrowth of a disturbed cognitive schema, such as
the view of oneself as bad (Layden et al. 1993). The patient would be
taught to recognize how this schema triggers self-destructive acts. The act
itself would be identified as a decision, and the patient would be encour-
aged to consider options. The act might also warrant directives (e.g.,
“Stop. Consider that it’s not good for you and doesn’t get you what you
want.”). At this point, of course, the cognitive approach deploys behav-
ioral techniques.
Beck and Freeman (1990) stressed three basic disturbed cognitions
held by borderline patients:
1. The world is dangerous and malevolent.

2. I am powerless and vulnerable.
3. I am inherently unacceptable.
Beck also believes that dichotomous thinking is particularly common

and problematic in borderline patients, who tend to perceive people,
feelings, or issues in terms of mutually exclusive categories rather than
seeing them as part of a continuum (Beck and Freeman 1990; Beck et al.
2004). He believes that the extreme (i.e., dichotomous) evaluation of sit-
uations is a cognitive handicap that leads borderline individuals to their
extreme emotional responses and actions (in contrast to positing these
individuals’ basic failure of emotional or impulse regulation)—responses
and actions that are then accompanied by rapid shifts to the opposite
view. Beck postulates that the basic assumptions, the dichotomous think-
ing, and an unstable sense of identity are the particularly problematic

cognitive features of borderline patients. Beck suggests that reducing
and eliminating dichotomous thinking is an early goal of therapy. Still, he
notes that the cognitive problems are so “encrusted” that a year is more
likely to be needed than the 12–24 sessions that can be effective with
other diagnostic types.
Leading cognitive therapists recognize that borderline patients’
transference offers a way to identify cognitive schemas that can control
their relationships (e.g., “My therapist wants to control me”) or can be
the triggers for their reactions within relationships (e.g., a therapist’s late-
ness triggers abandonment fears). J. E. Young (1994) and Beck et al.
(2004) even suggest that cognitive therapists may need to do work (cog-
nitive therapy) on themselves because of their countertransferences. All
the pioneering cognitive therapists give attention to the therapist’s need
to invest energy in establishing a trusting and collaborative working rela-
tionship.
Westen (1991) has helped to bridge the conceptual and clinical gaps
between the cognitive-behavioral and psychoanalytic approaches. Joining
Beck and Freeman, he reconceptualized the borderline patient’s defenses
(e.g., splitting and projection) as disturbed cognitive patterns. Whereas
Beck and Freeman (1990) refer to splitting as “dichotomous thinking,”
Westen identifies it as a misattributional style involving global or polarized
judgments about self or others. In this scheme, projection is seen as mis-
attributing one’s own motives to others. Westen then describes how cog-
nitive-behavioral techniques can be used within psychoanalytic therapy to
diminish the borderline patient’s overelaboration of affects and impulsive
behaviors. Essentially, Westen advocates a cognitive technique of labeling
(i.e., highlighting or setting apart) affect states (e.g., anger), words (e.g.,
“horrible” or “perfect”), or cognitions (e.g., “He distrusts me”) that can
subsequently be referred to as signals or red flags that will help that pa-
tient stop and think before reacting.
The description of empirically supported cognitive therapies that fol-
lows represents the most refined applications of cognitive-behavioral
therapy to BPD. It is notable, however, that other cognitive therapists are
currently exploring alternative concepts and techniques (Bienenfeld
2007; Paris 1994).
Schema-Focused Therapy
Although Jeffrey Young’s schema-focused therapy (SFT) has been in use
and in ongoing development since 1990 (J. E. Young 1994), it has re-
ceived relatively little attention within the BPD treatment literature until
the recent report of a randomized controlled trial that confirmed its ef-
ficacy (Giesen-Bloo et al. 2006). All the borderline patients did well in
this trial, but those receiving SFT did better than an alternative manual-
ized therapy (i.e., Kernberg’s transference-focused psychotherapy; dis-
cussed in Chapter 12). This randomized controlled trial put SFT firmly
on the map of effective options in treating BPD (Sidebar 11–1).
Sidebar 11–1: Schema-Focused Therapy:

Does It Work?
In a multicenter study from The Netherlands, 88 outpatient
borderline patients were randomly assigned to a con-
trolled trial comparing prolonged (3 years) individual psy-
chotherapies of two types: SFT and transference-focused
psychotherapy (Giesen-Bloo et al. 2006). Patients were
seen twice weekly by master’s-level psychologists with ex-
perience and commitment to their respective treatment.
In both cells, the therapists were trained and supervised by
advocates and experts. In the case of SFT, Jeffrey Young,
the treatment’s developer, performed this function. For
transference-focused psychotherapy, Frank Yeomans,
who had worked for many years with Otto Kernberg and
who had a major role in its manualization, performed this
function.
Outcomes were measured every 3 months but were
reported using survival analyses. Patients in both cells
showed significant reduction in borderline psychopathol-
ogy and improvement of quality of life and psychological
and personality traits that included more specific targets
for the respective therapies (e.g., self concepts, self-other
distinctions). Significant benefits were already evident in all
three domains for both forms of therapy by 1 year. How-
ever, by 3 years in all three major outcome domains, SFT
was significantly better than transference-focused psycho-
therapy—that is, in improvement of quality of life and in re-
duction of borderline psychopathology (especially in the
interpersonal and behavioral domains) and maladaptive
psychological and personality traits. SFT also outperformed
transference-focused psychotherapy in the slope (rate) of
improvement and had significantly fewer dropouts.
Some reservations about these conclusions are war-
ranted. First, a relatively healthy BPD sample was studied—
young outpatients who were willing to endure a 2-month
wait. How well these therapies would do if they were initi-
ated with more severely or more acutely disturbed border-
line patients is unclear. Second, generalizability was limited
by the length of the treatments. Three years is a long-term
therapy, and 61% of the SFT and 45% of the transference-
focused psychotherapy patients were continuing in their
respective treatments after 3 years. Insofar as many bor-

derline patients show clinically significant improvements
(more than 50% remit by 3 years) with far less specialized
therapies (Skodol et al. 2005; Zanarini et al. 2005), and this
study did not include a low-intensity treatment cell, the
added value of these specialized treatments still needs to
be established. Third, two significant factors probably bi-
ased the results in favor of SFT. The transference-focused
psychotherapy sample had more suicidality and self-harm
at baseline. This could have disrupted the transference-
focused psychotherapy contract setting and contributed
to more dropouts in this cell. In addition, fewer than half of
the transference-focused psychotherapy therapists were
adherent to the manual, whereas nearly all of the SFT ther-
apists were.
Despite these reservations, this study provides a signifi-
cant endorsement for the use of SFT. Jeffrey Young’s ther-
apy is now a significant contender in the increasingly
competitive arena for “best treatment” of BPD.
THEORY
A schema is a pattern imposed on one’s experience to help explain or in-
terpret it. Cognitive schemas develop early in life, and for people with
personality disorders, they are often inflexible and maladaptive. Ten spe-
cific maladaptive cognitive schemas for BPD that overlap with Beck’s have
been proposed by J. E. Young (1990): abandonment and loss, unlovabil-
ity, dependence, subjugation, lack of identification, mistrust, inadequate
self-discipline, fear of losing emotional control, guilt and punishment,
and emotional deprivation. These schemas are evaluated by a question-
naire and then examined to determine what triggers each schema, how it
is maintained, how it is avoided, and what would be an alternative and
more adaptive way of behaving.
In the manual for SFT, J. E. Young et al. (2003) move from these 10
cognitive schemas to describe five “modes” that characterize BPD. Modes
refer to those sets of schemas that are currently active states of mind that
govern a person’s immediate attitudes, responses, and behaviors. Young
developed this concept of modes out of his experience with borderline
patients because “the number of schemas and coping responses they
had—and their continuous shifts between them—was overwhelming”
(J. E. Young et al. 2003, p. 40). In doing this, Young’s concept has notable
similarities to the dynamic alternation between mental states that I have
proposed results from whether a borderline patients feels “held” (see
Chapter 1) and to the “self states” that another cognitive theory, devel-
oped by Ryle (2004), identifies in BPD patients. Ryle’s three self states in-
clude a deprived or an abused self that is vigilant for mistreatment, an

overly emotional self that defensively dissociates (detaches), and a self
that is unable to mentalize (i.e., to read oneself or others).
Table 11–2 describes the modes that typify BPD and how these might
relate to the DSM criteria. The first mode, Abandoned Child, is considered
the most central and useful perception for therapists to sustain their BPD
patient, and it is the mode patients are encouraged to spend much of
their therapy time within. The fifth mode, Healthy Adult, in borderline pa-
tients is very weak and undeveloped. When stressed, borderline patients
most often use the Detached Protector mode (Arntz et al. 2005).
BASIC PRACTICES
The Healthy Adult mode is initially and primarily a function served by the
therapist; by instruction and modeling, the therapist encourages the pa-
tient to internalize and grow in the ability to protect and soothe himself
or herself. This process is labeled “reparenting.” Therapists maintain
clear boundaries and set limits, but they are encouraged to reassure pa-
tients about caring, about their not leaving, and about the patients’ in-
herent goodness and to help patients stand up to the punitive parents in
their introjects and histories.
SFT claims to integrate behavioral, dynamic, attachment, and Gestalt
models. Therapists are quite active and interactive. They encourage pa-
tients to express needs and emotions (leave the Detached Protector mode);
learn coping skills; read written materials (by Young) about their sche-
mas; develop flash cards; actively play-act dialogues between modes or
with the therapist taking on one of the modes; and close their eyes, visu-
alize, and describe difficult experiences, including, eventually, trauma.
Altogether, SFT is an inventive, structured, and highly involving ther-
apy. The manual contains clear and clinically compelling characteriza-
tions of the borderline patients’ experiences. Suicidality and deliberate
self-harm are expected to diminish within 6 months, punitive introjects
are expected to soften by 1 year, and a capacity to sustain close relation-
ships can be expected by 3 years.
CAVEATS
Still, SFT prompts some questions. First, like DBT, SFT has its own con-
cepts and language, which, although understandable, may require con-
siderable immersion for therapists or patients to acquire the necessary
comfort and commitment levels. Again, like DBT, the therapy’s develop-
ers wrote as if they were unfamiliar with or disregarded the very extensive
prior literature about psychotherapies with borderline patients. Little ef-
fort was made to identify the considerable overlaps with prior contribu-
TABLE 11–2. Core schema modes of borderline personality

disorder
1. Abandoned Child—helpless, frightened, sad, no one cares,
alone, lost.
Related criteria: abandonment
2. Punitive Parent—attacking self, punishing self, “I’m bad,” harsh,
unforgiving.
Related criteria: deliberate self-harm, impulsivity
3. Detached Protector—disconnected/blocks feelings and needs,
avoids, distracts, dissociates.
Related criteria: abandonment, impulsivity, empty, identity
4. Angry Child—temper tantrums, rages, feels mistreated, rejected.
Related criteria: anger, impulsivity, affective instability
5. Healthy Adult—meets child’s unmet needs (e.g., protects the
Abandoned Child; limits the Angry Child).
Criteria of affective instability and identity both involve switching from
one mode to another.
Intensive, unstable relationship is particularly related to switching
between modes 1 and 4.
Source. Adapted from J.E. Young et al. 2003.
tions. Thus, the required immersion within SFT’s new concepts and
language may lend itself to cultism. Second, the explicit “reparenting”
function of therapists could encourage regressive transference enact-
ments. This danger is heightened by the apparent uncritical acceptance
of patients’ devaluative attitudes toward their actual parents and by the
lack of any family contacts or involvement during SFT.
OVERVIEW
SFT has emerged as a promising form of individual psychotherapy for
borderline patients. It has a vivid and understandable conceptualization
of borderline psychopathology, and it has proved to be teachable. Repli-
cation studies of its efficacy and further explication of its principles and
techniques are needed and will certainly be welcome.
Other Cognitive Therapies

Because cognitive therapies bridge the internal perspective of psychody-
namic theories with the external perspective of behavioral theories, they
may represent the arena where synthesis is most likely to occur. The ar-
rival of SFT is one indication of this synthesis, but other cognitive thera-
pies are also gaining a foothold. Notably, Anthony Ryle, a British
psychologist, has been developing a “cognitive analytic theory and ther-
apy.” Another cognitive approach, Systems Training for Emotional Pre-
dictability and Problem Solving, was described in Chapter 9. In the
following subsection, I describe another promising cognitive therapy
with preliminary empirical support.
MANUAL-ASSISTED C OGNITIVE TREATMENT

An interesting and potentially very valuable short-term therapy has been
developed in England that specifically targets deliberate self-harm and
other suicidal behaviors. Manual-assisted cognitive treatment (MACT)
was developed by two U.K. psychologists, Schmidt and Davidson (2004),
as a six-session therapy that incorporates elements of Linehan’s DBT,
Beck’s cognitive-behavioral therapy, and bibliotherapy (Scogin 2003).
Each session is structured around a chapter of a booklet covering 1) psy-
choeducation on self-harm and suicide attempts as well as a functional
analysis of the specific episodes of parasuicide (deliberate self-harm and
suicide attempts) of the patients; 2) emotion regulation strategies, in-
cluding crisis plan, distraction techniques, self-soothing techniques, or
ways to improve the moment; 3) problem-solving strategies; 4) cognitive
restructuring strategies and management of negative thinking; 5) man-
agement of substance use; and 6) relapse prevention strategies, including
advantages and disadvantages of engaging versus not engaging in delib-
erate self-harm, acceptance strategies, and relapse prevention plan.
Its brevity and format make MACT an attractive intervention with sig-
nificant public health implications, but its efficacy remains uncertain.
Two trials were conducted in England. The first failed to have much ef-
fect on parasuicidal behaviors but did decrease depression and hospital-
izations and improved future-oriented thinking (Evans et al. 1999). This
result prompted a larger multisite trial involving random assignment of
480 parasuicidal patients. This study also failed to show a benefit from
MACT compared with treatment as usual (Tyrer et al. 2003). Moreover,
when a subsample of 67 patients with BPD were examined, they actually
had a shorter time to parasuicide.
These rather discouraging results might have spelled the end of
MACT; however, trials conducted at McLean Hospital with a revised ver-
sion of MACT (i.e., McMACT) produced quite significant benefits
(Weinberg et al. 2006). These benefits may have been a result of the al-
legiance effects of having investigators who were therapists, narrowing
the selection to BPD patients with repeated deliberate self-harm (exclud-
ing those who were actively suicidal), or patients continuing treatment
for the entire six sessions. Notably, in the two British trials, the therapists’
motivation and competence were highly variable, and, remarkably, more
than half of the patients attended fewer than three sessions before drop-
ping out. In any event, these preliminary McMACT studies indicated that
this short-term intervention may yet prove valuable when therapists and
patients are selected appropriately.
The results of these trials combined with the evidence from longitu-
dinal research showing that BPD can remit quite quickly (Gunderson et
al. 2003) serve notice that focused short-term interventions may have a
more significant role in treating BPD than had ever been expected.
Summary
Cognitive-behavioral approaches to the care of borderline patients have
moved from the background of modalities into the foreground. The re-
markable empirical substantiation for DBT has excited a new cadre of en-
thusiastic clinicians, a public health debate about reimbursement, a new
standard for assessing competence, and a new intellectual ferment about
mechanisms of therapeutic action within therapies. DBT is now joined by
Young’s SFT, which also has received promising empirical validation. In
this chapter, I have attempted to place these treatments into their context
alongside other modalities and within the history of treatment develop-
ments for BPD. Other cognitive therapies that are less ambitious, but
more easily exportable, are in the process of being developed.
References
Aronson, 1986
46, 1981
patients with borderline personality disorder. Am J Psychiatry 158
(10 suppl):1–52, 2001
Arntz A, Klokman J, Sieswerda S: An experimental test of the schema
mode model of borderline personality disorder. J Behav Ther Exp
Psychiatry 36:226–239, 2005
Beck AT, Freeman AM: Cognitive Therapy of Personality Disorders. New
York, Guilford, 1990
Beck AT, Freeman A, Davis DD, et al: Cognitive Therapy of Personality
Disorders. New York, Guilford, 2004
Bienenfeld D: Cognitive therapy of patients with personality disorders.
Psychiatr Ann 2:133–139, 2007
Bohus M, Haaf B, Stiglmayr C, et al: Evaluation of inpatient DBT for bor-

derline personality disorder: a prospective study. Behav Res Ther
38:875–887, 2000
Chapman AL: Dialectical behavior therapy: current indications and
unique elements. Psychiatry 3(9):62–68, 2006
try 164:922–928, 2007
Coker S, Vize C, Wade T, et al: Patients with bulimia nervosa who fail to
engage in cognitive behavior therapy. Int J Eat Disord 13:35–40, 1993
Evans K, Tyrer P, Catalan J, et al: Manual-assisted cognitive-behaviour
therapy (MACT): a randomized controlled trial of a brief interven-
tion with bibliotherapy in the treatment of recurrent deliberate self-
harm. Psychol Med 29:19–25, 1999
Feldman RB, Guttman HA: Families of borderline patients: literal-minded
parents, borderline parents, and parental protectiveness. Am J Psychi-
atry 141:1392–1396, 1984
Friedman S, Chernen L: Discriminating the panic disorder patient from
patients with borderline personality disorder. J Anxiety Disord 8:49–
61, 1994
try 6:649–658, 2006
Goodpastor WA, Pitts WM, Snyder S, et al: A social learning approach to
group psychotherapy for hospitalized DSM–III borderline patients.
J Psychiatr Treat Eval 5:331–335, 1983
Gunderson JG, Lyoo IK: Family problems and relationships for adults
with borderline personality disorder. Harv Rev Psychiatry 4:272–278,
1997
Gunderson JG, Kerr J, Englund DW: The families of borderlines: a com-
parative study. Arch Gen Psychiatry 37:27–33, 1980
Gunderson JG, Bender D, Sanislow C, et al: Plausibility and possible de-
terminants of sudden “remissions” in borderline patients. Psychiatry
66:111–119, 2003
Hawkins KA, Sinha R: Can line clinicians master the conceptual complexi-
ties of dialectical behavior therapy? An evaluation of a State Department
of Mental Health training program. J Psychiatr Res 32:379–384, 1998
Hoffman P: Family intervention: DBT family skills training. Presented at
the Sixth International Congress on Disorders of Personality,
Geneva, Switzerland, September 1999
Hoffman P, Hooley JM: Expressed emotion and the treatment of BPD. In
Session 4(3):39–54, 1998
Kehrer C, Linehan MM: Interpersonal and emotional problem solving

skills and parasuicide among women with borderline personality dis-
order. J Personal Disord 10:153–163, 1996
line disorders, in Borderline Personality Disorders: The Concept, the
Syndrome, the Patient. Edited by Hartocollis P. New York, Interna-
tional Universities Press, 1977, pp 365–384
Koerner K, Dimeff LA: Further data—dialectical behavior therapy. Clin-
ical Psychology Science and Practice 7(1):104–112, 2000
Koons CR, Robins CJ, Tweed JL, et al: Efficacy of DBT in women veterans
with BPD. Behav Ther 32:371–390, 2001
Layden MA, Newman CF, Freeman A, et al: Cognitive Therapy of Border-
line Personality Disorder. Boston, MA, Allyn & Bacon, 1993
Linehan MM: Dialectical behavior therapy for borderline personality dis-
order: theory and method. Bull Menninger Clin 51:261–276, 1987a
Linehan MM: Dialectical behavioral therapy: a cognitive behavioral ap-
proach to parasuicide. J Personal Disord 1:328–333, 1987b
Linehan MM: Cognitive-Behavioral Therapy of Borderline Personality
Disorder. New York, Guilford, 1993a
Linehan MM: Skills Training Manual for Treating Borderline Personality
Disorder. New York, Guilford, 1993b
Linehan MM: Theory and treatment development and evaluation: reflec-
tions on Benjamin’s “models” for treatment. J Personal Disord 11:
325–335, 1997
Linehan MM, Heard HL: Impact of treatment accessibility on clinical
course of parasuicidal patients (letter). Arch Gen Psychiatry 50:157–
158, 1993
Linehan MM, Armstrong HE, Suarez A, et al: Cognitive-behavioral treat-
ment of chronically parasuicidal borderline patients. Arch Gen Psy-
chiatry 48:1060–1064, 1991
Linehan MM, Head HL, Armstrong HE: Naturalistic follow-up of a behav-
ioral treatment for chronically parasuicidal borderline patients. Arch
Linehan MM, Dimeff LA, Reynolds SK, et al: Dialectical behavior therapy
versus comprehensive validation therapy plus 12-step for the treat-
ment of opioid dependent women meeting criteria for borderline
personality disorder. Drug Alcohol Depend 67:13–26, 2002
Linehan MM, Comtois KA, Murray AM, et al: Two-year randomized con-
trolled trial and follow-up of dialectical behavior therapy vs therapy
by experts for suicidal behaviors and borderline personality disorder.
Lynch TR, Morse JQ, Mendelson T, et al: Dialectical behavior therapy for
depressed older adults: a randomized pilot study. Am J Geriatr Psy-
chiatry 11:33–45, 2003
McMain S: Dialectical behavior therapy for individuals with borderline

