Root Cause Analysis in Health Care

Includes Flash Drive!
Root Cause Analysis in

Health Care:
Tools and Techniques
F I F T H EDI T I O N
Executive Editor: James Parker
Project Manager: Lisa King
Associate Director, Publications: Helen M. Fry, MA
Associate Director, Production and Meeting Support: Johanna Harris
Executive Director, Publications and Education: Catherine Chopp Hinckley, MA, PhD
Joint Commission/JCR Reviewers for the fifth edition:
Lisa Buczkowski; Maureen Carr; Gerard Castro; Paul vanOstenberg; Ronald Wyatt, MD
Joint Commission Resources Mission
The mission of Joint Commission Resources (JCR) is to continuously improve the safety and quality of health care in the
United States and in the international community through the provision of education, publications, consultation, and
evaluation services.
Joint Commission Resources educational programs and publications support, but are separate from, the accreditation
activities of The Joint Commission. Attendees at Joint Commission Resources educational programs and purchasers of Joint
Commission Resources publications receive no special consideration or treatment in, or confidential information about, the
accreditation process.
The inclusion of an organization name, product, or service in a Joint Commission Resources publication should not be
construed as an endorsement of such organization, product, or service, nor is failure to include an organization name,
product, or service to be construed as disapproval.
This publication is designed to provide accurate and authoritative information in regard to the subject matter covered.
Every attempt has been made to ensure accuracy at the time of publication; however, please note that laws, regulations,
and standards are subject to change. Please also note that some of the examples in this publication are specific to the
laws and regulations of the locality of the facility. The information and examples in this publication are provided with
the understanding that the publisher is not engaged in providing medical, legal, or other professional advice. If any such
assistance is desired, the services of a competent professional person should be sought.
© 2015 The Joint Commission
Joint Commission Resources, Inc. (JCR), a not-for-profit affiliate of The Joint Commission, has been designated by The
Joint Commission to publish publications and multimedia products. JCR reproduces and distributes these materials under
license from The Joint Commission.
All rights reserved. No part of this publication may be reproduced in any form or by any means without written permission
from the publisher.
Printed in the USA 5 4 3 2 1
Requests for permission to make copies of any part of this work should be mailed to
Permissions Editor
Department of Publications and Education
Joint Commission Resources
1515 W. 22nd Street, Suite 1300W
Oak Brook, Illinois 60523 USA
[email protected]
ISBN: 978-1-59940-898-9
Library of Congress Control Number: 2009931192
For more information about Joint Commission Resources, please visit http://www.jcrinc.com.
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
CHAPTER 1: Root Cause Analysis: An Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Investigating Patient Safety Events: The Need for Comprehensive Systematic Analysis . . . . . . . . . . . 1
When Can a Root Cause Analysis Be Performed? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Variation and the Difference Between Proximate and Root Causes . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Benefits of Root Cause Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Maximizing the Value of Root Cause Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
The Root Cause Analysis and Action Plan: Doing It Right . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
CHAPTER 2: Addressing Sentinel Events in Policy and Strategy . . . . . . . . . . . . . . . . . 17

The Range of Adverse Events in Health Care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Signals of Risk: Near Misses and No‑Harm Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
The Joint Commission’s Sentinel Event Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Reasons for Reporting a Sentinel Event to The Joint Commission . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Required Response to a Reviewable Sentinel Event . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Joint Commission International’s Sentinel Event Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Related Joint Commission International Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Developing Your Own Sentinel Event Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Leadership, Culture, and Sentinel Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Early Response Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Event Investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Onward with Root Cause Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
CHAPTER 3: Preparing for Root Cause Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

Step 1 Organize a Team . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Step 2 Define the Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Step 3 Study the Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
CHAPTER 4: Determining Proximate Causes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

Step 4 Determine What Happened . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Step 5 Identify Contributing Process Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Step 6 Identify Other Contributing Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Step 7 Measure—Collect and Assess Data on Proximate and Underlying Causes . . . . . . . . . . . . . 71
Step 8 Design and Implement Immediate Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
iii
ROOT CAUSE ANALYSIS IN HEALTH CARE: TOOLS AND TECHNIQUES, Fifth Edition
CHAPTER 5: Identifying Root Causes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Step 9 Identify Which Systems Are Involved—The Root Causes . . . . . . . . . . . . . . . . . . . . . . . . 81
Step 10 Prune the List of Root Causes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Step 11 Confirm Root Causes and Consider Their Interrelationships . . . . . . . . . . . . . . . . . . . . . . 86
CHAPTER 6: Designing and Implementing an Action Plan for Improvement . . . . . . . 91