personality disorder and substance abuse: a randomized controlled
pilot study. Paper presented at the annual meeting of the Association
for the Advancement of Behavior Therapy, New Orleans, LA, Novem-
ber 2004
Miller AL, Rathaus JH, Linehan M, et al: DBT for suicidal adolescents.
Journal of Practical Psychiatry and Mental Health 3:78–86, 1997
Oldham JM. Guideline Watch: Practice Guideline for the Treatment of
Patients With Borderline Personality Disorder. Washington, DC,
Paris J: Borderline Personality Disorder: A Multidimensional Approach.
Robbins M: The adaptive significance of destructiveness in primitive per-
sonalities. J Am Psychoanal Assoc 36:627–652, 1988
Ryle A: The contribution of cognitive analytic therapy to the treatment of
borderline personality disorder. J Personal Disord 18:3–35, 2004
Safer DL, Telch CF, Agras WS: Dialectical behavior therapy for bulimia
nervosa. Am J Psychiatry 158:632–634, 2001
Schmidt U, Davidson K: Life After Self-Harm. Hove, East Sussex, En-
gland, Brunner-Routledge, 2004
Scogin FR: Introduction: the status of self-administered treatments. J Clin
Psychol 59:247–249, 2003
Shapiro ER: On curiosity: intrapsychic and interpersonal boundary for-
mation in family life. International Journal of Family Psychiatry 3:69–
89, 1982
Shearin EN, Linehan MM: Patient therapist ratings and relationship to
progress in dialectical behavior therapy for borderline personality
disorder. Behav Ther 23:730–741, 1992
Simpson EB, Pistorello J, Begin A, et al: Use of dialectical behavior ther-
apy in a partial hospital program for women with borderline person-
ality disorder. Psychiatr Serv 49:669–673, 1998
Skodol AE, Shea MT, McGlashan TH, et al: The Collaborative Longitudi-
nal Personality Disorders Study (CLPS): overview and implications.
J Personal Disord 19:487–504, 2005
Springer T, Silk KR: A review of inpatient group therapy for borderline
personality disorder. Harv Rev Psychiatry 3:268–378, 1996
1992
Stone M: The Borderline Syndromes. New York, McGraw-Hill, 1980
Swenson C: Kernberg and Linehan: two approaches to the borderline pa-
tient. J Personal Disord 3:26–35, 1989
Swenson C: How can we account for DBT’s popularity? Commentary re:

article by Scheel. Clinical Psychology Science and Practice 7(1):87–
91, 2000
Telch CF, Agras WS, Linehan MM: Dialectical behavior therapy for binge
eating disorder. J Consult Clin Psychol 69:1061–1065, 2001
Turner RM: Naturalistic evaluation of dialectical behavior therapy-
oriented treatment for BPD. Cogn Behav Pract 7:413–419, 2000
Tyrer P, Thompson S, Schmidt U, et al: Randomized controlled trial of
brief cognitive behaviour therapy versus treatment as usual in recur-
rent deliberate self-harm: the POPMACT study. Psychol Med 33:969–
976, 2003
Verheul R, Van Den Bosch LM, Koeter MW: Dialectical behaviour ther-
apy for women with borderline personality disorder: 12-month, ran-
domized clinical trail in The Netherlands. Br J Psychiatry 183:135–
140, 2003
Weinberg I, Gunderson JG, Hennen J, et al: Manual assisted cognitive
treatment for deliberate self-harm in borderline personality disorder
patients. J Personal Disord 20:482–492, 2006
Westen D: Cognitive-behavioral interventions in the psychoanalytic psy-
chotherapy of borderline personality disorders. Clin Psychol Rev 11:
211–230, 1991
Young DW, Gunderson JG: Family images of borderline adolescents. Psy-
chiatry 58:164–172, 1995
Young JE: Cognitive Therapy for Personality Disorders. Sarasota, FL, Pro-
fessional Resource Exchange, 1990
Young JE: Cognitive Therapy for Personality Disorders: A Schema-Focused
Approach, Revised Edition. Sarasota, FL, Professional Resources Press,
1994
Young JE, Klosko J, Weishaar ME: Schema Therapy: A Practitioner’s Guide.
New York, Guilford, 2003
Zanarini MC, Frankenburg FR, Hennen J, et al: The longitudinal course
of borderline psychopathology: 6-year prospective follow-up of the
phenomenology of borderline personality disorder. Am J Psychiatry
160:274–283, 2003
Zanarini MC, Frankenburg FR, Hennen J, et al: Psychosocial functioning
of borderline patients and Axis II comparison subjects followed pro-
spectively for six years. J Personal Disord 19:19–29, 2005
Chapter 12
PSYCHODYNAMIC
PSYCHOTHERAPIES
THE IDENTIFICATION OF borderline personality psychopathology arose

out of observations from within psychoanalytically oriented treatment set-
tings (Chapter 1), and the term borderline became widely used only after—
and I think because—Kernberg (1968) and Masterson (1971) promoted
hopes that skilled, intensive, long-term psychodynamic psychotherapies
could bring about curative, basic structural changes. In the 30 years follow-
ing Kernberg’s and Masterson’s publications, the literature was flooded
with articles and books on the problems, processes, and potential for such
therapies. Psychotherapies remain central to most treatment plans. As
noted, during the getting-started first phase (Chapter 10), the therapist’s
early tasks usually involve case management activities. In this chapter, I dis-
cuss the processes within subsequent phases of long-term psychodynamic
therapies that allow internal psychological changes of a more basic, struc-
tural, and enduring nature to occur. These phases are as follows:
• Phase 2: A relational alliance

• Phase 3: Positive dependency
• Phase 4: Secure attachment, the working alliance, and consolidation
of self
Mirroring the changes that have occurred within psychoanalytic con-

ceptions about processes of change (Stern et al. 1998), the psychoana-
lytic literature relevant to borderline patients has moved from ego-
psychology or object relations theories and techniques into relational per-
spectives that now attach more significance to processes of engagement
and attachment. Although the idea that processes within the therapy re-
279
lationship are themselves therapeutic has been central to most psychody-

namic therapies, as noted in Chapter 11, schema-focused therapy (SFT)
also includes this conception.
With respect to the framework of the five therapeutic functions de-
scribed in Chapter 3 of this book, those that are most specific to psycho-
analytic therapies are involvement and actualization. Involvement refers to
the process by which a level of closeness, trust, and intimacy develops that
exceeds what patients have previously experienced and that gives pa-
tients new abilities to form such relationships with others. Actualization re-
fers to the process by which patients develop recognition and acceptance
of themselves as unique. It develops out of increased awareness and un-
derstanding of oneself. This function has special significance for border-
line patients, whose recognition of feelings or motives is impaired and
whose sense of self is often confused or distorted.
In this chapter, I review some aspects of the literature on psychoanalytic
psychotherapies and the increasingly sophisticated research on their effi-
cacy. Two basic processes of change are described: those related to rela-
tional experiences and those related to learning. These psychotherapeutic
processes are sometimes referred to respectively as corrective attachment and
as insight (or as changes in cognitive schema). I then identify three phases
of change (phases 2–4) that allow therapists and patients to recognize
whether therapies are progressing or have reached an impasse. Common
types of impasse in each phase are described.
Pre-Empirical Developments
In the 1970s and early 1980s, many conferences were held featuring ex-
pert psychoanalytic therapists who detailed competing theories and tech-
niques that they believed were most effective. During this period, the
enthusiasm for the value of psychoanalytic therapies with borderline pa-
tients was at its peak. In a thoughtful commentary, Aronson (1985)
pointed out that the authors responsible for this peak were all narrowly
analytic and rarely focused on issues of diagnosis, attrition, treatment fail-
ure, or limitations of their model.
No debate captured more attention than those featuring Kernberg
and Kohut/Adler (Sidebar 12–1 and Table 12–1).
Sidebar 12–1: Kernberg Versus Kohut/Adler:

The Debate of the 1970s
Kernberg’s (1967, 1968) seminal papers on borderline pa-
tients bridged object relations theory (the concepts of in-
trojection, self-representation, projective identification) and
Psychodynamic Psychotherapies ❘ 281
instinct theory (vicissitudes of basic aggressive drives). At

about the same time, Heinz Kohut’s (1971) seminal contribu-
tions to self psychology (with its concepts of mirroring, self-
objects, and transmuting internalizations) became tied to
pathological narcissism and became a counterpoint to
Kernberg’s theories about development and therapeutic
technique. These theoretical contributions became part of
a larger intellectual debate. Kernberg’s construct of border-
line personality organization subsumed what Kohut called
pathological narcissism, and Kohut’s concept of patholog-
ical narcissism offered different ways to explain borderline
phenomena such as rage, feelings of entitlement, and intol-
erance of aloneness.
Behind these debates was a fundamental difference:
whether the inappropriate or excessive expressions of anger
shown by borderline or narcissistic patients were under-
standable as reactive to insults to a fragile self (i.e., narcissis-
tic injury) or as inadequately modulated expressions of an
aggressive drive.
The theoretical and clinical controversy provided ma-
terials for a burgeoning number of books on psychoanalytic
therapies. Among these, the books by Brandschaft and
Stolorow (1987), Chessick (1977), and Volkan (1987) aligned
themselves with self psychology, and those by Goldstein
(1985), Grotstein (1988), Masterson (1972, 1976), Rinsley
(1982), and Searles (1986) were more aligned with object re-
lations theory. In the more recent literature, Kernberg’s the-
ory increasingly stands alone. In contrast to Kernberg’s
theory and to TFP (Clarkin et al. 2006), both mentalization-
based therapy (Bateman and Fonagy 2004) and SFT (Young
et al. 2003) see borderline anger as reactive and defensive.
Against the backdrop of the debate between the Kernbergian and Ko-
hutian/Adlerian models, several experts wrote accounts about psychody-
namic psychotherapies that were more pragmatic and eschewed either
theoretical pole (e.g., Benjamin 1993; Gabbard and Wilkinson 1994; Gun-
derson 1984, 1996; Kroll 1988; McGlashan 1993; Paris 1998; Stone 1990,
1993). My own experience indicated that there were often distinctions in
the use of each model: the supportive techniques advocated by Adler are
needed early in therapy, are crucial for therapies done with support from
other modalities, and are the techniques essential for making borderline
patients feel cared for and become attached (in a relational alliance). It
also seemed apparent that the interpretations advocated by Kernberg can
be essential for managing early negative transferences, that they become
increasingly valuable over time, and that, as noted by Kernberg, they are
crucial to helping borderline patients recognize and own unacknowl-
TABLE 12–1. The debate

KERNBERG KOHUT/ADLER
(Transference analysis) (Corrective relationships)

Confront Validate
Limit Avail yourself
Interpret Empathize
Aggression intrinsic Aggression reactive
Conflict model Deficit model
edged and misdirected aggression. But—as noted by Kernberg’s many

critics, and a point emphasized by Gabbard et al. (1994)—the acceptance
of such interpretations depends on the presence (even though transitory)
of a working alliance (to be discussed later in this chapter in “Phase 4: Se-
cure Attachment, the Working Alliance, and Consolidation of Self”).
Outcome Studies: Nonrandomized Trials

Since 1968, at least 56 books have been written by psychoanalysts about
treatment of borderline personality disorder (BPD) by psychoanalytic psy-
chotherapies (data from Library of Congress database search). (Of course,
a vastly larger number of additional journal articles and book chapters are
devoted to the subject.) In most instances, these books reported on inten-
sive (three sessions a week or more) therapies conducted with a focus on
insights gained by examination of the relationship. In some instances, psy-
choanalysis proper (i.e., the parallel, lying-down arrangement) has been
described (Abend et al. 1983; Volkan 1987). Four books described a total
of 26 detailed case studies (Abend et al. 1983; Chessick 1977; Searles 1986;
Volkan 1987). Of these 26 case studies, 18 patients did not meet criteria for
BPD (Gunderson and Gabbard 1999), although they did have the primi-
tive defenses and identity issues consistent with the broader concept of
borderline personality organization (see Chapter 1). By DSM standards,
the 18 non-BPD cases would meet criteria for other types of personality dis-
order, primarily narcissistic (n=6) or schizoid (n=4).
Subsequent to these case series, a number of naturalistic prospective
follow-along studies have explored what benefits can be expected from
long-term psychoanalytic therapies (Table 12–2).
The first and most widely influential of these studies was the Menninger
Psychotherapy Research Project (MPRP; Kernberg 1972; Kernberg et al.
1972; Wallerstein 1986). Most of the 42 patients in this naturalistic prospec-
tive study had severe personality disorders. In the study design, patients
were assigned to receive psychoanalytic, expressive (i.e., investigative, in-
sight oriented, emotion generating), or supportive (i.e., directive, defense
reinforcing, emotion inhibiting) forms of psychotherapy. Hospitalization
was a concurrent context for most patients in the course of the treatments.
Horwitz (1982) examined the outcomes of a subgroup of 16 patients
of the original 42 who might qualify for a diagnosis of borderline person-
ality (see Table 12–2). Of these 16 patients, 5 were considered successes,
5 were considered unchanged, and 6 were thought to have become
worse. The type of treatment the subgroup sample received was divided
equally between psychoanalysis and psychotherapy. All 6 borderline pa-
tients who had received psychoanalysis alone got worse. Two patients who
received expressive therapy did well, but they were considered to have
higher baseline levels of ego strength. All 5 of the successful outcomes oc-
curred in patients who had received supportive-expressive psychother-
apy. Nevertheless, 4 of these 5 still had evidence of primary-process
thinking on psychological tests conducted at the follow-up evaluation.
This study remains the most significant evidence that psychoanalysis per
se is contraindicated for borderline patients.
Still, borderline patients who have attained internal controls and sta-
ble role functioning may want to deepen their therapy and, in effect, ex-
pand their object relatedness by learning to examine the increased
frustrations and projections that are invited by going on the couch. If
they have achieved these capabilities, they are by then, arguably, no
longer borderline. In my limited experience, such patients can then go
on to achieve considerable further growth, but it becomes impossible to
sort out the effects of aging (i.e., BPD’s natural course).
A second set of studies from McLean Hospital in the 1980s was
prompted by the question whether it would be possible to evaluate the ef-
ficacy of psychoanalytic therapy through controlled outcome research.
Drawing on lessons learned from a previous experience in conducting
such a study with schizophrenia subjects (Gunderson et al. 1984; Stanton
et al. 1984), we were determined as a first step to document that major
structural benefits occur from psychoanalytic psychotherapy and to
establish reasonable estimates of how often such benefits last and how
long such benefits take to occur. As noted in Chapter 10 of this book,
Waldinger and I found that even expert, published psychoanalytic ther-
apists frequently had high rate of dropouts and rarely (10%) judged their
therapies with borderline patients to have ended successfully (Waldinger
and Gunderson 1984).
The second approach took place at McLean, where many therapists,
like those in the Menninger study, had been practicing intensive long-term
284
TABLE 12–2. Summary of outcome studies
STUDY SAMPLE TYPE OF PSYCHOTHERAPY THERAPISTS
❘
Menninger Psychotherapy N =16 who met criteria for BPO Assigned to receive psychoanalysis (n=6) M.D. and Ph.D.

Research Project (Horwitz and were referred for long- or “expressive” (n=5) or “supportive” with extensive
1982; Kernberg 1972; term treatment with (n=5) psychotherapy on basis of clinical experience
Kernberg et al. 1972; intermittent use judgment; frequency=≥3/wk for
Wallerstein 1986) of hospitalization expressive and 1/wk for supportive
McLean Psychotherapy N =60 hospitalized patients “Psychodynamic” without M.D.’s and Ph.D.’s
Engagement Project meeting DIB/DSM criteria for standardization of technique, theory, of varying
(Gunderson et al. 1989, BPD; all were starting a new or intensity; most were seen once experience,
1997; Najavits and psychotherapy weekly including trainees
Gunderson 1995).
Northwestern University N =23 outpatient “borderline- Unstandardized, “generic”; most Unstated, many
Department of Psychology psychotic” subjects from patients seen ≤ once weekly by
(Howard et al. 1986) various study samples trainees
New South Wales N =30 outpatients meeting Standardized, “self psychological” with M.D. and Ph.D.
(Stevenson and Meares DSM-IV criteria for BPD intensive supervision; frequency was trainees
1992, 1999; Meares et al. twice weekly
1999).
Note. BPD= borderline personality disorder; BPO=borderline personality organization; DIB = Diagnostic Interview for Borderline Patients;
DSM-IV=Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (American Psychiatric Association 1994).
therapies for many years. When McLean staff were invited to identify cases
that they thought had gone on to curative changes, surprisingly few could
do so. Five cases were subsequently detailed in case reports (Waldinger
and Gunderson 1989). This study showed that borderline patients could
undergo curative changes in long-term therapies (4–7 years). Given this
duration of treatment, and in the light of current knowledge about the un-
expectedly good course of BPD, it is not possible to safely infer that the cur-
ative changes were due to the therapies. In any event, the more instructive
finding was to discover how rarely such curative changes occur.
The third part of this investigation was a naturalistic prospective study
with unselected therapists of variable experience. This study confirmed
that dropouts were common (Gunderson et al. 1989) and that there was
an overall variability in outcomes (Najavits and Gunderson 1995; Sabo et
al. 1995). As a result of these studies, we concluded that even with senior,
experienced therapists, major successes were unusual and took many
years and that, without being able to identify what qualities of therapist
and borderline patient made effectiveness possible, controlled outcome
research on psychoanalytic therapy with borderline patients was not yet
feasible.
A third outcome study was reported by Howard et al. (1986) as part of
a meta-analysis in which an overall relation between number of sessions
and successful outcome was documented. A smaller sample at their own
clinic, 23 patients grouped as “borderline-psychotic,” required a signifi-
cantly greater number of sessions to achieve improvement than did pa-
tients who were depressed or anxious. In what was usually once-weekly
therapies, about half of the “borderline-psychotic” sample who remained
in therapy “had improved” by 6 months, about 75% by 1 year, and nearly
90% by 2 years. Although this study had a nonstandardized threshold for
improvement and did not identify what percentage of patients discontin-
ued therapy because they were not improving, the seemingly impressive
rate of improvement now seems less surprising in view of the natural
course of this disorder.
The fourth outcome study came from New South Wales, Australia.
Stevenson and Meares (1992, 1999) conducted a naturalistic prospective
study of the effectiveness of psychodynamic psychotherapy. The thera-
pies were conducted twice weekly by young trainees who received exten-
sive supervision. The 30 borderline patients received 12 months of
therapy, during which only 16% dropped out after testifying to a very suc-
cessful engagement process (Meares et al. 1999). At the 1-year follow-up,
these patients had a significant decline in hospital use; in episodes of self-
harm (from 3.77 per year to 0.83 per year); and in the mean number of
DSM-IV (American Psychiatric Association 1994) BPD criteria, with 9 pa-
tients (30%) below the threshold for BPD diagnosis. Again, these results
might appear remarkably positive, and they were significantly better than
a waiting-list control group (Meares et al. 1999), but the results appear far
less impressive given current knowledge about the natural course of BPD
(Skodol et al. 2005; Zanarini et al. 2005).
Transference-Focused Psychotherapy
Psychodynamic or psychoanalytic therapies with borderline patients have
been the target of growing criticism and skepticism since dialectical behav-
ior therapy (DBT) emerged with empirical validation in the early 1990s
(Sidebar 12–2). Nowhere was this more evident than for Kernberg’s model
of transference-focused psychotherapy (TFP)—a model introduced by a
psychoanalytic leader and already controversial within that community.
The heroic efforts required to manualize TFP deserve note. The effort be-
gan with the still theory-laden and clinically rich draft of a manual pub-
lished by Kernberg and colleagues (1989). The draft then underwent
considerable further development, giving way over the next 15 years to a se-
ries of increasingly pragmatic, specific, and operationalized versions (Clar-
kin et al. 1999, 2006; Koenigsberg et al. 2000; Yeomans et al. 2002). This
effort was led by Clarkin, a distinguished psychotherapy researcher. Perhaps
pushed by the example set by Linehan for DBT (see Table 11–1), TFP also
has now identified a hierarchy of treatment goals. As shown in Table 12–3,
like DBT, TFP’s hierarchy begins by addressing suicidal threats—also in-
cluding violent behaviors—and then moves to other treatment-interfering
behaviors. As a result of this manualization of TFP, this form of psychoana-
lytic psychotherapy became uniquely capable of having its efficacy tested.
Sidebar 12–2: Kernberg Versus Linehan:

The Debate of the 1990s
The emergence in 1993 of DBT as the first empirically vali-
dated treatment for BPD created a clinical and scientific
precedent that threatened the favored status previously
accorded psychoanalytic therapies. Linehan quickly
achieved national recognition and the appreciation of
academic leaders and behaviorally oriented clinicians.
Kernberg had by this time risen to the top of the psychoan-
alytic guild and retained an uncompromising conviction
about his treatment’s superiority on the basis of its theoret-
ical depth and comprehensiveness.
In clinical interviews and at numerous well-attended
conferences, the two leaders presented startling contrasts.
Kernberg presented as a small, balding, dark-suited man
whose old-world formality and manners were combined
with his intellectual intensity and seriousness. His talks were

often from transcripts, he loved to describe his theory, and
occasionally he offered vivid vignettes highlighting the im-
portance of primitive defenses, the vicissitudes of aggres-
sion, and the value of an interpretive stance. In contrast,
Linehan presented as a indomitable woman whose confi-
dence and strength invariably dominated whatever occa-
sion she attended. She used expressive gestures and
PowerPoint slides while speaking. Although Linehan also
was committed to her theory centering on emotional dys-
regulation, invalidating environment, and the value of an
in-the-moment emotional availability, she primarily loved
to dwell on her research.
Both Linehan and Kernberg were acutely aware that
they competed for leadership. Both were reluctant to iden-
tify either the overlaps or the merits of each other’s ap-
proach. Kernberg dismissed DBT as superficial. Linehan
dismissed TFP with “where’s the data?” Linehan unflinchingly
challenged claims for treatment efficacy that were without
supportive data. In the process, she challenged the psycho-
analytic trade guild, in which membership required many
years of extensive and private training and whose thera-
peutic credibility was based on compelling theories and
handed-down wisdom by senior practitioners. Kernberg
saw DBT as superficial in its theory, its behavioral focus, and
its capacity to effect “structured” changes. Still, although
never explicitly acknowledged, the inherent weakness of
the claims of efficacy that were based solely on Kernberg’s
personal experience and the fidelity of his acolytes was rec-
ognized. Linehan’s example and model and chiding all in-
centivized TFP advocates to puts its efficacy to the test.
As noted in Chapter 11, a major project in Holland has compared the

benefits from 3 years of Kernberg’s TFP (Clarkin et al. 1999) with the
benefits of 3 years of SFT. As with prior research in which different types
of similarly intensive psychotherapy are compared, the outcomes seemed
likely to be similar. Still, as noted in Chapter 11, the borderline patients in
both cells significantly improved, but those receiving SFT did better than
those receiving TFP (Giesen-Bloo et al. 2006). This interpretation is com-
plicated by the fact that TFP patients began with more suicidality and self-
harm and by the failure of 51% of the TFP therapists to be adherent—
doubtless undermining effectiveness and contributing to the higher rate
of TFP dropouts. One certain result is that Dutch borderline patients will
benefit from having a cadre of well-trained therapists as a nucleus who
can teach others and thereby eventually elevate the overall standards of
psychotherapeutic care.
TABLE 12–3. Hierarchy of targets for transference-focused

psychotherapy
1. Suicide or homicide
2. Overt threats to treatment continuity
Dishonesty or deliberate withholding
Contract breaches
In-session acting-out
Between-session acting-out
Source. Adapted from Clarkin et al. 1999.
TFP Versus DBT Versus Supportive Psychotherapy

A long-awaited, methodologically rigorous, and clinically more meaning-
ful test of TFP’s efficacy has now been completed (Clarkin et al. 2007).
This study randomly assigned 90 BPD patients to receive twice-weekly
TFP, standard DBT (1.5 hours of group and 1 hour of individual), or
once-weekly supportive therapy for 1 year. Therapists in all three cells
were trained to competence and then sustained adherence for their
model. All three treatment cells proved more-or-less equally effective in
major sectors of outcome: depression, anxiety, social relations, and Glo-
bal Assessment of Functioning. The treatments differed in other re-
spects. TFP and DBT patients had more improved suicidality than did
supportive therapy patients. TFP and supportive therapy patients im-
proved more than did DBT patients on anger. Surprisingly, outcome in
the areas of BPD psychopathology and health care use has not been re-
ported. No differences were found in sustaining these gains over a 1-year
follow-up.
This was a remarkable study with complicated and significant impli-
cations:
1. The reported outcomes were comparable to those achieved by DBT,

thus justifying TFP’s long-standing claim of efficacy for treating BPD.
This result, as noted in the review of earlier studies, was by no means
predictable.
2. The results of this study underscore the question of what are the key
processes of change. The fact that a supportive psychotherapy did as
well as either DBT or TFP suggests that support may be the necessary
and sufficient vehicle for change. It is a reminder that in the original
MPRP, supportive techniques (and presumably attitudes) were essen-
tial for successful outcomes. The supportive therapy deployed in this
study (Appelbaum 2005) unfortunately is not representative of usual

supportive therapy. It was actually a very dynamically sophisticated
“supportive-expressive” therapy that differed from TFP primarily by of-
fering more reassurance, self-disclosure, and occasionally directives.
Also, the access to as-needed sessions for patients in this cell meant that
the final number of sessions was similar to that for TFP and DBT.
3. This was the first study in which a psychoanalytic psychotherapy (with
its emphasis on interpretation and transference) was operationalized
to a point that competence and adherence were measurable. This
flies in the face of psychoanalytic protestations that their therapy in-
volves too much subtlety and creativity to be straitjacketed by a man-
ual. Although the decades of effort required to manualize TFP testify
to why it may have seemed impossible, the success sets a new standard
by which psychoanalytic therapies for other patient types should be
measured.
With the overall similarity of the outcomes from all three treatments
in mind, one distinction is being assigned much significance: patients
who received TFP scored higher on measures of reflectiveness and coher-
ence. Both variables are scored from the Adult Attachment Interview,
and reflectiveness is thought to be a proxy for Fonagy’s concept of men-
talization (see Chapter 1, Sidebar 1–4). This finding could be important
insofar as it provides some evidence of specificity for the claim that psy-
choanalytic therapies offer deeper changes in the ways people think. It is
thought that improved mentalization (more accurate at complex assess-
ments of self and others) is the mechanism for change in symptoms and
behaviors.
Overview of Change Processes

Randomized controlled trials for psychotherapies primarily examine the
early stages of longer-term treatments. The studies of DBT and TFP focus
on the first year. Even in the 3-year study of SFT (vs. TFP; Chapter 11),
most of the patients continued their therapies. In this section, I describe
the processes of change that might be expected in the course of long-
term therapies.
Table 12–4 offers a schematic survey of the processes of change that
can be expected over the course of long-term successful therapies with bor-
derline patients. It mirrors changes described in Chapter 3 (Figure 3–2)
but elaborates on those changes by adding processes within the therapy.
The timeline may vary considerably, but the sequence in which
changes occur is quite predictable. Above all, the table highlights that
TABLE 12–4. Indices of change in long-term psychotherapies for patients with borderline personality disorder
290
(BPD)
PHASE
❘
1: GETTING STARTED 2: RELATIONAL ALLIANCE 3: POSITIVE DEPENDENCY 4: SECURE ATTACHMENT

(0–3 MONTHS) (1 MONTH–1 YEAR) (1 YEAR–2 OR 3 YEARS) (2 OR 3 YEARS–?) RESULT
Change target ↓symptoms (moods) ↓ self-destructiveness ↓maladaptive ↓splitting, leading to ↓emptiness
↓ impulsivity interpersonal problems owning anger ↑friends
↓projection ↓Avoidance
Therapeutic Contractual alliance Relational alliance Relational alliance Working alliance Secure
relationship Agreed-upon goals Therapist valued Therapy valued Separation anxiety attachment
and roles Dependent/anxious Dependent/positive
Counterdependent
Major issues Action, symptoms, Affect recognition and Misattribution, Negative transference Internal locus
fearfulness tolerance assertiveness Reentering competition of control
Anger and denial of Accepting neediness Fear of aggression Developmental issues
anger Anger projected Anger projected Trauma, self-image
Projection
Therapist Interactive Clarifies maladaptive Identifies conflicts and Interprets conflicts N/A
activities Responsive responses to feelings misattributions and transference
Educates and clarifies (e.g., frustration) Supports functional Confronts avoidance
Validates and empathizes capabilities
Develops formulation Connects present to past
Outcome Patient likes and is Capable of low-demand Capable of low-demand Capable of competi- Patient does
engaged by therapist social role relationships tion, friendships not have BPD
when change ceases, it is wise for therapists to seek consultation rather

than assume that further change is not possible.
Several studies have examined the alliance between borderline patients
and their therapists. As noted, this topic has been central to many psycho-
analytic discussions (see Sidebar 3–2). In the McLean prospective repeated-
measures study of 35 BPD patients who were beginning individual psycho-
therapies, a rather steady improvement in the relational type of alliance was
observed (Figure 12–1) (Gunderson et al. 1997). Although in the Men-
ninger Treatment Intervention Project the long-term course of the working
alliance resulted in overall improvements in one patient (Gabbard et al.
1988, 1994), the level of alliance in three other patients was still found to
fluctuate dramatically, even after they were in therapy for several years (Hor-
witz et al. 1996). Of note in both studies is that the initial alliance scores were
higher than expected (Gunderson et al. 1997; Horwitz et al. 1996).
Phase 2: A Relational Alliance

After a patient has become engaged in therapy (identified as phase 1 in
Table 12–4 and in Chapter 10), phase 2 begins. In this phase, the primary
FIGURE 12–1. Change in therapist and patient ratings of their al-

liance during years of therapy.
BPD= borderline personality disorder.
Source. Adapted from Gunderson et al. 1997.
learning process involves behavioral control, and the primary relational

process involves advancing from the contractual form of alliance, as dis-
cussed in Chapter 10, to establishing the second level of alliance, a rela-
tional alliance (see Table 12–4). With the borderline patient, a relational
alliance involves likability, common goals, reliability, and hope for a bet-
ter future (Table 12–5). In effect, the therapist becomes a selfobject
(Stolorow 1995) or a transitional object (Giovacchini 1984; Modell 1963)
with potential to offer corrective experiences.
The primary therapeutic techniques that make a relational alliance pos-
sible involve showing interest, conveying feasible expectations, possessing
resilience in the face of opposition, and, above all, as emphasized by Adler
(1986) and reported by Stevenson and Meares (1992), deploying empathy
and validation. Empathy involves acknowledgment of patients’ dilemmas
(e.g., “What a difficult situation,” or “I can see why you were undecided”)
and especially of their feeling states (e.g., “You must have been scared,” or
“You seem angry about that”). Empathy is often complicated by patients’
fears that their feelings either are evidence of their “badness” or will be un-
acceptable to others. Fonagy and Target (1996) emphasized the corrective
power of empathetic interventions for those borderline patients whose feel-
ings as children have been ignored, mislabeled, or rejected. Patients learn
to observe themselves by being observed; to paraphrase Winnicott (1967),
they discover themselves in their mother’s (or therapist’s) eyes. They also
learn a useful new way to label and accept part of their experience.
Although patients’ initial reaction to feedback about themselves is likely
to be ambivalent, usually suspicious, and sometimes hostile, it helps to start
by making observations at the surface. Uninvited observations by the clini-
cian indicate that the clinician is attending to the task of helping patients
learn about themselves. I often comment about facial expressions: “You look
worried” or “You seemed sad when you talked about….” Actively identifying
a patient’s apparent feelings is most important when the patient looks either
fearful or angry—feelings that may be hard to recognize or talk about and ei-
ther of which, if not noted, can result in flight. Affirmation or disclosure by
therapists about their own feelings, done with discretion, also helps develop
TABLE 12–5. Components of a relational alliance

Likability: responsive, real
Commonality: goals, interests, values
Reliability: conscientiousness, predictability
Hopefulness: change is possible; goals are attainable
better mentalization capacities (Fonagy 1991, 1995; Fonagy et al. 1995): learn-
ing to represent feelings and affects in one’s mind without action. It also
presages a very important, recurrent thematic process in long-term therapies
whereby borderline patients connect behaviors to events, to feelings, and to
their thoughts.
Validation involves actively reinforcing the reality of borderline pa-
tients’ perceptions and identifying the adaptive functions served by their
defenses and behaviors. Of particular delicacy is the balance between lis-
tening sympathetically to disclosures of past mistreatment and, while val-
idating the experience of unfairness, not assuming the validity of the
realities as described (Gunderson and Chu 1994). This balance can be
difficult, either because the natural impulse is to convey support or be-
cause the borderline patient so clearly wants you to. It is usually sufficient
to convey that the patient’s life sounds as if it were awful and that you can
understand why, under such circumstances, he or she behaves as he or
she characteristically does. Although these processes of empathic iden-
tification of feelings, validating their significance, and tracing their
connections to events and actions have been given a developmental per-
spective by Fonagy’s observations, the same interventions also arose as
central processes within the cognitive-behavioral therapies of Linehan
(1993, 1997) and Young (1990; Young et al. 2003) on the basis of only
their clinical experiences.
Being liked occurs early (see Chapter 10) and is of value, but prima-
rily because it helps create the engagement needed for a relational alli-
ance. The valuation of a therapist is most directly a result of the therapist’s
empathy and validation. These activities make the therapist what is
termed a good object. But valuing therapy derives from learning experience.
Interpretations or confrontations that bring to the patient’s attention
problems in himself or herself are activities that risk the therapist’s be-
coming a bad object. Still, by 3–6 months, the value of the tasks in therapy
should be evident in patients’ reports that they have learned new things
about themselves (Gunderson et al. 1997). Indeed, I like to underscore
the therapy’s task, understanding oneself, from the very first session by
making observations about a patient and inquiring about whether he or
she has learned anything new.
Empirical data that help identify key therapeutic processes are avail-
able from the MPRP. As noted earlier, this study distinguished between two
types of psychotherapy: expressive (i.e., investigative, insight oriented,
emotion generating) and supportive (i.e., directive, defense reinforcing,
emotion inhibiting). The original interpretation by Kernberg et al.
(1972), buttressed by Guttman’s formidable statistical techniques, was that
the data indicated that expressive techniques and processes were effective.
A reanalysis by Horwitz (1974) suggested that patients with borderline per-

sonality organization who had a strong therapeutic alliance with their
therapists did improve significantly in supportive therapy. Wallerstein’s
(1986) further reanalysis of the data determined that in actual practice al-
most all the therapies offered a less expressive—that is, less psychoanalytic
and more supportive—approach than the study design called for. More-
over, patients frequently switched between treatment modes and between
therapists. The study appears to yield quite different conclusions concern-
ing the value of expressive therapy for BPD. Kernberg has persisted in
viewing expressive therapy as superior, whereas for Wallerstein and others,
the results pointed toward the critical role of supportive elements.
Of further note is the significance of relational factors emphasized in
Stevenson and Meares’ psychodynamic therapies. Stevenson and Meares
(1992; Meares et al. 1999) defined their brand of therapy (and offered in-
tensive supervision to ensure adherence) as self psychological (Kohutian/
Adlerian; see Sidebar 12–1), with particular emphasis on empathic con-
nection. The exploratory component focused on the identification of the
inevitable triggers that disrupt that sense of connectedness. Thus, the em-
pirical evidence reinforces the conclusion that supportive, attachment-
generating interventions are critically important for successful psycho-
therapies (Chapter 10). Notably, the evidence offered from both Waller-
stein (1986) and Stevenson and Meares (1992) that supportive, empathic
interventions are critically important is consistent with the larger shifts in
modern psychoanalytic thinking, which now, starting with Langs’s “biper-
sonal field” (Langs 1976) and Gill’s “dyadic relationship” (Gill 1979), have
accepted the corrective power of relational processes (Lyons-Ruth et al.
1998; Stern et al. 1998). For borderline patients, this is a prototypical type
of transitional object relatedness that can provide a holding environment
(Modell 1976).
On reading Adler or Stevenson and Meares, there is a notable ab-
sence of concern about and focus on boundaries and behavioral control
issues in contrast with the attention given to these issues by Kernberg,
Linehan, and most others. This first year in therapy is when the border-
line patient’s testing behaviors, or boundary violations (see Chapter 4),
are most likely to occur. The primary benefits to be expected in this sec-
ond phase of the first year of therapy are behavioral: fewer impulsive, des-
perate, and self-destructive behaviors.
Although much attention has been given in the literature to the role
of confrontations or limits in facilitating behavior change, I believe that
therapist activities in the area of learning or insight can also greatly con-
tribute to attaining the primary behavioral control and safety objectives
of the first year.
The following vignette illustrates some of the processes that typify this
phase of therapy, when the relational alliance is being built alongside a
task orientation. This material illustrates my efforts to convert the mean-
ing of depressive symptoms into maladaptive defensive phenomena—
that is, convert the patient’s depression into a meaningful communica-
tion of needs and fears. I use it also to introduce how a concept of mental
deficits can link symptoms to meanings and how therapy can be transfor-
mative.
Vignette
Ms. AA was 6 months into her thrice-weekly psychotherapy. She had
started as an inpatient, moved through 3 months of partial hospital, and
now was an outpatient. She appeared looking pale and thin, walked slowly
to her chair, seemed distracted, and didn’t look at me.
Therapist: You look depressed. [a comment about a feeling]
Ms. AA: I am.
Therapist: What’s going on.... how do you understand this? [a ques-
tion]
Ms. AA: I don’t.
Therapist: I’m surprised you don’t relate it to what we talked about
last time (i.e., having started work).... [a linking comment that creates a
coherent narrative]
(Silence)
Therapist: Then, do you relate becoming depressed to starting work?
[a question]
Ms. AA: No.
Therapist: [now I question her response to therapy, to me]: Does that
mean that you think what I’ve been pointing out, interpreting, and even
predicting about your depression isn’t correct? [I’ve been saying since we
started that every time she takes a step toward more responsibility and less
patient care, it represents a big threat to her and impels her to seek more
supports.]
Ms. AA (irritably interrupting): Yes, I know (rolls eyes disdainfully),
every step forward will be followed by 10 backward. I think that’s just your
theory.
Therapist: That theory helps explain why you’re depressed: why tak-
ing a step—not a little one, by the way [here I provide validation]—like
your new job would predictably cause you to feel deprived and feel in
need of more help. Unfortunately, to my mind, by becoming dysfunc-
tional, you may evoke caring responses that you could otherwise attain
more readily than you believe.
Ms. AA’s depression can still be seen as an Axis I disorder: somatic

symptoms (sleepless, lost appetite, anergic) with morbid preconceptions.
Still, unlike earlier depressive episodes, at this time Ms. AA is responsive
to my efforts, even if she seems to disagree with the content of my re-
marks. My responses to her involve interpreting her depression as 1) a re-
gression in the face of abandonment and separation fears (including

from the therapist) engendered by the job; 2) an angry attack on the
meaning given to her depressive experiences in the therapy; and 3) a
communication of her need for added support.
I deploy a series of responses to Ms. AA that are escalating in their in-
trusiveness and in their provocativeness—that is, in being difficult for her
to ignore. First commenting, then questioning, and then asking a more
affectively charged and interpersonally meaningful question: “Do you
dismiss what we’ve talked about?” When I end up repeating the interpre-
tation that put her depression into an interpersonal context—as fear of
loss and an expression of need—the patient is affectively engaged with
me. This aspect of the interaction highlights my readiness to get involved.
In addition, even though Ms. AA seemed unable to see connections
between her depression and recent events, and even though she failed to
validate my interpretive efforts, I believe these seemingly futile activities
will make it increasingly difficult for her to ignore such connections in the
future. These activities will change Ms. AA’s thinking from teleological to in-
tentional. Whether the mechanism is conceptualized as education, trans-
muting internalization, or identification with the aggressor, or whether
the effect is to subtly reconfigure the brain’s neurophysiological response,
the result is to introduce a disruptive new way of thinking. I believe that
the prospect this offers—that her depressive experience can develop per-
sonal meaning—is an irresistibly hopeful and compelling message.
In this process, I am adding small increments to her capacity for
thinking or introspection that will eventually allow her to be able to put
words to her experience. A bridging conceptualization is offered by Fon-
agy’s term mentalization, mentioned earlier in this chapter. Ms. AA’s
childhood experience did not arm her with ways to label or identify her
moods or her motivations. Her feelings and motivations during child-
hood either were mislabeled or were left unidentified, with the result that
she is unable to conceptualize or communicate her inner experiences.
From this perspective, the recurrence of her depression is a testimonial
to the fact that she has not yet developed the ability to recognize the sub-
tle precursory feelings or cognitions related to her depressed moods
(e.g., her anxiety about the therapist’s potential lack of interest or the
early signals of her yearnings for supportive attention) that could foretell
and forestall the emerging depressive state of mind.
The most central issue for helping borderline patients’ insights to oc-
cur during the second phase involves helping patients understand how
their wishes for caring attention prompt their interpersonal demands
and evoke the rejections or anger that they fear (an issue repeatedly ad-
dressed with Ms. AA). This issue is important for no one more than the
therapist, who must help patients accept that their wish for caring atten-
tion is understandable and acceptable and that having those wishes frus-
trated prompts many of their behavior problems. Although this issue
sounds like transference analysis, it is usually first identified in situations
outside the therapy: for example, “I knew that when your mother went on
vacation you were likely to start drinking” or “When you leave the halfway
house, as much as you hate it, it is going to represent a big loss for you.”
Such interpretations of meanings assigned to events by borderline pa-
tients underscore the therapist’s role as an interested observer.
Other interpretations involve the defensive role of behaviors: for ex-
ample, “You know when you yell that your husband will comply” or “Tak-
ing these drugs prevents you from feeling weak.” Again, these are not
transference interpretations; they are designed to increase self-awareness,
and in the process, they help patients to appreciate the therapist’s ability
to make their life more understandable. The primary use of transference
interpretations involves the borderline patients’ subtle or indirect expres-
sions of hostility. Hostilities can become endemic if not addressed, but
hostility needs to be identified in a natural, instructive way, without imply-
ing that the patient has offended or scared the therapist—rather, the ther-
apist can invite a more direct expression.
When interpretations are met by hostility, the patient’s feelings need
to be respected, but a therapist ought not to be apologetic; making ob-
servations is essential to a therapist’s ability to be helpful. Indeed, I offer
patients such observations in a psychoeducational way and buttress my
observations by citing how well known and familiar such patterns of re-
sponse are. In this way, the interpretation becomes neutralized. In much
this same way, I believe Benjamin (1993) combines education with dy-
namic formulation, and Young et al. (2003) actually includes written
summaries of patients’ alternative modes (see Chapter 11).
Obviously, the belief held by many borderline patients that “psycho-
therapy might help” after 6 months is enhanced by advances made both in
the relational alliance and by any actual learning that has taken place. Al-
though the former is essential, it should never be considered sufficient. By
the end of 1 year, the patient should be involved in therapy and attached to
the therapist (see Table 12–4). This is another sign that the patient has
fully achieved the goal of a relational (affective and empathic) alliance.
Continued involvement and investment by the therapist—as shown by
reliability, interest, and good judgment—evokes hope about the relation-
ship that, in the first phase of treatment, is often experienced as dangerous
vulnerability (“I’ll get hurt, rejected,” etc.). Still, most borderline patients
consciously entertain the idea, some if not most of the time, that “this ther-
apist cares.” This idea only gradually becomes a conviction after actual ex-
perience with the therapist, quite independently of whatever idealized or

devalued attributions are professed about the therapist. (In response to re-
quests for reassurance about caring, it is best to tell patients that the only
way to know is through experience.) Unrealistic sexual or nurturant ex-
pectations (i.e., transferences) may fuel a borderline patient’s attachment
and often will endure well through the second year of a therapy. The qui-
eter conviction about the therapist’s caring for the patient is the bedrock
for establishment of the relational alliance, and it develops slowly but in-
exorably from the largely nonverbal and nonspecific experiences in the re-
lationship (i.e., reliability, concerned attention, acceptance).
Phase 3: Positive Dependency

Between 6 and 18 months of therapy, as has been shown in the previous ex-
amination of successful therapies (Waldinger and Gunderson 1989;
Wallerstein 1986), positive dependency should have evolved. Dependency
does not necessarily mean wanting to be told what to do; it primarily in-
volves extreme sensitivity to the therapist’s moods, attitudes, and absences.
This type of relationship gets established with nondynamic clinicians too,
such as cognitive-behavioral therapists, case managers, or psychopharma-
cologists, although in these relationships the dependency is more apt to in-
clude expectations of direction and reassurance. In essence, the therapist
has become a transitional object (Sidebar 12–3).
Sidebar 12–3: Transitional Objects:

From Concept to Phenomenon
Winnicott (1953) identified the phenomenon in which chil-
dren struggling with the recognition that their caregivers
are separate from, not extensions of, themselves adopt in-
animate transitional objects, whose presence can diminish
their anxieties and whose absence causes great distress.
This use has been confirmed (Roig et al. 1987). Although the
use of transitional objects is not an uncommon phenome-
non in normal development, it is particularly common
(about 70%) for patients with BPD, and it is significantly more
common among them than among patients with other—
most particularly, antisocial—personality disorders (Arkema
1981; Cardasis et al. 1997; Horton et al. 1974; Morris et al.
1986). The sustained attachment to transitional objects in
adults remains one of the simplest and most pathogno-
monic indicators for the diagnosis of BPD to nursing staffs,
who bear witness to the importance placed on such ob-
jects (e.g., timeworn dolls, blankets, pandas) when border-
line patients come to stay in hospital or residential settings.
Many adults who are not borderline also have transi-

tional objects. This observation may help clinicians assign
less stigma and adopt a more curious approach to this
phenomenon. The emergence of new transitional objects
during the course of long-term therapy may even be
viewed as a sign of relatedness that reflects improvement.
Without question, this has been my experience in multiple
instances in which patients have obtained a pet that
serves this role. It is as if they are beginning to “generalize”
their experience with me into more trust of themselves. In
another instance, an improving patient was willing and
able to give her transitional object (a stuffed animal) to her
child “to take care of.”
Many borderline patients will deny being dependent on a therapist.

Acknowledging their dependency is a cause of anxiety—particularly sep-
aration (“I’ll be left”) and paranoid (“I’ll be mistreated”) anxieties—that
may evoke considerable defensive use of devaluation. Patients typically
confess that “my therapist means too much,” thereby reflecting their de-
pendency and their apprehension about it. Marking progress into phase
3, patients should no longer be denying a dependent attachment.
Under these circumstances—when dependency is acknowledged and
the therapist is valued—patients are less resistant to self-disclosure and
more responsive to learning from a therapist’s observations. Many of the
testing behaviors and boundary problems that characterized the first year
are significantly diminished. The work of connecting feelings to situa-
tions and behaviors remains central. Similarly, the recurring themes are
needing caring attention and learning how frustration can be managed
without action. These themes can now be more easily addressed within
the context of patients’ responses to their therapists. In this period of
therapy, the exchanges can be quite intense, and a therapist’s composure
and containment can usually provide the needed holding without need-
ing case management (parameters) or a second modality. Learning to
think about the relation of cause and effect, regarding both feelings and
interpersonal relationships, introduces delays of impulse discharge or
avoidance behaviors. This learning to “think first” helps build affect tol-
erance. Mentalization, or being able to conceptualize, like any new habit,
requires much repetition to be internalized (speaking psychologically)
or to be embedded as new neural circuits (speaking biologically).
The Menninger study of alliance included a microanalysis of taped ses-
sions from 39 BPD patients that documented multiple shifts in the level of
collaboration (working alliance) within sessions (Allen et al. 1990). This
work showed that advances apparent at one point in a session or even over
longer periods will be dramatically reversed and then slowly regained.

Nevertheless, I believe that the ontogeny of the alliance is a dialectic pro-
cess, in which the more mature working form of alliance progressively be-
comes more resilient and persistent over the course of therapy (see Figure
12–1), whereas the regressions from collaboration become less long-lived
and less dramatic. Within this iterative process, broad generalizations can
be made about the time frame by which signs of a developing alliance
should be noted. In phase 2, signs of a developing working alliance involve
the ability to hear feedback without flight and with ability to think about
the feedback. The absence of such signs by 6 months is sufficiently trou-
blesome that the viability and effectiveness of the therapy become ques-
tionable. For therapists, as Gabbard et al. (1994) pointed out, the rapid
vicissitudes of the working alliance within sessions require therapists to be
deft, resourceful, and adaptive in how they respond. In particular, Gab-
bard et al. noted that interpretations are “high-risk, high-gain” interven-
tions. This statement is especially true for the transference interpretations
believed by Kernberg to be central.
Of value within the hurts and confusion resulting from projections,
misunderstandings, and intense feelings is an ongoing review of what
transpires in sessions in the interaction between patient and therapist.
Not only does this mean a review within sessions of what was said, meant,
and so forth; this process is also greatly assisted by having sessions tape-
recorded (Robbins 1988)—a technique introduced by Martin Orne to
help Sylvia Plath (who probably had BPD; see Freed 1984). The encour-
agement to tape-record sessions is sometimes resisted, but once it is be-
gun, borderline patients are usually quite responsive to what they can
learn. Tape recording makes possible a quite specific clarification of what
“really” occurred (it is very nice to have a borderline patient volunteer
that he or she understood what you said or why you said it). In addition,
the tapes also serve (as can other office items) as concrete extensions of
the therapist’s involvement and attention between sessions—that is, as
transitional objects (see Sidebar 12–3).
Therapists will allow themselves to be transitional objects and will
want to make the silent functions that they serve as explicit as possible.
Thus, for example, just as the therapist’s task in the first year was helping
the patient to understand that his or her actions stemmed from feelings
and relational needs, in this second year (and third phase), it is valuable
for the therapist to help the patient to identify what the patient depends
on the therapist for. The essential component of this process will involve
issues of not being alone and of feeling connected—in effect, issues in-
volving object constancy. Patients’ ability to recognize this need to avoid
aloneness will make such experiences more easily managed. Table 12–6
shows a hierarchy of ways in which prolonged separations from therapists

can be managed. Although interpretations of intolerance of aloneness
run the risk of imposing theory on a patient’s experience, it is worth mak-
ing these interpretations early and often because awareness of this di-
lemma can so effectively diminish unnecessary regressive responses and
unnecessarily heroic acts of availability by therapists (see Gunderson
1996 and Sidebar 12–4).
Still, most interpretive or confrontational activity in the second year
remains in the domain of connecting feelings and behaviors to interper-
sonal situations. Even though this activity occurs increasingly within the
relationship to the therapist, it still does not quite conform to transfer-
ence interpretations; the focus of activity is learning to know oneself in
new ways and to be more of aware of the therapist’s mental states.
TABLE 12–6. Hierarchy of transitional options for use during

therapist absences
1. Therapist accessible by telephone
As-needed call
Prescheduled call
2. Therapist substitutes: coverage by colleagues
Prescheduled meetings
Meetings to be requested by the patient as needed
3. Therapist-associated transitional objects
Tape-recorded sessions
Notes from the therapist
Cognitive-behavioral directives (“what to do”)
Items from the therapist’s office
4. Self-initiated coverage options
Increased contact with friends or relatives
Increasing social networking (e.g., events, clubs)
Distracting oneself (e.g., travel, movies)
Note. These options are generally needed only for absences of more than a
week. Options are listed hierarchically from most soothing to least.
Source. Reprinted from Gunderson JG: “The Borderline Patient’s Intolerance of
Aloneness: Insecure Attachments and Therapist Availability.” American Journal of
Psychiatry 153(6):752–758, 1996. Used with permission.
Sidebar 12–4: Is Regression Therapeutic?

The Two Margarets
Although the capacity for borderline patients to regress in
unstructured situations (from Rorschach tests to workplaces)
was one of the first defining characteristics of borderline pa-
tients (Gunderson and Singer 1975), the clinical implications
of this capacity have always teased therapists. There is the
standard-issue warning: regressions are a danger in unstruc-
tured therapies. This danger was forcefully detailed by psy-
choanalysts who described the development of psychotic
transferences in analyses (Hoch and Polatin 1949) or even
psychotherapies (Frosch 1970; Zetzel 1971). Equally strong
warnings were voiced against the danger of regression
within unstructured milieu programs (Adler 1973; Knight 1953).
In contrast, in another tradition in psychoanalytic writ-
ing, regression is considered a necessary component of
transference—necessary for a full transference to become
convincingly evident in both patient and analyst. What is
more difficult to trace is the idea that the therapist/analyst
might then provide a corrective experience by fulfilling the
patient’s transference needs. Yet this is evident in Balint’s
(1992) book The Basic Fault: Therapeutic Aspects of Regres-
sion and is illustrated quite dramatically in practices de-
ployed, with apparent success, by D.W. Winnicott in his
treatment of Margaret Little (Little 1981).
Winnicott spoke openly with Dr. Margaret Little about
the importance of completing a “full regression,” even
though Dr. Little worked as a therapist more or less contin-
uously during the 8 years of her treatment (Little 1981). Win-
nicott responded to “the terrified child” that Little became
“like” when she was in analysis. He was hesitant to note Lit-
tle’s strength because in her “borderline” state (Little 1981,
p. 24), she feared it meant losing him. He held her hands
and extended the length of sessions when she was silent.
He visited her daily at her home when she was ill. When she
became terrified, he held her head and interpreted it as re-
living her birth. He condemned her mother while explicitly
providing the nurturance he felt Little’s mother had failed
to give, and to encourage Little’s regression, he hospital-
ized her during one of his absences. Little reported that as a
result of the therapy, she went on to a more satisfying and
stable interpersonal life and a successful career as a train-
ing analyst. She explicitly credited the power and learning
of her “full regression” for enabling her to “find and free
[her] true self” (Little 1981, p. 38).
Thirty-some years after Margaret Little’s therapy, an-
other female psychiatrist, Margaret Bean-Bayog, became
infamous for efforts that were equally heroic, equally cre-
ative, and equally regressive. Bean-Bayog, too, did this to

treat a very bright, promising, and tormented Harvard med-
ical student with BPD. Here too, the therapist invited the pa-
tient’s regression via addressing “the child within.” She read
children’s books, inscribed gifts “for the baby,” made her-
self available over vacations, arranged hospitalizations dur-
ing her absences, and deliberately tried to provide Paul, her
patient, with the nurturant attention he allegedly lacked
from his mother. Paul did not improve. He had a series of
hospitalizations that included a course of electroconvulsive
therapy, and after about 3 years, this therapy eventually
terminated. When Paul, unlike Margaret Little, ended up
dead 6 months later (ostensibly by suicide but possibly from
cocaine overdose), his family pursued legal action. As a re-
sult, the case became the featured topic of national news
and two books (Chafetz and Chafetz 1994; McNamara
1994), and also as a result, Margaret Bean-Bayog regretta-
bly lost her license to practice medicine.
Debates about the merits of regression are not related
exclusively to borderline patients. The merits of regression
also were actively debated in an earlier literature about
psychotherapy with schizophrenic patients (Gunderson
and Mosher 1975). Underlying the claims for the effective-
ness of long-term hospitalization were the claims of correc-
tive action that the holding or parenting experience given
during such hospitalizations could provide for patients of
many diagnostic types. Even now, reparenting is offered to
some otherwise treatment-resistant anorexic patients. Most
relevant is that “reparenting” is explicitly described as a
part of what SFT (Chapter 11) offers borderline patients.
In my view, what Margaret Bean-Bayog did was brave
and potentially defensible—if all else had failed. In the cur-
rent state of our knowledge about the course of BPD and
the benefits from effective therapies, it would be very diffi-
cult to justify. Perhaps Bean-Bayog just had the wrong pa-
tient, but I doubt it: regressive experiences clearly can be
powerfully harmful for borderline patients.
The Margaret Little case offers a rationale that regres-
sions might be powerfully helpful. However, I am skeptical
about Margaret Little’s conclusions. I suspect that an inten-
sive schedule of sessions in which Little received the caring
attention of an older, idealized man could have accom-
plished as much—maybe more quickly—without the re-
gressive transference enactments. At present, there may
still be a role for deliberately regressive psychotherapy, but
because of the dangers, regressions should be encour-
aged only when other therapies have failed and when the
regressions are conducted with informed patient collabo-
ration and ongoing professional consultations.
A patient of mine once told me, “A need fulfilled will go

away.” I smiled skeptically, but sometimes I wonder.
The conclusion of phase 3 can occur as early as the end of the second
year of therapy and usually occurs within year 3. At this point, the border-
line patient has acquired a capacity for stable, supportive relationships
and a capacity for stable, low-demand work (see Table 12–4). At this
point, many borderline patients can successfully leave therapy. They can
get on with their lives if they have the good fortune of having established
stable, supportive living or working situations. It is not unusual, for exam-
ple, for borderline patients to find a romantic partner, or even a spouse,
whose presence can greatly diminish the relational needs served by a
therapist. Others find stable supports from extended families, self-help
groups, or church communities that are sufficient. Borderline patients
are still insecure about rejections, fearful about separations, and prone to
cut themselves, drink, binge, rage, or withdraw in the face of conflicts.
However, such reactions are less severe and less prolonged than before
therapy or during phase 1. But patients are still unable to rely on a con-
sistent inner locus of control; they remain too reactive (defiant or com-
pliant) toward external pressures.
Phase 4: Secure Attachment, the Working

Alliance, and Consolidation of Self
At this point, the psychotherapeutic techniques are no longer very spe-
cific to the borderline patient’s psychopathology, except that the issues
remain those unique to this diagnostic group. Although this phase is the
least essential for mental rehabilitation, it is of the most indefinite dura-
tion (see Table 12–4).
A stable and increasingly secure relationship has formed, and a col-
laborative working alliance can generally be assumed. The capacity for a
secure attachment to the therapist may at last become evident, meaning
an attachment in which absences may cause anxiety or objections but do
not require substitutes or any therapist-associated objects (see Table 12–
6). The relationship is no longer contaminated by fears of rejection or
abandonment; and criticisms, although unwanted, can be responded to
effectively.
The direct expression of hatefulness toward the therapist that in
Kernberg’s theory is needed to remedy core psychopathology may occur
during this phase. This behavior remains, in my experience, a critical
process in rendering a borderline patient nonborderline. This process is
not always possible: deeply ingrained moralistic prohibitions or deficient
intellectual or organizational capabilities can prevent it. Nor does this

process usually occur in the cathartic way that I had imagined. Rather, it
is more likely to occur in the form of direct, cruel indictments over a long
time, for which the therapist has become a safe container.
In this phase of therapy, long-denied problems with early trauma can
be revisited usefully, or the developmental origins of distortions in body
image can be explored. Such issues may take years to open up and to have
the needed desensitization or resolution occur. This process involves a
patient’s obtaining a coherent narrative of his or her life, without major
gaps, thereby consolidating a sense of self.
Entering competition is always both desirable and conflictual for bor-
derline patients insofar as it triggers fears of aggression and of rejection.
In addition to clarifying such fears, therapists often need to actively urge
borderline patients to compete. Competition requires that borderline
patients take initiatives on their own behalf without guilt.
The acquisition of stable, nonsexual, intimate relationships is almost
certainly a sign that someone is no longer borderline. The conclusion of
this phase is marked by the patient’s fullness of life—his or her investment
in work and satisfaction from it and from relationships outside therapy.
Impasses
Individual psychotherapies rarely achieve the initial and mutual goal of
curative—or at least basic personality structure—changes. Table 12–7
identifies the common reasons that impasses occur. The reasons vary
within each phase of treatment.
Notably, whereas too much frustration in phase 2 causes dropout, too
little causes regression. Because inexperienced therapists tend to worry
too much about frustrating and thereby losing patients, they may become
targets for devaluation and dismissal by borderline patients; or, more of-
ten, these therapists create a chronically dependent and potentially re-
gressive relationship (see Sidebar 12–4). A major concern that all
therapists need to be aware of is the capacity for borderline patients to re-
gress in therapies that are too unstructured or seductive. This issue usu-
ally is not as obvious as that described in Sidebar 12–4, on the two
Margarets. The issue more often takes the form of a patient’s silent belief
that his or her therapist is doing and will continue doing for the patient
what he or she found lacking in his or her early parental relationships: lis-
tening kindly and empathically and offering an opportunity to be under-
stood nonjudgmentally, spiced by some sound advice. Although this is
not exactly a transference (insofar as the patient’s attribution is exactly
what the well-meaning therapist would say he or she is intentionally
TABLE 12–7. Major reasons for impasses

(“can’t leave, can’t progress”): chronicity
PHASE 2 PHASE 3 PHASE 4
(1 MONTH–1 YEAR) (1 YEAR–2 OR 3 YEARS) (2 OR 3 YEARS–?)
Too much frustration Inadequate relational Insufficient task

focus (transference) (exploratory)
orientation
Too little frustration Insufficient vocational Insufficient support for
expectations nontherapy
relationships (for life
to replace therapy)
Insufficient attention Overvaluation of the Underestimation of the
to functioning role of psychotherapy potential for normality
doing), the therapist is serving a parental role without examining its

meaning and in the process may be perpetuating expectations for rela-
tionships that are unrealistic. Hence the patient is grateful and depen-
dent, but he or she is making no progress in the capacity for mature
relationships outside the therapy. This type of therapeutic impasse, in my
experience, is usually accompanied by a failure to address the patient’s
functional impairment. Such therapists may be too ready to sympathize
with their patients’ complaints about employers who are misunderstand-
ing or their complaints about being assigned the same performance stan-
dards as other, less-handicapped people.
On the other side of this dilemma is the reality that too much frustra-
tion will cause flight. “Too much frustration,” with some borderline pa-
tients, may be very little. Some will perceive that any lack of reassurance,
lack of indication of care, or sign of inattention is evidence of a therapist’s
cruelty or disinterest. The concept of a negative therapeutic reaction is ap-
propriate here. In these circumstances, readiness for psychotherapy
needs to be evaluated. When “too much frustration” means an intoler-
ance for confrontations or interpretation, then either less use of these
techniques is required or the role of a second modality becomes critical.
Early in the course of psychotherapies (phase 1 or phase 2), this second
modality not only buffers transference distortions but also fulfills other
and complementary goals (other modalities, e.g., sociotherapeutic or
pharmacological).
As a further step-down refinement of the principle of split treatment,
it is sometimes helpful to introduce a second individual clinician during
the course of an intensive individual psychotherapy that is already under
way. Stiver (1988) discussed how the introduction of an ongoing consul-

tant (for 2–6 months) can mitigate the regressive (or, I would add, rage-
ful) transferences that can develop. A further variation of this that I have
seen successfully used is the introduction of an intermittent cotherapist.
The primary problem with such arrangements is not the danger of split-
ting but the potential that such use of a cotherapist may, if unexamined,
verify the patient’s belief in the destructiveness of his or her hostilities.
Summary
Much of the clinical literature about psychoanalytic therapies has wres-
tled with the relative merits of supportive, attachment-enhancing inter-
ventions and those that are more explicitly insight-enhancing ones. In
this chapter, I underscore the necessity of both the relational and the
learning components if therapies are to be successful. Both empirical
work and clinical experience document the overriding importance of
supportive forms of interventions (e.g., empathy, validation, reassurance,
clarification) during the second phase of therapy if the therapy is con-
ducted within the agreed-on usual framework. Nonetheless, it is of critical
importance, even early in therapies, to underscore the tasks of therapy: to
learn about oneself and to change as a result of what is learned. When this
task orientation is combined with the development of a trusting and de-
pendent relationship, the borderline patient will be able to function, and
the therapy will move into a third phase. During this period, the focus in
sessions is often on the borderline patient’s learning to identify his or her
feelings and how they relate to the therapist’s behaviors or words. The
gradually improved ability to understand feelings correctly and to accept
unsupportive feedback (e.g., interpretation, confrontations, impatience,
criticisms) enables the patient to form stable relationships. The fourth
phase of therapy will be more fully insight oriented, and the patient’s
ownership of hostilities and resolution of developmental failures allow
the previously borderline patient to compete and to take independent,
self-serving initiatives.
The hope that borderline patients can undergo curative change from
psychodynamic psychotherapies is justified, but such change rarely oc-
curs. It is critically important that therapists appreciate the sequence of
changes and their approximate timetable so as not to foreshorten unwit-
tingly this long-term process.
References
Abend SM, Porder MS, Willick MS: Borderline Patients: Psychoanalytic
Perspectives. New York, International Universities Press, 1983
Adler G: Hospital treatment of borderline patients. Am J Psychiatry 130:

32–35, 1973
Aronson, 1986
Allen JG, Gabbard GO, Newsom GE, et al: Detecting patterns of change
in patients’ collaboration within individual psychotherapy sessions.
Psychotherapy 27:522–530, 1990
Appelbaum AH: Supportive psychotherapy, in The American Psychiatric
Publishing Textbook of Personality Disorders. Edited by Oldham JM,
Skodol AE, Bender DS. Washington, DC, American Psychiatric Pub-
lishing, 2005, pp 335–346
Arkema PH: The borderline personality and transitional relatedness. Am
J Psychiatry 138:172–177, 1981
Aronson TA: A historical perspective on the borderline concept: a review
and critique. Psychiatry 48:209–222, 1985
Balint M: The Basic Fault: Therapeutic Aspects of Regression. Evanston,
IL, Northwestern University Press, 1992
Bateman A, Fonagy P: Psychotherapy for Borderline Personality Disor-
der: Mentalization-Based Treatment. Oxford, England, Oxford Uni-
versity Press, 2004
Brandschaft B, Stolorow RD: The borderline concept: an intersubjective
viewpoint, in The Borderline Patient, Vol 2. Edited by Grotstein JS,
Solomon MF, Lang JA. Hillsdale, NJ, Analytic Press, 1987, pp 103–126
Cardasis W, Hochman JA, Silk KR: Transitional objects and borderline
Chafetz GS, Chafetz ME: The Bizarre Relationship Between a Prominent
Harvard Psychiatrist and Her Suicidal Patient. New York, Crown, 1994
Chessick RD: Intensive Psychotherapy of the Borderline Patient. New
sonality. New York, Wiley, 1999
sonality: Focusing on Object Relations. Washington, DC, American
Psychiatric Publishing, 2006
try 164:922–928, 2007
erations in the treatment of a borderline patient. Int J Psychoanal 72
(Pt 4):639–656, 1991
malfunction in borderline personalities. Int J Psychoanal 76 (Pt
1):39–44, 1995
Fonagy P, Target M: Playing with reality: theory of mind and the normal
development of psychic reality. Int J Psychoanal 77 (Pt 2):217–234,
1996
Fonagy P, Leigh T, Kennedy R, et al: Attachment, borderline states and
the representation of emotions and cognitions in self and other, in
Emotion, Cognition, and Representation (Rochester Symposium on
Developmental Psychopathology, Vol 6). Edited by Cicchetti D, Toth
SL. Rochester, NY, University of Rochester Press, 1995, pp 371–414
Freed AO: Differentiating between borderline and narcissistic personal-
ities. Soc Casework 65:395–404, 1984
Frosch J: Psychoanalytic considerations of the psychotic character. J Am
Psychoanal Assoc 18:24–50, 1970
Borderline Patients. Washington, DC, American Psychiatric Press,
1994
Gabbard GO, Horwitz L, Frieswyk S, et al: The effect of therapist inter-
vention on the therapeutic alliance with borderline patients. J Am
Psychoanal Assoc 36:697–727, 1988
Gabbard GO, Horwitz L, Allen JG, et al: Transference interpretation in
the psychotherapy of borderline patients: a high-risk, high-gain phe-
nomenon. Harv Rev Psychiatry 2:59–69, 1994
try 6:649–658, 2006
Gill MM: The analysis of the transference. J Am Psychoanal Assoc
27(suppl):263–288, 1979
Giovacchini PL: The psychoanalytic paradox: the self as a transitional ob-
ject. Psychoanal Rev 71:81–104, 1984
Goldstein WN: Introduction to Borderline Conditions. New York, Jason
Aronson, 1985
Grotstein JS: Transitional phenomena and the dilemma of the me/not
me interface, in The Solace Paradigm: An Eclectic Search for Psycho-
logical Immunity. Edited by Horton PC, Gewirtz H, Kreutter KJ. Mad-
ison, CT, International Universities Press, 1988, pp 59–74
758, 1996
Gunderson JG, Chu JA: Treatment implications of past trauma in border-
line personality disorder. Harv Rev Psychiatry 1:75–81, 1994
47:679–704, 1999
Gunderson JG, Mosher LR (eds): Psychotherapy of Schizophrenia. New
Gunderson JG, Singer M: Defining borderline patients: an overview. Am
J Psychiatry 132:1–10, 1975
Gunderson JG, Frank AF, Katz HM, et al: Effects of psychotherapy in
schizophrenia, II: comparative outcome of two forms of treatment.
Schizophr Bull 10:564–598, 1984
1989
Gunderson JG, Najavits LM, Leonhard C, et al: Ontogeny of the thera-
peutic alliance in borderline patients. Psychother Res 7:301–309,
1997
Hoch P, Polatin P: Pseudoneurotic forms of schizophrenia. Psychiatr Q
23:248–276, 1949
Horton PC, Louy J, Coppolillo HP: Personality disorder and transitional
relatedness. Arch Gen Psychiatry 30:618–622, 1974
Horwitz L: Clinical Prediction in Psychotherapy. New York, Jason Aron-
son, 1974
Horwitz L: Clinical and projective assessments of borderline patients. Pa-
per presented at a symposium, The Borderline Patient: A Multiaxial
View, Department of Psychiatry, University of Miami School of Med-
icine, Key Biscayne, FL, 1982
Horwitz L, Gabbard GO, Allen JG, et al: Borderline Personality Disorder:
Tailoring the Psychotherapy to the Patient. Washington, DC, Ameri-
can Psychiatric Press, 1996
Howard KI, Kopta SM, Krause MS, et al: The dose-effect relationship in
psychotherapy. Am Psychol 41:159–164, 1986
15:641–685, 1967
Kernberg O: Final report of the Menninger Foundation’s Psychotherapy
Research Project: summary and conclusions. Bull Menninger Clin
38:181–195, 1972
Kernberg OF, Burstein ED, Coyne L, et al: Psychotherapy and psycho-
analysis: final report of the Menninger Foundation’s Psychotherapy
Research Project. Bull Menninger Clin 36:1–275, 1972
Kernberg O, Selzer M, Koeningsberg HW, et al: Psychodynamic Psycho-
Knight R: Borderline states. Bull Menninger Clin 17:1–12, 1953
Koenigsberg HW, Kernberg OF, Stone MH, et al: Borderline Patients: Ex-
tending the Limits of Treatability. New York, Basic Books, 2000
Kohut H: The Analysis of the Self. New York, International Universities
Press, 1971
Kroll J: The Challenge of the Borderline Patient: Competency in Diagno-
sis and Treatment. New York, WW Norton, 1988
Langs R: The Bipersonal Field. New York, Jason Aronson, 1976
Linehan MM: Theory and treatment development and evaluation: reflec-
tions on Benjamin’s “models” for treatment. J Personal Disord 11:
325–335, 1997
Little MI: Transference Neurosis and Transference Psychosis. New York,
Jason Aronson, 1981
Lyons-Ruth K, Bruschweiler-Stern N, Harrison AM, et al: Implicit rela-
tional knowing: its role in development and psychoanalytic treat-
ment. Infant Ment Health J 19:282–289, 1998
Masterson J: Treatment of the adolescent with borderline syndrome (a
problem in separation-individuation). Bull Menninger Clin 35:5–18,
1971
Masterson J: Psychotherapy of the Borderline Adult. New York, Brunner/
Mazel, 1976
McGlashan TH: Implications of outcome research for the treatment of
borderline personality disorder, in Borderline Personality Disorder:
Etiology and Treatment. Edited by Paris J. Washington, DC, Ameri-
can Psychiatric Press, 1993, pp 235–260
McNamara E: Sex, Suicide, and the Harvard Psychiatrist. New York,
Pocket Books, 1994
Meares R, Stevenson J, Comerford A: Psychotherapy with borderline pa-
tients, I: a comparison between treated and untreated cohorts. Aust
N Z J Psychiatry 33:467–472, 1999
Modell AH: Primitive object relationships and the predisposition to
schizophrenia. Int J Psychoanal 44:282–291, 1963
Modell AH: “The Holding Environment” and the therapeutic action of
psychoanalysis. J Am Psychoanal Assoc 24(2):285–307, 1976
Morris H, Gunderson JG, Zanarini MC: Transitional object use and bor-
derline psychopathology. Am J Psychiatry 143:1534–1538, 1986
Najavits LM, Gunderson JG: Better than expected: improvements in bor-
derline personality disorder in a 3-year prospective outcome study.
Compr Psychiatry 36:296–302, 1995
Paris J: Working With Traits: Psychotherapy of Personality Disorders.
Northvale, NJ, Jason Aronson, 1998
Rinsley D: Borderline and Other Self Disorders. New York, Jason Aron-
son, 1982
Robbins M: Use of audiotape recording in impasses with severely dis-
turbed patients. J Am Psychoanal Assoc 36:61–75, 1988
Roig E, Roig C, Soth N: The use of transitional objects in emotionally dis-
turbed adolescent inpatients. Int J Adolesc Youth 1:45–58, 1987
Sabo AN, Gunderson JG, Najavits LM, et al: Changes in self-destructive-
ness of borderline patients in psychotherapy: a prospective follow-up.
J Nerv Ment Dis 183:370–376, 1995
Searles HF: My Work With Borderline Patients. Northvale, NJ, Jason
Aronson, 1986
Skodol AE, Shea MT, McGlashan TH, et al: The Collaborative Longitudi-
nal Personality Disorders Study (CLPS): overview and implications.
J Personal Disord 19:487–504, 2005
Stanton AH, Gunderson JG, Knapp PH, et al: Effects of psychotherapy in
schizophrenia, I: design and implementation of a controlled study.
Schizophr Bull 10:520–563, 1984
Stern DN, Sander LW, Nahum JP, et al: Noninterpretive mechanisms in
psychoanalytic therapy: the “something more” than interpretation.
Int J Psychoanal 79:903–921, 1998
1992
Stevenson J, Meares R: Psychotherapy with borderline patients, II: a pre-
liminary cost benefit study. Aust N Z J Psychiatry 33:473–477, 1999
Stiver IP: Developmental dimensions of regression: introducing a con-
sultant in the treatment of borderline patients. McLean Hospital
Journal 13:89–107, 1988
Stolorow RD: An intersubjective view of self-psychology. Psychoanalytic
Dialogues 53:393–399, 1995
Stone MH: Treatment of borderline patients: a pragmatic approach. Psy-
chiatr Clin North Am 13:265–285, 1990
Stone MH: Abnormalities of Personality: Within and Beyond the Realm
of Treatment. New York, WW Norton, 1993
Volkan VD: Six Steps in the Treatment of Borderline Personality Organi-
zation. New York, Jason Aronson, 1987
Waldinger RJ, Gunderson JG: Effective Psychotherapy With Borderline
Patients: Case Studies. Washington, DC, American Psychiatric Press,
1989
Wallerstein R: Forty-Two Lives in Treatment. New York, Guilford, 1986
Winnicott DW: The mirror role of mother and family in child develop-
ment, in The Predicament of the Family: A Psychoanalytic Sympo-
sium. Edited by Lomas P. London, Hogarth Press, 1967, pp 26–33
Yeomans FE, Clarkin JF, Kernberg OF: A Primer for Transference Fo-
cused Psychotherapy for the Borderline Patient. Northvale, NJ, Jason
Aronson, 2002
Young JE: Cognitive Therapy for Personality Disorders. Sarasota, FL, Pro-
fessional Resource Exchange, 1990
Zanarini MC, Frankenburg FR, Hennen J, et al: Psychosocial functioning
of borderline patients and Axis II comparison subjects followed pro-
spectively for six years. J Personal Disord 19:19–29, 2005
Zetzel ER: A developmental approach to the borderline patient. Am J Psy-
chiatry 127:867–871, 1971
Chapter 13
FUTURE CONSIDERATIONS
PATIENTS WITH borderline personality disorder (BPD) are an impos-

ing presence within clinical populations. Although academics still argue
about the validity of the diagnosis, it is now firmly accepted by clinicians
(B. Pfohl, K. Silk, C. Robins, M. Zimmerman, and J. Gunderson, “Atti-
tudes Towards Borderline PD: A Survey of 752 Clinicians,” unpublished
data, May 1999). This is a significant achievement because use of this cat-
egory began quite tentatively in the 1970s and only became official in
1980. However, some glaring voids in our knowledge remain. There is
still a lack of good epidemiological data to document the incidence, prev-
alence, and natural comorbidities of BPD. There is also a lack of careful
documentation of the enormous social and public health significance of
BPD (i.e., how its effect extends into legal, child care, social service, and
educational systems). The diagnosis is still relatively new, and, within the
current historical context of radical changes in health care and the neu-
rosciences, the diagnostic construct and the treatments recommended
for these patients will undoubtedly undergo further change in the years
ahead. In this chapter, I anticipate what these changes might involve.
Treatment Implications
Development of Specialists and Special Services
This book’s recurrent themes have implications for the care of border-
line patients. First, a range of services is usually needed. It is and always
has been rare for any individual with BPD to make major gains from any
one therapy. The hope that one person or one modality might effect such
gains has most frequently been attributed to individual psychotherapy
delivered by experts. Yet accounts of successful individual therapy show
that it has almost always involved other modalities (considered adjunc-
tive), such as group therapies, hospitalization, medications, or family
315
meetings. Even with an optimal, skilled individual therapist (primary cli-

nician), treatment with other modalities will facilitate change and ad-
dress some problems better than individual therapy could.
A less explicit but even more compelling theme in this book is that in
the absence of standards of clinical care, bad things often happen: regres-
sions, psychotic transferences, noncompliance, violence, escalating self-
destructiveness, and lawsuits. Borderline patients achieved a bad reputa-
tion because when treated wrongly, they become worse, and their hostil-
ities and behavior problems become unmanageable. Because of this, the
American Psychiatric Association (2001) could justify developing prac-
tice guidelines for BPD treatment despite the relative absence of empir-
ically tested treatments.
An implication of these basic themes is that specialists and specialized
clinical services are needed. A proposal was outlined by the European
branch of the International Society for the Study of Personality Disorders
(ISSPD) in 1998 to develop identifiable centers in each European coun-
try for specialized training, research, and treatment services for person-
ality disorders, with obvious reference to borderline patients (ISSPD
Newsletter 1998). Such centers are clearly justifiable by virtue of the high
prevalence (15%–20%) of borderline patients in clinical services every-
where that they have been looked for (Gunderson 1999). Table 13–1
identifies the components included at McLean Hospital’s Center for the
Treatment of Borderline Personality Disorder. Unfortunately, centers for
BPD treatment that provide all levels of care and skilled, enthusiastic, col-
laborative practitioners in multiple modalities are very rare.
From this perspective, knowledgeable and comprehensive treatment
for BPD is rarely available. In the United States, most major departments of
psychiatry or major psychiatric hospitals and an increasing number of de-
partments of psychology have staff members who are identified as experts
on BPD, but very few have a BPD-related specialty service. Most of the time,
borderline patients have their treatment administered piecemeal by peo-
ple or services without skill, experience, or enthusiasm. This deplorable sit-
uation is even worse in most countries outside the United States.
Economic, scientific, and clinical forces all push for the development
of specialized services and clinical specialists to treat BPD. The economic
forces will be further pushed by documentation of the enormous public
health costs of this diagnostic group, from the perspectives of both health
care utilization (see Bender et al. 2001; Walker et al. 1999) and vocational
disability (e.g., Skodol et al. 2002). The economic incentive is magnified
by the current cost-benefit consciousness. Even managed care organiza-
tions, which dread long-term care, recognize that the revolving door in
hospitalizations and emergency departments signifies the failure and
Future Considerations ❘ 317
TABLE 13–1. Components of McLean Hospital’s borderline

personality disorder (BPD) specialty center
Special services at all levels of care primarily designed for BPD patients
Clinicians primarily dedicated to the care of BPD
Expertise (as measured by recognized leadership, e.g., invited talks,
publications)
All modalities available (minimally including psychopharmacology and
individual and group therapies)
Research on BPD
Special training programs for treating BPD
Family support services
Public advocacy and education
cost-ineffectiveness of nonspecialized and short-term treatments. Insofar

as the first year of treatment can make serious changes in self-destructive
and suicidal behaviors, it can greatly reduce health care costs. The costs
to society of unemployment and interpersonal conflicts far exceed health
care costs. To address these problems requires longer-term treatments
with both rehabilitative and interpersonal goals.
The scientific forces that encourage specialization for BPD are pro-
pelled by the increasing body of evidence that diagnosis-specific, manual-
guided therapies for BPD invariably exceed generic types (treatments as
usual) in their efficacy (and cost-effectiveness).
The clinical forces that push for specialized services are described
throughout this book: many clinicians do not like working with these pa-
tients; poorly conceived and executed therapies easily and frequently re-
sult in patients getting worse, with anger and despair for patients and
families and sometimes with subsequent liability issues for the treaters.
Above all, the proper care of these patients at every level of the health
care system, and with every modality, involves very distinctive features cus-
tomized to the particular needs of this patient group.
Standards of Care for BPD

In 1997, the development of BPD practice guidelines was authorized by
the American Psychiatric Association, and in fall 1998, the committee,
chaired by John Oldham, began the task of reviewing literature. The
committee’s initial draft was completed in spring 2000, and the resulting
guidelines were published in 2001 (American Psychiatric Association).
The same literature that guided the development of those guidelines also
informed the development of the earlier edition of this book (Gunder-

son 2001). The impressive growth in knowledge about treating BPD that
has prompted this current edition indicates that the 2001 American Psy-
chiatric Association practice guidelines are somewhat outdated. This is
especially true regarding psychopharmacology, the need for psychoedu-
cation, the use of short-term interventions, the value of skills training,
and empirical support for competing forms of psychotherapy.
The standards of care outlined in Table 13–2 reflect clinical evidence
that is consistent with—and increasingly buttressed by—empirically based
knowledge. Readers will recognize that the rationale for these standards is
reflected in the content of this book. Adoption of these standards would
require radical changes in the current health service system. It is probably
feasible to expect these standards to be met only in specialty centers (as
described earlier). The types and quality of care currently available to BPD
patients are highly variable. Without standards of care, a natural selection
process will continue, with the risk that cost containment will be more val-
ued than cost-benefit.
Diagnostic Implications
DSM-V
Some of the current personality disorders may be better conceptualized as
extreme variants of normal personality, whereas others may be better con-
ceptualized as spectrum variants of Axis I disorders (Figure 13–1). As
DSM-V emerges on the horizon, there has been some interest in trying to
put BPD within a dimensional system (Krueger et al. 2007). BPD’s exten-
sive validation, its severity of impairment, and, above all, the specificity of
its treatment needs, as documented throughout this book, distinguish it
from most other personality disorders, and this should be reflected by its
place in DSM-V. The safest way to do this is to retain BPD as a disorder of
the self (a personality disorder) but to take measures that ensure that it re-
ceives a priority akin to Axis I disorders for treatment and reimbursement.
Clinical Theories: The Search for the Core Psychopathology

Table 13–3 identifies major theories about the core (i.e., basic, underly-
ing, requisite) psychopathology of BPD that has derived from clinical
observations. Some of these theories have attributed core psychopathol-
ogy to constitutional (i.e., temperamental) variables with predominantly
genetic origins. Others have emphasized environmentally induced
character failures. These theories have all emphasized failures in early
preborderline children’s experience with their primary caregivers. Advo-
cates for any of these theories, as noted throughout this book, use them
TABLE 13–2. Standards of care for borderline personality disorder

(BPD)
BPD patients and their significant others should receive
psychoeducation about this diagnosis and its treatment (Chapters 1
and 3).
Treatments should be tailored to meet goals for change agreed to by the
BPD patient (Chapters 3, 4, and 6).
BPD patients should have a primary clinician (Chapter 4) who is
experienced with borderline patients or is under skilled supervision.
Impulsive BPD patients should have two or more collaborating
components in their treatment (Chapters 3–6, 8, 9) until they are
stabilized in the community.
The least restrictive level of care consistent with safety and social
rehabilitation should be used (Chapter 5).
BPD patients should be offered medications with the explicit
expectation of partial relief and with plans to test the effects of tapered
dosage every few months thereafter (Chapters 6 and 7).
Self-injurious patients should be offered cognitive-behavioral skills
training (Chapters 4 and 11).
Psychotherapy should be provided by therapists who are trained to give
BPD-specific therapies or who are under skilled supervision
(Chapter 10).
Psychodynamic psychotherapy should be reserved for those BPD
patients without disabling social and vocational impairments
(Chapter 12).
Note. Standards of care as of 2008. All chapters cited are in this book.
to organize their clinical practices. Increasingly, these theories are be-

coming subjects for research. Mahler and Kaplan (1977), for example,
showed that rapprochement failures, assigned a central role by Master-
son, are neither necessary nor sufficient cause for BPD. The theory of af-
fective dysregulation was disconfirmed by testing emotional responses to
excitatory photographs and finding that borderline subjects’ responses
were similar to those of nonborderline control subjects (Herpetz et al.
1999). On a more positive note, work on attachment failures by Fonagy et
al. (2000) and others (Lyons-Ruth 1998) seems to confirm their potential
explanatory power.
A significant recent development is the effort to identify the core di-
mensions of psychopathology from the scientifically based perspectives
320
❘
FIGURE 13–1. Relation of borderline personality disorder with other personality disorders (PDs).
Source. Adapted from Gunderson and Phillips 1995.
TABLE 13–3. The core psychopathology of borderline personality

disorder: seminal clinical theories
CORE PSYCHOPATHOLOGY THEORIST ATTRIBUTION
Excessive aggression Kernberg (1967) T or E

Abandonment anxieties Masterson (1972) E
secondary to
rapprochement failures
Affective dysregulation Klein (1975); Stone (1979); T
Akiskal (1981); Linehan
(1993); Livesley (1993);
Zanarini and Frankenburg
(1994)
Intolerance of aloneness Adler and Buie (1979); E
secondary to attachment Fonagy (1991); Benjamin
failures (1993); Gunderson (1994,
1996)
Impulse dyscontrol Links et al. (1999) T
Complex posttraumatic Herman (1992) E
stress disorder
Note. E= primarily an environmentally caused disposition; T=primarily a tem-
peramental disposition.
of neurobiology or academic psychology. Almost all such scientifically

based hypotheses about BPD’s core psychopathology involve formula-
tions of a core temperament—that is, the heavily genetically determined
way in which a child perceives and reacts to the environment.
Torgersen et al. (2000) estimated that BPD’s level of heritability was
0.68—a level that is close to the estimated heritability for bipolar disorder
(Kendler and Greenspan 2006). This finding has prompted the recon-
ceptualization of BPD as a “brain disease” that takes its place beside other
major and severe “real” mental illnesses.
Public Awareness and Advocacy

The dramatic expansion of communications via the Internet is making
information available about private concerns, such as psychiatric disor-
ders, that were previously uninvestigated because of fear or shame. The
privacy and distance that Internet communication allows has provided
isolated borderline individuals with a means of finding support from
peers (Ginther 1997; Silk 1997) and has encouraged families and others
to learn about this condition and find treatment resources. The Appen-
dix at the end of this book provides an index of the many types of infor-
mation available for consumers. This information is rapidly expanding in
amount, variety, and utilization.
Bridging the gap between clinicians and families has far-reaching con-
sequences. That family members usually do not define themselves as pa-
tients (Chapter 8) means that they can easily convert their parental hopes
for better treatment and their alliance with the treating clinicians into ad-
vocacy. This has been dramatically evident in the success of national or-
ganizations such as the National Alliance on Mental Illness (NAMI) and
the Depression and Bipolar Support Alliance (DBSA), formerly known as
the National Depressive and Manic-Depressive Association, which have
worked relentlessly to decrease stigma and increase research into mental
illnesses. The same need has long been evident for borderline patients:
they have a chronic disease of great public health significance that enor-
mously burdens families who too often are deeply shamed by it.
In 1994, Valerie Porr founded an organization in New York City, Treat-
ment and Research Advancements Association for Personality Disorder
(TARA APD), dedicated to advocacy. In 1995, graduates (“veterans”) of
McLean Hospital’s psychoeducational multiple-family groups (described
in Chapter 8) and of Perry Hoffman’s dialectical behavior therapy (DBT)
for families at the New York Presbyterian Hospital, Westchester Division,
formed similar organizations: the New England Personality Disorder As-
sociation (NEPDA) and the National Education Alliance for Borderline
Personality Disorder (NEA-BPD), respectively. These organizations have
drawn attention to the little-recognized plight of this population.
Both NEA-BPD (and its companion group NEPDA) and TARA APD
have helped convert the despair of BPD patients into protests and public
action. Their advocacy has largely been responsible for NAMI’s recent
adoption of BPD. These efforts will be buttressed by research document-
ing how comorbidity with BPD accounts for much of the resistance to
treatment of anxiety and mood disorders—disorders that have far more
recognition, research, and insurance support—and by documentation of
the enormous costs to society of BPD, both in disability and in liability.
The initiatives of both NEPDA and TARA APD include
• petitioning the National Institute of Mental Health and NAMI to

adopt BPD as a priority—long overdue for a population that consti-
tutes 15%–20% of clinical populations
• sponsoring talks that are open to the public
• publicizing newsletters and Web sites to disseminate information—
including research findings and events whereby parents and inter-
ested others may learn more and, more importantly, take action on
behalf of the mentally ill
• developing self-help groups modeled on Alcoholics Anonymous and
the DBSA.
Everyone in the mental health disciplines should applaud the efforts

of NEPDA and TARA APD. It has taken strength for families to become
spokespersons for the very offspring who often condemn them. Without
doubt, the goals of these organizations are the same as those of clinicians
who labor in different ways to be helpful.
Public awareness and advocacy has received support from two publi-
cized events. In the best-selling book Diana in Search of Herself: Portrait of a
Troubled Princess, her biographer, Sally Bedell Smith (1999), identified the
famous princess as having BPD. Also in 1999, Susanna Kaysen’s (1993)
book Girl, Interrupted became an acclaimed movie, in which the border-
line diagnosis, although viewed skeptically by Kaysen, was much dis-
cussed. Still, BPD awaits the public recognition that could come from a
living celebrity who has or had BPD and who would be willing to become
a spokesperson and work to educate the public about this disorder and
destigmatize it (Sidebar 13–1).
Sidebar 13–1: Were a Famous Borderline Person

to Go Public…
It would be a major step forward for someone with fame to
identify himself or herself publicly as having BPD. Of course,
most people with this disorder are either too dysfunctional
to achieve fame or, if able to achieve it, too ashamed of or
insecure about their inner devils to want to go public. The
most famous person until Princess Diana to have been al-
leged to have this disorder was Marilyn Monroe. Her life has
been the subject of many biographies, all of which capture
her desperation, her impulsivity, and her series of clinging,
demanding, ultimately failed relationships, culminating in
suicide. Borderline personality has been inferred in others,
such as Sylvia Plath and Judy Garland.
Should such a famous person go public with the illness,
he or she could at best evoke a public outcry for better rec-
ognition, less stigma, more research, and better clinical ser-
vices. Princess Diana’s desperate life evoked sympathy. It
was only after her death that her biographer identified her
as having BPD and chronicled how her public persona re-
flected her inner emptiness and her lack of access to good
treatment (despite wealth and position).
As welcome as the coming out of a much-beloved
spokesperson would be, it seems more likely that a famous
person with BPD would surface because of venting the

negative side of this personality—for example, through un-
reasonable demands or violence. This behavior might of
course evoke a counterproductive public attitude toward
the community of borderline patients that the celebrity
could come to represent.
The ability of famous individuals to go public effec-
tively as an advocate for both more compassionate atti-
tudes and better clinical services probably requires that
they have already benefited significantly from treatment.
There is increasing reason to believe that better-quality
care will yield a new generation of borderline patients who
privately, if not publicly, will become advocates for them-
selves. One of them also may become famous.
The Borderline Personality Disorder Research Foundation

In 1999, a remarkable gift was given by an anonymous Swiss donor. The
Borderline Personality Disorder Research Foundation (BPDRF) was es-
tablished to support research on BPD. The foundation initially funded
four sites that are using multidisciplinary, highly sophisticated methods
to examine the neurobiology, genetics, and psychotherapeutics of BPD.
The foundation subsequently supported a refereed competition for re-
search funds that generated numerous applications. The very significant
study of transference-focused psychotherapy (Chapter 12) and the pre-
liminary study conducted at McLean Hospital with a revised version of
manual-assisted cognitive treatment (Chapter 11) are both products of
this foundation’s support. As a side effect, this initiative has engaged
world-class scientists who otherwise knew little about BPD in stimulating
and guiding the research initiative and in legitimizing the field. An im-
mediate dividend of the financial and intellectual strength of this initia-
tive has been the heightened awareness by the National Institute of
Mental Health, pharmaceutical companies, and other foundations of the
need for more research attention to this extremely needy and underrep-
resented patient population. A longer-term effect of the BPDRF is that it
provided the funding base for a group of talented and much-needed ju-
nior investigators whose contributions to BPD are just beginning to be
seen. Regrettably, at this point it is unclear whether the BPDRF will con-
tinue to serve the pivotal role that it has had.
Summary
Since the 1970s, an industrious and diverse group of clinicians and scien-
tists have created a space for BPD in the minds of the mental health com-
munity. The results of these efforts are evident in the greatly expanded
body of knowledge about this disorder. In this book, I have reviewed the
advances in treatment of BPD, and it makes clear that we already know
enough to significantly improve the prognosis of these patients. Knowl-
edge will continue to grow. The more immediate task is to implement
and disseminate what is already apparent. For those tasks, this book is in-
tended to provide a template.
The space created for the borderline diagnosis exists on another,
more abstract level. For the mental health field, it provides a needed asy-
lum for creative theory building and research. This disorder has thus far
warded off conceptual reductionism or the constraints of entrenched
standards of care. BPD is not the intellectual or clinical property of psy-
choanalysts, of psychologists or psychiatrists, of researchers or clinicians,
or of theoreticians or practitioners. All are part “owners” who remain vi-
tally necessary contributors. This book is intended to communicate the
excitement and challenge of being a part of this community.
In the years ahead, a different task awaits. It is of critical importance to
the welfare of these patients that their tragedies—and their potential for
change—enter the collective mind of the larger society of which the men-
tal health community is only a small part. Successfully attaining this much
higher level of collective consciousness will be assisted by initiatives from
those who are prepared to become public advocates. Attaining this goal
ultimately rests on the still unquenched and seemingly inexhaustible ap-
peal for rescue that, to their credit, remains the public marker for these
patients.
References
Adler G, Buie DH Jr: Aloneness and borderline psychopathology: the
possible relevance of child development issues. Int J Psychoanal
60:83–96, 1979
46, 1981
158(suppl):1–52, 2001
erations in the treatment of a borderline patient. Int J Psychoanal 72
(Pt 4):639–656, 1991
Fonagy P, Target M, Gergely G: Attachment and borderline personality

disorder: a theory and some psychiatric evidence. Psychiatr Clin
North Am 23:103–122, 2000
Ginther C: The Web as resource for patient advocacy groups. Psychiatric
Times 14(11), November 1997
Gunderson JG: Building structure for the borderline construct. Acta Psy-
chiatr Scand Suppl 379:12–18, 1994
758, 1996
Gunderson JG: Personality disorders, in The New Harvard Guide to Psy-
chiatry. Edited by Nicholi AM. Cambridge, MA, Harvard University
Press, 1999, pp 337–357
Gunderson JG: Borderline Personality Disorder: A Clinical Guide. Wash-
ington, DC, American Psychiatric Publishing, 2001
Gunderson JG, Phillips KA: Personality disorders, in Comprehensive
Textbook of Psychiatry, 6th Edition. Edited by Kaplan HI, Sadock BJ.
Baltimore, MD, Williams & Wilkins, 1995, pp 1425–1441
Herman J: Trauma and Recovery. New York, Basic Books, 1992
Herpetz SC, Kunert HJ, Schwenger UB, et al: Affective responsiveness in
borderline personality disorder: a psychophysiological approach.
Am J Psychiatry 156:1550–1556, 1999
ISSPD Newsletter, Vol 3, No 1, December 1998
Kaysen S: Girl, Interrupted. New York, Random House, 1993
Kendler KS, Greenspan RJ: The nature of genetic influences on behavior:
lessons from “simpler” organisms. Am J Psychiatry 163:1683–1694,
2006
15:641–685, 1967
Klein D: Psychopharmacology and the borderline patient, in Borderline
1975, pp 75–92
Krueger RF, Skodol AE, Livesley WJ, et al: Synthesizing dimensional and
categorical approaches to personality disorders: refining the re-
search agenda for DSM-V Axis II. Int J Methods Psychiatr Res 16
(suppl 1):S65–S73, 2007
Links PS, Heslegrave R, van Reekum R: Impulsivity: core aspect of bor-
derline personality disorder. J Personal Disord 12:1–9, 1999
Livesley WJ, Jang KL, Jackson DN, et al: Genetic and environmental con-
tributions to dimensions of personality disorder. Am J Psychiatry
150:1826–1831, 1993
Lyons-Ruth K: Implicit relational knowing: its role in development and
psychoanalytic treatment. Infant Ment Health J 19:282–289, 1998
Mahler M, Kaplan L: Developmental aspects in the assessment of narcis-

sistic and so-called borderline personalities, in Borderline Personal-
ity Disorders: The Concept, the Syndrome, the Patient. Edited by
Hartocollis P. New York, International Universities Press, 1977, pp
71–86
Silk K: Email, transitional relatedness, and borderline patients. Paper
presented at the 5th International Congress on the Disorders of Per-
sonality, Vancouver, BC, Canada, June 1997
Skodol AE, Gunderson JG, McGlashan TH, et al: Functional impairment
in patients with schizotypal, borderline, avoidant, or obsessive-
compulsive personality disorder. Am J Psychiatry 159:276–283, 2002
Smith SB: Diana in Search of Herself: Portrait of a Troubled Princess.
New York, Times Books/Crown Publishing, 1999
North Am 2:577–594, 1979
Torgersen S, Lygren S, Øien PA, et al: A twin study of personality disor-
ders. Compr Psychiatry 41:416–425, 2000
Walker EA, Unutzer J, Rutter C, et al: Costs of health care use by women
HMO members with a history of childhood abuse and neglect. Arch
Appendix
PSYCHOEDUCATIONAL
RESOURCES: PRINTED MATERIALS,
VIDEOS, FILMS, AND WEB SITES
Printed Materials
Overviews
“Borderline Personality Disorder.” Journal of the California Alliance for the
Mentally Ill, Vol. 8, No. 1, 1997.
Wide-ranging and very readable comments from experts, families,
and persons with borderline personality disorder (BPD).
Borderline Personality Disorder Demystified: An Essential Guide for Understand-
ing and Living With BPD, by R.O. Friedel. New York, Avalon Publish-
ing Group, 2004.
An experienced and compassionate psychiatrist shares his wisdom.
Borderline Personality Disorder: What You Need to Know, by J. Gunderson. Bel-
mont, MA, McLean Hospital, 2002.
A concise, informative summary.
I Hate You, Don’t Leave Me: Understanding the Borderline Personality, by J.J.
Kreisman and H. Straus. New York, Avon, 1991.
Readable and instructive; the first book for lay people.
Imbroglio, by J. Cauwels. New York, W.W. Norton, 1992.
Scholarly and understandable; a bit dated.
Life at the Border: Understanding and Recovering From the Borderline Personality
Disorder, by L. M. Heller. Okeechobee, FL, Dyslimbia Press, 1999.
A biological perspective; overestimates the role of medications.
New Hope for People With Borderline Personality Disorder, by N.R. Bockian, V.
Porr, and N. E. Villagran. Roseville, CA, Prima Publishing, 2002.
A readable, informative book incorporating the new knowledge
about improved treatments and improved prognosis.
329
“The Pain of Being Borderline,” by M. Zanarini, F. Frankenburg, J. Gunder-

son, and others. Harvard Review of Psychiatry, Vol. 6 (Nov.–Dec. 1998),
pp. 201–207.
A vivid description of negative thoughts and feelings reported by
individuals with BPD.
Understanding and Treating Borderline Personality Disorder: A Guide for Profes-
sionals and Families, by J. G. Gunderson and P.D. Hoffman. Washing-
ton, DC, American Psychiatric Publishing, 2005.
A readable, up-to-date set of overview chapters prepared by experts,
patients, and families.
Family Issues
The Family Crucible, by A. Y. Napier and C. Whitaker. New York, Bantam
Books, 1978.
How one family member can bring covert family issues to light.
“Family Guidelines,” by J. G. Gunderson and C. Berkowitz. Belmont, MA,
New England Personality Disorder Association (NEPDA), McLean
Hospital, 2002.
Concise directions on how to improve the family environment.
Siren’s Dance: My Marriage to a Borderline: A Case Study, by A. Walker. Em-
maus, PA, Rodale, 2003.
A readable and dramatic account of a spouse’s experience.
Sometimes I Act Crazy: Living With Borderline Personality Disorder, by J.J. Kreis-
man and H. Straus. Hoboken, NJ, Wiley, 2004.
The readable sequel to I Hate You, Don’t Leave Me: Understanding the
Borderline Personality is directed to families.
Stop Walking on Eggshells: Taking Your Life Back When Someone You Care About
Has Borderline Personality Disorder, by R. Kreger and P. T. Mason. Oak-
land, CA, New Harbinger Publications, 1998.
An encouraging manual for skill building.
Instructive Books
A Bright Red Scream: Self-Mutilation and the Language of Pain, by M. Strong.
New York, Penguin Books, 1998.
Diana in Search of Herself: Portrait of a Troubled Princess, by S.B. Smith. New
York, Times Books/Crown Publishing, 1999.
Readable, insightful glimpse of the distinction between public per-
sona and internal strife.
Eclipses: Behind the Borderline Personality Disorder, by M.F. Thornton. Madi-
son, AL, Monte Sano Publishing, 1997.
Appendix: Psychoeducational Resources ❘ 331
Effective Psychotherapy With Borderline Patients, by R. J. Waldinger and J. G.

Gunderson. Washington, DC, American Psychiatric Press, 1989.
Case reports of successful long-term therapies.
Girl, Interrupted, by S. Kaysen. New York, Random House, 1993.
A best-seller. Highlights hospital experiences, treatment impasses,
black-and-white thinking.
I’m Not Supposed to Be Here: My Recovery From Borderline Personality Disorder,
by R. Reiland. Milwaukee, WI, Eggshell Press, 2002.
Lost in the Mirror: An Inside Look at Borderline Personality Disorder, 2nd Edi-
tion, by R. Moskovitz. Dallas, TX, Taylor Publications, 2001.
Borderline patients can recognize themselves. Vivid and compas-
sionate.
Marilyn: A Biography, by N. Mailer. New York, Grosset and Dunlap, 1973.
History’s most celebrated exemplar of BPD.
Search for the Real Self, by J.F. Masterson. New York, Free Press, 1990.
Readable psychodynamic approach.
Starry Starry Night: Life and Psychiatric History of Vincent van Gogh, by E. van
Meekeren. Amsterdam, The Netherlands, Benecke NI, 2003.
A careful review of the artist’s interpersonal and psychiatric history.
Newsletters
TARA Times. From TARA, 23 Greene St., New York, NY 10013.
Excellent accounts of public health costs and of advocacy initia-
tives and opportunities.
Videos
“Back From the Edge.” BPD Resource Center. Cambridge, MA, Lichten-
stein Creative Media, 2007.
Vivid, articulate, evocative first-person accounts.
“Beyond the Borderline.” Western Psychological Association. Albany, NY,
Olive Tree Productions, 1998.
“Borderline Syndrome: A Personality Disorder of Our Time.” Narrated by
Maureen Stapleton. Albany, NY, Olive Tree Productions, 1988.
An instructive introduction.
Films
Bliss (with Craig Sheffer and Sheryl Lee). Triumph Films/Stewart Pic-
tures, 1997.
A poignant look at a lost soul with a history of childhood sexual abuse.
Fatal Attraction (with Glenn Close). Paramount Pictures, 1987.
Frightening portrait of abandonment rage.
Girl, Interrupted (with Winona Ryder and Angelina Jolie). Columbia Pic-
tures, 1999.
More vivid than the book (see earlier book listing).
Lethal Weapon (with Mel Gibson). Warner Bros., 1987.
Captures identity disturbance.
Looking for Mr. Goodbar (with Diane Keaton). Paramount Pictures, 1977.
Captures emptiness, thrill seeking, and good/bad split self.
Play Misty for Me (with Clint Eastwood). Universal Studios, 1971.
Torment by others who resist being possessed.
Taxi Driver (with Robert De Niro and Jodie Foster). Columbia Pictures,
1976.
Web Sites
(prepared by Maria Daversa, Ph.D., and Marc Walter, M.D.)
BPD Central
http://www.bpdcentral.com
Information—organized by Randi Kreger; provides basic information
about BPD; especially useful for consumers, partners, and par-
ents
Links—to national and international organizations, research and
treatment, newsgroups, legal help, and regional support groups
Referral source—Find a Therapist, libraries, hiring an attorney, and tele-
phone support groups
Lists—many videos, books, articles with a bias toward work by Randi
Kreger, the author of the site
BPD Sanctuary
http://www.mhsanctuary.com/borderline
Information—supportive and hopeful testimony by and for borderline
patients; provides many links to BPD communities (chat rooms,
bulletin boards, blogs, open forums); information for consumers
and clinicians; also has Ask the Therapist and Ask the Experts sec-
tions
Links—for clinicians, consumers, a family section with a family chat
room, bulletin boards, and resources specific for families
Referral source—provides a list of doctors and therapists by state who
treat BPD, a link to 1-800-Therapist, and toll-free resources and
hotlines
Lists—extensive bookstore separated by topics related to BPD and
general mental health, plus a list of 75 articles
BPD today
http://www.borderlinepersonalitytoday.com
Information—provides information about BPD for consumers, fami-
lies of consumers, and clinicians
Links—bookstore; library; discussions for families; questions and an-
swers with experts; and resources for clinicians, families, consum-
ers, spiritual support, and volunteers
Referral source—provides a link to 1-800-Therapist
Lists—many articles and books related to BPD, families, children,
medications, child abuse, and self-harm
BPD World
http://www.bpdworld.org
Registered charity for people with BPD in the United Kingdom
(founded and created by Joshua Cole in March 2003; Registered
Charity Number: 1111750)
Information—provides information about BPD, relationships, and
treatment options
Links—information about BPD, including theories and causes; gen-
eral mental health, child abuse, depression, eating disorders, and
self-harm
Referral source—free counseling service, advocacy, and telephone sup-
port is provided by centers in United Kingdom
Lists—has selection of free printed material on BPD, cognitive-behav-
ioral therapy, dialectical behavior therapy, self-harm, and crisis in-
formation; however, available only in United Kingdom
DBT Psychotherapy Network

http://www.dbtnetwork.com
Information—related to only the dialectical behavior therapy program
in New Orleans, LA. Explains BPD and cognitive-behavioral ther-
apy and dialectical behavior therapy treatment modalities.
Links—none
Referral source—none
Lists—none
Helen’s World of BPD Resources

http://www.bpdresources.com
Information—provides general information about BPD, information
for families, BPD-specific treatment modalities, and articles
Links—minimal links to sites about general mental health and co-
occurring disorders
Lists—minimal number of articles and books related to coexisting dis-
orders, BPD in children and adolescents, dialectical behavior
therapy, BPD, families, medications, psychotherapies, and fea-
tures of BPD
Mental Help Net

http://www.mentalhelp.net
Information—well organized and frequently updated. Can find BPD
under the topic “Personality Disorders”; limited to DSM criteria
and selective overview of treatment (i.e., psychotherapy, hospital-
ization, medications).
Links—questions and answers, Weblog entries
Referral source—a “Find/Locate” menu is helpful to locate therapists,
self-help groups, on-line counseling, telephone numbers, and
medication information
Lists—provides a “Read and Listen” menu of essays, e-books, current
news, book reviews, podcasts, magazine articles, and professional
interviews
National Alliance on Mental Illness (NAMI)

http://www.nami.org
Information—can find BPD under the topic “Inform Yourself” (click on
“Other Illnesses”); provides information about DSM criteria, eti-
ology, co-occurring disorders, treatment, self-harm, medications
Links—information on treatment modalities for BPD, “A BPD Brief”
(Gunderson 2006), substance use disorders, self-injury, family
connection reading list, mental illness discussion groups; also Na-
tional Education Alliance for Borderline Personality Disorder
and National Institute of Mental Health Web sites
Referral source—for state and local affiliates that provide support, edu-
cation, information, referral, and advocacy
Lists—provides a library link but recommends only nine books
Palace net
http://www.palace.net/~llama/psych/bpd.html
Information—limited to several major researchers, clinicians, and
DSM
Links—none
Referral source—list of treatment programs from the National Institute
of Mental Health
Lists—none
Souls Self Help Central

http://www.soulselfhelp.on.ca
Information—supportive and hopeful, provides a definition of BPD
and information related to abandonment, etiology, attachment
issues, identity, and relational styles. Also provides extensive in-
formation about dialectical behavior therapy and on-line dialec-
tical behavior therapy skills training classes.
Links—e-mail lists for BPD, dialectical behavior therapy, and BPD for
children and adolescents, codependency, and self-harm behavior
Lists—mostly articles and books written by A.J. Mahari, the author of
the site
Web 4 Health
http://www.web4health.info
Information—a primarily ICD-10 view of BPD, can submit questions
online for responses by BPD experts (from Europe), uses a ques-
tion-and-answer format with answers taken from peer-reviewed
articles; is difficult to navigate and hard to find topics
Links—to sites about general health, mental health, hotlines, library
and research, blogs, psychological testing, and mental health
search engines
Referral source—European BPD experts provide free “Ask an Expert”
online services
Lists—none
INDEX
Page numbers printed in boldface type refer to tables or figures.
Page numbers followed by an s refer to sidebars.
Page numbers followed by a v refer to vignettes.
Abandonment fears, in DSM-IV American Psychiatric Association. See

criteria, 11, 12 also Diagnostic and Statistical
ABERG (Acceptance-Based Emotion Manual of Mental Disorders
Regulation Group Intervention), liability guidelines, 147s
209, 214 practice guidelines, 263, 316, 317–
skills training modules, 210–211 318
ACT (Assertive community couples therapy, 198
treatment), 130 pharmacotherapy, 143, 158,
Actualization, 68, 69, 71, 76, 280 159, 165
Adler, G., 78, 83s, 302s Amitriptyline, 150v
debate with Kernberg, 280s–281s, Anger, 11, 14, 272
281, 282 impulsivity and, 14–15, 44, 46,
intolerance of aloneness, 12, 18 159v–160v, 159–162, 162
attachment failures, 19s–20s, Anorexia, co-occurrence with
321 borderline personality disorder,
treatment approach, 65, 121s, 259, 40, 41, 49
292, 294 Antidepressants, for management of
Adolescents affective dyscontrol/depression,
age at onset, 10 164
diagnosis in, 27 anger/impulsivity, 162
liability example, 93s anxiety, 169
Adoption studies, 4–5 cognitive perceptual dyscontrol,
Advocacy, 321–323, 323s–324s 167
Affect, 7, 76, 77–78. See also Emotions Antipsychotics, 139–140, 159, 163
Affective dyscontrol/instability, 11, for management of
13–14, 162v–163v, 162–165, 164 affective dyscontrol/
Age, at onset, 10 depression, 163, 164
Aggression, goal to reduce, 73 anger/impulsivity, 161, 162
Alcohol abuse, 93s, 145v anxiety, 168, 169
co-occurrence with borderline cognitive-perceptual
personality disorder, 40, 41 dyscontrol, 166–167,
Aloneness, 18s–20s, 18–22, 21 167
Alprazolam, 160 suicide risk, 97, 97
337
Antisocial personality disorder, 38, 39 Bipolar II disorder, 38, 39

borderline personality disorder comparison with borderline
and, 56–59, 57, 58v personality disorder, 43, 44,
co-occurrence with borderline 45
personality disorder, 40, 41 co-occurrence with borderline
overlapping characteristics personality disorder, 40, 41
between borderline Bipolar spectrum, borderline
personality disorder and, 56, personality disorder and, 43–47,
57 44v–45v, 45, 46s–47s
Anxiety, 167–169, 168v Borderline personality disorder
pharmacotherapy, 169 (BPD)
Anxious/ambivalent attachment, age at onset, 10
19s aloneness, 18s–20s, 18–22, 21
Aripiprazole, for management of antisocial personality disorder
cognitive-perceptual dyscontrol, and, 56–59, 57, 58v
165, 167 attachment issues, 20, 20s, 21
Assertive community treatment bipolar II disorder, 39, 41, 43, 44,
(ACT), 130 45
Attachment, 20, 20s, 21 bipolar spectrum and, 43–47, 44v–
secure, 290, 304–305 45v, 44, 45, 46s–47s
Avoidant personality disorder, 38, clinical theories, 321
39 cognitive-behavioral therapies,
Axis I disorder, 38, 153 253–278
Axis II disorder, 23–24 cognitive-perceptual symptoms,
17
Bean-Bayog, Margaret, 302s–303s concepts, 3, 4
Beck, A.T., 254, 267–268, 270, 273 definition, 3
Behavior demographics, 10
anxious/ambivalent, 19s depression and, 40–43, 41, 42
change, 74, 75, 76 development of the borderline
disorganized, 19s–20s construct, 2, 5–9, 7, 8
quality-of-life-improving, 259 diagnosis, 1–36
therapy-interfering, 259 in adolescents, 27
Benzodiazepines, for management of explanation of, 27–28
affective dyscontrol/depression, misuses of, 22–24, 23s
164 origins of, 1–5, 2, 2s–3s,
anger/impulsivity, 162 4
anxiety, 169 resistance to, 187–188
cognitive-perceptual dyscontrol, differential diagnosis, 37–64
167 changing construct, 38,
Bills, 232v, 232–233 40
Bipolar I disorder, co-occurrence comorbidities, 40–59, 41
with borderline personality overall function, 37–38
disorder, 40, 41 DSM-IV criteria, 9–17, 11
Index ❘ 339
eating disorders and, 49–51 resources, 329–335

epidemiology, 9, 10 somatoform/somatization
experience of, 13s disorders and, 52–53
family interventions and split treatment, 107–110
therapies, 177–205 substance abuse and, 51–52, 51v–
future considerations, 315–327 52v
diagnostic implications, 318– treatment. See also specific therapies
321, 320, 321 affects or emotions, 77–78
DSM-V, 318, 320 countertransference, 84–85
public awareness and advocacy, goals, 72–77, 74s–76s, 75
321–323, 323s–324s historical overview, 65–67
research, 324 impulse/action pattern,
standards of care, 317–318, 78–79
319 initial structuring, 80–82, 82
treatment implications, 315– multiple modalities and step-
317, 317 down services, 69, 71–72
group therapy, 207–223 overview, 65–88
guidelines for relationship with treaters, 76,
psychopharmacological 80
treatment, 144 resistance to, 188
as iatrogenic disorder, 15s–16s social function impairment,
implementation of changes in 79–80
treatment, 105–106 sociotherapies, 72, 73
individual psychotherapies, 225– step-down services, 69
252 therapeutic alliance, 82–84, 83,
levels of care, 113–137 83s–84s
narcissistic personality disorder therapeutic processes, 67–71,
and, 53s–55s, 53–56, 54, 68, 69
55v transition from treatment to
from organization to syndrome to therapy, 188–189
disorder, 5–9, 7, 8 World Health Organization
pharmacotherapy adaptation, 1
clinical practices, 139–155 Borderline Personality Disorder
patient attitudes, 149 Research Foundation (BPDRF),
selection of medications, 157– 324
175 Borderline personality organization,
treatment guidelines, 144 2, 3, 4, 5
posttraumatic stress disorder and, The Borderline Syndrome (Grinker
47–49, 48v, 49v et al.), 4
prevalence, 10 Boundaries, 106–107, 108
psychodynamic psychotherapies, therapeutic frame, 231–236
279–313 therapist, 231, 241
regression, 302s–304s Bowlby, J., 18s–20s
research studies and clinical BPD. See Borderline personality
reports, 8 disorder
BPDRF (Borderline Personality operant conditioning

Disorder Research Foundation), applications, 255–256
324 overview, 253–255
Bulimia, 40, 41, 49–50 schema-focused therapy,
268–272
Carbamazepine, 160 Cognitive-perceptual dyscontrol,
for management of 165–167
affective dyscontrol/ pharmacotherapy, 167
depression, 164 symptoms, 16–17, 17, 166v
anger/impulsivity, 162 Cognitive therapies. See Cognitive-
anxiety, 169 behavioral therapies
cognitive-perceptual Communication, goals for families,
dyscontrol, 167 73
depression, 78 Community meetings,
Caregivers 126–127
abandonment, 124s Confidentiality, 93s
attachment to, 18, 19s–20s, 186s, outside-of-group contact, 219, 220v
318 progress reports, 75s, 77s
transitional objects, 58, 298s Containment, 67, 68, 69, 76. See also
patients as, 125s Holding and holding
psychopharmacologist as, 145, environment
148, 152 Contingency clarification,
resistance, in family therapy, 181– 255–256
183 Contracting
spouse as, 198 roles, 83, 228–231, 229
therapist as, 38, 95, 101s safety, 99, 99s–101s
Case administrator. See Primary Contractual (behavioral) alliance,
clinician 83, 90–91, 91, 229, 290
Case management, 89–112, 234. level III care, 122
See also Primary clinician level IV care, 120
in intensive outpatient care, 132– Countertransference, 84–85, 110, 248
133 diagnosis and, 16s, 23s, 28, 38, 59–
Changes, expected in treatment, 72– 60
80, 75, 76 families and, 180–181
Chronic pain syndrome, 168v group therapies and, 131, 218
Citalopram, 143v–144v, 165 psychopharmacologists and,
Cognition, 7 149–151, 150
Cognitive-behavioral therapies, 253– Couples therapy. See Marital therapy
278. See also Dialectical behavior Crises
therapy; Systems Training for guidelines for management of
Emotional Predictability and safety, 100
Problem Solving interventions, 120
manual-assisted cognitive Cutting, 26s. See also Self-injurious
treatment, 273–274 behavior
Index ❘ 341
Day treatment, 114, 116, 127–130, Diana, Princess, 323, 323s–324s

128s–129s Diana in Search of Herself: Portrait of a
DBSA (Depression and Bipolar Troubled Princess (Smith), 323
Support Alliance), 322 DIB (Diagnostic Interview for
DBT. See Dialectical behavior therapy Borderline Patients), 5
Death. See also Suicidal behavior Disorganized/disoriented
risk of, 104 attachment, 19s–20s
pharmacotherapy and, 150v Disturbed relationships, in DSM-IV
Defensiveness, 73 criteria, 10–13, 13s
Depakote (divalproex), 150v Divalproex (Depakote), 150v
Depersonalization, 11, 17 DSM. See Diagnostic and Statistical
Depression, 78, 163v, 295v Manual of Mental Disorders
borderline personality disorder Dysthymia, co-occurrence with
and, 40–43, 41, 42 borderline personality disorder,
Depression and Bipolar Support 40, 41
Alliance (DBSA), 322
Derealization, cognitive-perceptual Eating disorders, 40, 41, 49–51
symptom, 11, 17 Education. See Psychoeducation
Detached protector, 272 Emotions
Diagnostic Interview for Borderline expected changes, in treatment,
Patients (DIB), 5, 12, 18, 284 76, 77–78
Diagnostic and Statistical Manual of regulation, 257
Mental Disorders (DSM) skills training modules, 210
DSM-III, 1, 5, 6, 9–17, 11 Empathy, 292, 293, 294. See also
DSM-IV, 9–17, 11 Relational (empathic) alliance;
DSM-IV-TR, 27 Validation
DSM-V, 318, 320 borderline personality disorder,
Dialectical behavior therapy (DBT), and treatment
78, 89, 228, 256–267. See also versus bipolar disorder, 45
Linehan, M.M. versus narcissistic personality
applications, 262–263 disorder, 56
basic services, 257–261, 259, 259v– dialectical behavior therapy, 260
261v family therapy, 186
empirical support, 261–262 interpersonal group therapy,
individual versus group 215s
components, 263–265, 264v– Kohut/Adler approach, 282
265v therapeutic
limitations, 266 goals, 229
overview, 266–267 stance, 106, 107, 121s, 246
skills groups, 210–211, 212 therapist
theory, 257 functions, 68, 290
training standards, 247, 249, 262 qualities, 240s, 242
versus transference-focused Emptiness, feelings of
psychotherapy, 288–289 in DSM-IV criteria, 11, 12–13
Environment, 28, 96, 321. See also Fluoxetine (Prozac), 141s, 150, 150v,
Holding and holding 161
environment Fonagy, P., 15s, 18s, 20s, 236, 319,
antisocial personality disorder, 321. See also Mentalization;
58 Mentalization-based therapy
family, 79, 118s, 179, 186, 189, 192 empathy, 292, 293
invalidating, 257, 258, 287s family therapy, 189, 191, 194
narcissistic personality disorder, Frosch, John, 16
56
Gabapentin, for management of
Families. See also Group therapy; anxiety, 168
Psychoeducation General psychiatric management,
goals, in sociotherapies, 73 263
interventions and therapies, 177– Girl, Interrupted (Kaysen), 65–66, 323
205 Grinker, Roy, 2, 2s–3s, 3–4, 4
establishing an alliance, 187– Group meetings, level III care, 126–
189 127
history, 178–180, 180 Group therapy, 127, 207–223. See also
initial family meetings, 183– Families
187, 184s, 186s–187s common problems, 218–221
marital or couples therapy, disruptive group behaviors,
197–201, 198v–199v, 200v 218–219, 218v–219v
psychoeducational family outside-of-group contacts,
therapy, 189, 189–196, 219v–220v, 219–221
191s, 192–193, 193s, 195s– silence, 221, 221v
196s dialectical behavior therapy,
resistance, 181v–182v, 181–183, component of, 258
182s–183s engaging patient in, 208–209
therapists and goals, in sociotherapies, 73
countertransferences, group structure, 216–218
180–181 duration, 216–217
multiple-family group, 190 leadership, 217–218
problem identification, 184–185 length, 216
punitive parents, 272 size, 216
resources for, 330 indications, 207–208
single-family interventions, 189– psychodynamic group therapies,
190 214–216
Family therapy, 189, 189–196, 191s, interpersonal group
192–193, 193s, 195s–196s. See also psychotherapy, 214–215,
Psychoeducation 215s
early stage, 191, 191s, 192–193, mentalization-based group
193s therapy, 215–216
late state, 195–196, 195s–196s skills training groups, 209–210,
middle stage, 194–195 209–214
Films, for psychoeducation, 331–332 Guilt, in families, 186s–187s
Index ❘ 343
Histrionic personality disorder, 38, 39 therapeutic frame, 231–236

Holding and holding environment, bills, 232v, 232–233
18s–19s, 20, 21, 21, 77s, 77–78 frequency, 233–234, 234, 235
containment, 67, 68, 69, 76 scheduling, 234–235
family therapy, 181, 201 seating, 235–236
frequency of sessions, 234, 264, 303s therapists, 236–243
interpersonal group, 214, 216 qualifications, 236–241, 237v–
level III care, 122, 126, 129–130 238v, 238, 239, 240s–241s
structure as, 70, 100s qualities, 241–243, 242s–243s,
therapist as, 299, 233 244
Hopelessness, 78 training, 249
Hospital (level IV) care, 68, 69, 114, Intensive outpatient (level II) care,
115–122, 119v 68, 69, 114, 130–133, 131v
changes in prior therapies, 120 assertive community treatment,
goals and components, 116–117, 130
119–120 goals and components, 116–117,
long-term, 115s, 118s–119s 131–133, 132
medical, neurological, and self-assessment groups, 131–132
psychological evaluations, 120 International Society for the Study of
psychotherapeutic technique, 121s Personality Disorders (ISSPD),
staff, 121–122 316
structure, 120–121 Interpersonal group psychotherapy,
treatment plan, 120 214–215, 215s. See also Group
therapy; Psychodynamic
Iatrogenic, borderline personality psychotherapies
disorder as, 15s–16s dialectical behavior therapy,
Ideas of reference, 17 contrast to, 258
Identity disturbance, 16 duration, 216–217
Illusions, cognitive-perceptual level II care, 132, 133
symptom, 11, 17 Interventions
Impasses, in psychotherapy, 305–307, multiple-family group, 190
306 single-family, 189–190
Impulsivity, 11, 14–15, 44, 46 Invalidation, 257
in borderline construct, 6, 7 Involvement, 68, 69, 70–71, 76
change in, 78–79 ISSPD (International Society for the
Individual psychotherapies, 225–252. Study of Personality Disorders),
See also specific psychotherapies 316
contracting roles, 228–231, 229
definition, 225 Kernberg, Otto, 38, 57, 215s, 300. See
dropouts, 227–228 also Transference-focused
goals, 235 psychotherapy
prerequisites, 225–227 anger and aggression, 14, 304, 321
qualities of effectiveness, 247–248, borderline personality
248, 249 organization, 2, 3, 4, 22
readiness, 226 boundaries, 294
Kernberg, Otto (continued) Linehan, M.M., 14, 229, 293. See also
debate Dialectical behavior therapy
with Kohut/Adler, 280s–281s, debate with Kernberg, 286s–287s
281–282, 282 skills training groups, 211, 212
with Linehan, 286s–287s split treatment, 109–110
identity disturbance, 16 therapist
Menninger Psychotherapy credentials, 240s–241s
Research Project, 282–283, qualities, 241, 242s–243s,
284, 293 245
splitting, 12 Listening skills, goal of
containment of, 121s sociotherapies, 73
therapist Lithium, 150v, 160
credentials, 240s–241s Little, Margaret, 302s, 303s
qualities, 242s–243s Living skills, level III goal, 123
therapy Losing and Fusing (Lewin and Schulz),
contracting for, 99s 71
prerequisites for, 227
Klonopin, 145 Mack, John, 24
Kohut, Heinz, 78, 280s–281s, 281, MACT (Manual-assisted cognitive
282 treatment), 273–274
Magical thinking, cognitive-
Lamotrigine, 160 perceptual symptom, 17
for management of Major depressive disorder, 38, 39, 40–
affective dyscontrol/ 43, 41, 42
depression, 164 group therapy, 208
anger/impulsivity, 162 Manners, goal of sociotherapies,
anxiety, 169 73
cognitive-perceptual Manual-assisted cognitive treatment
dyscontrol, 167 (MACT), 273–274
Leadership, of group therapy, 217– MAOIs. See Monoamine oxidase
218 inhibitors
Learned associations, 255 Marital therapy, 184s, 197–201
Levels of care, 68, 69, 113–137, 116 acting-out couples, 198v–199v
family interventions, 177 caregiving spouse-type couples,
functions, 114 200, 200v
level I (outpatient), 133–134 mutual projective-type couples,
level II (intensive outpatient), 199–200, 200v
130–133 therapeutic processes, 201
level III (residential/partial Masterson, J.F., 12, 18, 279, 281s, 319,
hospital/day treatment), 321
122–130 family therapy, 178, 179, 197
level IV (hospital), 115–122 McLean Hospital, 227–228, 273–274,
selection or change of, 114–115, 283, 284, 285, 316, 324, 317
116 Medications. See Pharmacotherapy
Liability, 92, 93s, 147s Men, in group therapy, 208
Index ❘ 345
Menninger Psychotherapy Research Narcissistic personality disorder, 38,

Project, 107, 282–283, 284, 39
293 borderline personality disorder
Mentalization, 20s. See also Fonagy, P. and, 53s–55s, 53–56, 54, 55v
attributions exercise, 194 co-occurrence with borderline
positive dependency, 299 personality disorder, 40, 41
relational alliance, 292–293 National Alliance on Mental Illness
transference-focused therapy, (NAMI), 322, 334
289 National Depressive and Manic-
Mentalization-based therapy, 128s– Depressive Association, 322
129s, 247, 248, 249, 281s. See also National Education Alliance for
Fonagy, P. Borderline Personality Disorder
group therapy, 215–216 (NEA-BPD), 322–323
Modell, A., 18, 19s National Institute of Mental Health
Models (NIMH), 5, 178
borderline personality disorder, NEA-BPD. See National Education
179 Alliance for Borderline
pharmacotherapy Personality Disorder
diagnostically-focused, 158 NEPDA (New England Personality
subsyndromal, 158 Disorder Association), 322
symptom-targeted, 158–169 Neurotic personality organization, 3,
Monoamine oxidase inhibitors 4
(MAOIs), 163–165 New England Personality Disorder
for management of Association (NEPDA), 322
affective dyscontrol/ Night care, 114, 116, 122–130
depression, 164 NIMH (National Institute of Mental
anger/impulsivity, 162 Health), 5, 178
anxiety, 169
cognitive-perceptual Obesity, co-occurrence with
dyscontrol, 167 borderline personality disorder,
Mood stabilizers, 160 40, 41
for management of Outpatient (level I) care, 68, 69, 114,
affective dyscontrol/ 133–134, 134v
depression, 164 goals and components, 116–117
anger/impulsivity, 162
anxiety, 169 Paranoid experiences, cognitive-
bipolar disorder and/or perceptual symptom, 17
borderline personality Parents, “good cop”/”bad cop” roles,
disorder symptoms, 46 189, 191s, 193s. See also Families
cognitive-perceptual Paroxetine (Paxil), 152v
dyscontrol, 167 Patients
choice of therapist, 244
Naloxone, 168 therapeutic engagement, 243–246
NAMI (National Alliance on Mental Paxil (paroxetine), 152v
Illness), 322, 334 Perphenazine (Trilafon), 150v
Pharmacotherapy. See also individual Primary clinician, 80, 145v,

drug names 145–146
clinical practices boundaries, violations, and setting
attitudes, of patients, 148, 149 limits, 106–107, 108
contraindications and case management, 89–112
discontinuance, 151–153, choice of, 244
152v confidentiality, 93s
guidelines for treatment, contractual alliance, 90–91
144 countertransferences and, 180–
history, 139–140 181
psychiatrist's role, 143–147, for dialectical behavior therapy,
143v–144v, 145v, 147s 262
role of medications, 140–142, engaging patients and, 208–209
141s–142s liability issues, 92, 93s
symptom chasing, 148 “no-therapy therapy,” 94
transference- options during absence, 301
countertransference issue, patient feedback and, 307
149–151, 150v, 150 qualifications, 90, 236–241, 237v–
selection of medications, 157–175 238v, 238, 239, 240s–241s
overdose, 159v–160v proposed credentialing,
pharmacotherapy models, 157– 242
158 qualities, 241–243, 242s–243s,
symptom-targeted model, 158– 244
169 quality assurance, 240s–241s
PISA. See Psychosocial/ relationship management,
Psychoeducational Group 93–94
Intervention for People With responsibilities, 90–92, 91
Recurrent Suicide Attempts supervision, 110
Plath, Sylvia, 300 training, 242, 249
Positive dependency, phase of usual practices, 93s
therapy, 290, 298s–299s, 298– Principle of triangulation, 126
304, 301, 302s–304s Progress reports, 74s–77s
Posttraumatic stress disorder (PTSD), Prozac (fluoxetine), 141s, 150, 150v,
38, 39, 168v 161
borderline personality disorder Psychic vs. somatic anxiety,
and, 47–49, 48v, 49v 167–169
co-occurrence with borderline Psychodynamic psychotherapies,
personality disorder, 40, 41 279–313
Practice Guideline for the Treatment of group therapies, 214–216, 215s
Patients with Borderline Personality impasses, 305–307, 306
Disorder (American Psychiatric outcome studies, 282–286, 284
Association), 263, 316, 317–318 overview of change processes,
couples therapy, 198 289–291, 290
pharmacotherapy, 143, 158, 159, positive dependency, 298s–299s,
165 298–304, 301, 302s–304s
Index ❘ 347
pre-empirical developments, Residential (level III)/partial

280s–281s, 280–282, 282 hospital/day treatment, 68, 69,
relational alliance, 291, 291–298, 114, 122–130, 123v
292, 295v behavioral and interpersonal
secure attachment, 304–305 traits, 125
suicidal behavior, 78–79 goals and components, 116–117,
transference-focused 123–125, 123s–125s
psychotherapy, 286s–287s, relational alliance, 122
286–289, 288 staff, 125–126
Psychoeducation, 81–82, 82. See also structure, 126–130
Families; Family therapy day treatment, 127–128
family therapy, 179–180, 180, 185, empirical support,
189, 189–196 128–129
rationale for patients, 82 group meetings, 126–127
responsibility of primary clinician, night care, 129–130
91
skills training modules, 210–211 Safety management, 94–102
Psychosocial/Psychoeducational contracting for safety, 99,
Group Intervention for People 99s–101s
With Recurrent Suicide Attempts feeling unsafe, 98v–99v, 98–102,
(PISA), 209, 213–214 99s–101s, 101v–102v
duration, 216 guidelines, 100
skills training modules, 210–211 prevention, 97–98
Psychotic personality organization, 3, responsibility of primary clinician,
4 91
PTSD. See Posttraumatic stress suicide assessment, 94–97, 96v,
disorder 97
Public awareness, 321–323, Schema-focused therapy (SFT), 248,
323s–324s 268–272
basic practices, 271
Quality assurance, 240s–241s caveats, 271–272
Quality-of-life-improving behavior, frequency, of sessions, 233
259 modes, 245–246, 272
outcomes, 269s–270s
Reality testing, lapses in, 16–17 overview, 272
Recreational/expressive groups, 127 theory, 270–271, 272
Regression, 302s–304s training, 249
Relational (affective/empathic) transference-focused therapy,
alliance, 83, 122, 290, 291, 291– contrast to, 281s
298, 292, 295v Schizoaffective disorder, 208
Relationships Schizoid personality disorder,
disturbed, 10–13, 13s 38, 39
management, 93–94 Schizotypal personality disorder, 38,
with treaters, 76, 80 39
unstable, 10, 11, 12 Secure attachment, 290, 304–305
Selective serotonin reuptake psychodynamic psychotherapy,

inhibitors (SSRIs), 142v 290, 307
for management of psychopharmacology, 146,
affective dyscontrol/ Split treatment, 107–109, 146–147,
depression, 164 147s
anger/impulsivity, 162 SSRIs. See Selective serotonin
anxiety, 169 reuptake inhibitors
cognitive-perceptual Staff. See also Primary clinician
dyscontrol, 167 in hospital, 121–122
Self-assessment groups, level II care, in residential treatment center,
131–132 125–126
Self-awareness, goal of STAIRWAYS program, 213. See also
sociotherapies, 73 Systems Training for Emotional
Self-injurious behavior, 24–25, 25, Predictability and Problem
26s, 48v Solving
functions, 25 Standards of care, 317–318, 319
impulsivity, 78–79 STEPPS. See Systems Training for
self-mutilation, 11, 14–15, 42v, Emotional Predictability and
46s–47s, 166v Problem Solving
Sexual abuse, 9, 48, 49v, 50, 145v Stevenson, J., 121s, 259, 292, 294
SFT. See Schema-focused therapy New South Wales study, 284, 285–
Silence, 221, 221v 286
Single-family interventions, 189–190 Stewart, Martha, 53s–55s
Skills training groups, 209–214, 210– Structure, 68, 69, 70, 76, 80–82
211 hospital (level IV), 120–121
Sleep disorders, 48v intensive outpatient (level II),
Sleeping pills, 162v–163v 131–133
Social function, sphere of change, 76 residential (level III), 126–130
Social skills, goal of sociotherapies, 73 therapeutic frame, 231–236, 248
Sociotherapies, 72, 73 bills, 232–233
goals, 73 frequency, 233–234, 234, 235
Somatic anxiety, 167–169 scheduling, 234–235
Somatization, co-occurrence with seating, 235–236
borderline personality disorder, Substance abuse, 58v
40, 41 borderline personality disorder
Somatoform/somatization disorders, and, 51–52, 51v–52v
borderline personality disorder co-occurrence with borderline
and, 52–53 personality disorder, 40, 41
Splitting, 78, 80, 107–109 Suicidal behavior, 14–15, 48v, 237v–
defense, 3, 45, 45, 214, 268 238v, 259, 288
dialectical behavior therapy, 264 acute-on-chronic model, 96, 96v,
expressed in DSM-IV criteria, 10, 97
12, 14 assessment, 94–97, 96v, 97
family therapy, addressed in, 191, dialectical behavior therapy, 259
193s DSM-IV criteria, 11, 14–15
Index ❘ 349
impulsivity, 78–79 Transference-focused psychotherapy

preventive stance, 97–98 (TFP), 228, 286s–287s, 286–289,
principle of false submission, 102– 288. See also Kernberg, Otto
105 contracting for therapy, 229–231
response to recurrent, 102–105, dialectical behavior therapy,
103v, 104v–105v contrast to, 288–289
Support, 67–70, 68, 69, 76, 186, 186s– frequency, 233
187s training, 249
Symptom chasing, 148 Transitional objects, 18, 18s–19s, 58,
Systems Training for Emotional 298s–299s
Predictability and Problem medications as, 152
Solving (STEPPS), 209, therapist as, 292, 294, 298, 300,
212–213 301
duration, 216–217 Transitional relatedness, 18
skills training modules, 210–211 Trauma, 6, 7, 7, 28. See also
Posttraumatic stress disorder
TARA APD (Treatment and Research dialectical behavior therapy, 265
Advancements Association for family unavailability, 179
Personality Disorder), 322 schema-focused therapy, 271
Temperament, 14, 321 secure attachment phase, 290,
family guilt, 186s 305
overlap with other personality Treaters, relationship with, 76, 80
disorders, 38, 43, 57 Treatment and Research
TFP. See Transference-focused Advancements Association for
psychotherapy Personality Disorder (TARA
Therapeutic alliance, 82–84, 83. APD), 322
See also Relational alliance Tricyclic antidepressants, for
engaging patient in, 243–246 management of
forms of, 83 affective dyscontrol/depression,
myths about, 83s–84s 164
Therapist. See Primary clinician anger/impulsivity, 162
Tolerance of others, goal of anxiety, 169
sociotherapies, 73 cognitive-perceptual dyscontrol,
Topiramate, 160 167
for management of Trilafon (perphenazine), 150v
affective dyscontrol/
depression, 164 Unstable relationships, in DSM-IV
anger/impulsivity, 162 criteria, 10, 11, 12
anxiety, 169
cognitive-perceptual Vaillant, George, 23, 23s
dyscontrol, 167 Validation, 68, 68, 117, 226
Toronto Alexithymia Scale, 213– dialectical behavior therapy, 258–
214 259
Transference-countertransference, invalidation, 257
84–85, 149–151, 150, 150v relational alliance, 292, 293, 295v
Valproate, 160 Web sites, for psychoeducation, 332–

for management of 335
affective dyscontrol/ WHO (World Health Organization), 1
depression, 164 Winnicott, D.W., 18, 18s–19s, 58, 67,
anger/impulsivity, 162 292, 298s, 302s
anxiety, 169 Working (cognitive/motivational)
cognitive-perceptual alliance, 83, 83s, 100s, 225, 254,
dyscontrol, 167 290, 299–300, 304–307
van Gogh, Vincent, 46s–47s family therapy, 187
Videos, for psychoeducation, World Health Organization (WHO), 1
331 Worthlessness, 78
Vocational rehabilitation
intensive outpatient (level II) care, Young, Jeffrey E., 254, 293. See also
131 Schema-focused therapy
residential (level III) care, 123,
123s–125s Zanarini, M.C., 13s, 52, 140, 321

Borderline

Uploaded by

Copyright:

Available Formats

Borderline

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Borderline

Uploaded by

Copyright:

Available Formats

BORDERLINE PERSONALITY

John G. Gunderson, M.D.

About the Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi

1 The Borderline Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . 1

2 Differential Diagnosis: Overlaps, Subtleties, and

4 Case Management: The Primary Clinician . . . . . . . . 89

5 Levels of Care: Indications, Structure, Staffing . . . . 113

7 Pharmacotherapy: Selection of Medications . . . . . 157

8 Family Interventions and Therapies . . . . . . . . . . . . . 177

9 Group Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207

11 Cognitive-Behavioral Therapies: Dialectical

12 Psychodynamic Psychotherapies . . . . . . . . . . . . . . 279

13 Future Considerations . . . . . . . . . . . . . . . . . . . . . . . . 315

John G. Gunderson, M.D., is Professor of Psychiatry at Harvard Medical

Paul S. Links, M.D., F.R.C.P.C., is Professor of Psychiatry at the Univer-

THIS BOOK IS A SEQUEL to Borderline Personality Disorder: A Clinical

Three theories guide most clinicians—biological, cognitive-behav-

Chapter 12 is devoted to psychodynamic (i.e., psychoanalytic) psy-

THE BORDERLINE DIAGNOSIS

THE BORDERLINE PERSONALITY DISORDER (BPD) diagnosis en-

Origins of the Diagnosis

FIGURE 1–1. Development of the borderline construct, I.

patients needed to avoid regressing. After Knight, the term borderline to

Sidebar 1–1: Where Were the Borderline Patients

derline psychopathology is a by-product of social changes

FIGURE 1–2. Concepts of borderline disorders.

the publication of The Borderline Syndrome in 1968, Grinker and colleagues

Shifts in the Borderline Construct:

The identification of the borderline personality became very wide-

FIGURE 1–3. Development of the borderline construct, II.

thesis and implementation of what has been learned, and, as described

• Establishing a course that is distinctive from that of either psychotic or

FIGURE 1–4. Number of research studies and clinical reports on

FIGURE 1–5. Number of books on borderline personality disorder,

Therefore, although boundaries remain blurred into a variety of

An Explication of the DSM-IV Criteria

TABLE 1–1. Demographics of borderline personality disorder

1. Unstable, intense relationships. This criterion describes the interper-

TABLE 1–2. DSM-III-R and DSM-IV diagnostic criteria for

Source. Adapted from American Psychiatric Association: Diagnostic and Statistical

gratifying, or devalued, if ungratifying. Kernberg (1967) is responsi-

pression (Westen et al. 1992) and linking the borderline individual’s

Sidebar 1–2: The Subjective Experience

Linehan and other cognitive-behavioral clinicians have adopted the

tient’s behavioral specialty. This criterion is so prototypical of persons

Sidebar 1–3: Borderline Personality as

problems with noncompliance and maintaining bound-

8. Identity disturbance. This criterion is derived from Kernberg’s descrip-

and hallucinogenic phenomena (e.g., “I thought I heard my

TABLE 1–3. Prevalence of cognitive/perceptual symptoms in

A Clinical Synthesis: Intolerance of Aloneness

Sidebar 1–4: British Developmentalists:

ing—that is, serving as a container for—the borderline

and Buie 1979; Benjamin 1993; Fonagy 1991; Gunderson

Table 1–4 shows how borderline patients’ feelings of being securely

TABLE 1–4. How borderline personality disorder (BPD) patients’

haviors may be potentially lethal, indicating a readiness to die unless the

Misuses of the Borderline Diagnosis

psychiatrists because of its role in explaining mood disorders that prove

Sidebar 1–5: “Wisdom Is Never Calling