Step 12 Explore and Identify Risk Reduction Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Step 13 Formulate Improvement Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
Step 14 Evaluate Proposed Improvement Actions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Step 15 Design Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Step 16 Ensure Acceptability of the Action Plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Step 17 Implement the Improvement Plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Step 18 Develop Measures of Effectiveness and Ensure Their Success . . . . . . . . . . . . . . . . . . . . . . 113
Step 19 Evaluate Implementation of Improvement Efforts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
Step 20 Take Additional Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Step 21 Communicate the Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
CHAPTER 7: Tools and Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

What Is Lean Six Sigma? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Affinity Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Brainstorming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Capability Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
Change Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Change Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Check Sheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
Control Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Failure Mode and Effects Analysis (FMEA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Fishbone Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Flowchart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
Gantt Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Histogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Kaizen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Multivoting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Operational Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Pareto Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Relations Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Run Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Scatter Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
SIPOC Process Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
Stakeholder Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Standard Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Value Stream Mapping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
iv
Contents
CHAPTER 8: Root Cause Analysis Case Studies from the Field . . . . . . . . . . . . . . . . . 175

Root Cause Analysis of Serious Adverse Events Among Older Patients in the
Veterans Health Administration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Using Root Cause Analysis and Form Redesign to Reduce Incorrect Ordering of HIV Tests . . . . 186
Counting Matters: Lessons from the Root Cause Analysis of a Retained Surgical Item . . . . . . . . . 194
Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
v
vi
Introduction
Patient safety events can cause serious harm or death. adverse events in acute care hospital admissions range from
They affect anyone. To address and prevent these threats, 2.9% in the United States to 5.0%–10.0% in the United
health care organizations must dig deep to unearth the root Kingdom, 7.5% in Canada, 12.9% in New Zealand, and
cause(s) and develop solutions that address the problems 16.6% in Australia.2
from a systems perspective.
Although these reports and chart reviews illuminate the
Indeed, the very presence of patient safety events indicates problem, it is virtually impossible to know how many
a continuing paradox in contemporary health care. Despite patients suffer as a result of health care system failures;
remarkable advances in almost every field of health care, however, any single patient safety event is a cause for
the occurrence of errors, or failures—the term used increas- concern. These events can result in tragedy for individuals
ingly instead of errors—persists. When such failures harm served and their families, add costs to an already over-
patients, the results can be heartbreaking. Most failures and burdened health care system, adversely affect the public’s
sentinel events—that is, a patient safety event (not primarily perception of an organization, and lead to litigation. They
related to the natural course of the patient’s illness or can also deeply affect health care professionals who are
underlying condition) that reaches a patient and results in dedicated to the well-being of their patients.
death, permanent harm, or severe temporary harm—are the
result of system and process flaws. These flaws are often not Health care organizations, then, have no choice but to
immediately apparent and require investigation answer one key question: Why do these errors or failures
continue to occur?
The prevalence of patient safety events had been thrust into
the limelight with the watershed report To Err Is Human: To answer this, a comprehensive systematic analysis must
Building a Safer Health System, published in 2000 by the be done. The most commonly used form of comprehensive
Institute of Medicine (IOM). The IOM report, however, systematic analysis among Joint Commission–accredited
was just the tip of the iceberg. More reports followed, organizations is root cause analysis—a process for identi-
illustrating the need to improve the quality of care being fying the basic or causal factor(s) underlying variation in
delivered in the United States. For example, researchers performance, including the occurrence or possible occur-
at Johns Hopkins Children’s Center and the US Agency rence of a sentinel event—and all of its related tools. Root
for Healthcare Research and Quality reviewed 5.7 million cause analysis can be used to uncover the factors that lead
records of patients younger than 19 years of age from 27 to patient safety events and move organizations to deliver
states who were hospitalized in 2000. Of the 52,000 chil- safer care.
dren identified by the researchers as being harmed by unsafe
medical care during their hospital stays, 4,483 suffered a Although health care organizations in the United States
fatal injury.1 often use root cause analysis to help improve quality en
route to accreditation, such analysis has many broader
Quality-of-care issues such as these are a problem for applications around the world. High-quality care is high-
hospitals around the world. According to a 2007 report, quality care, whether it is delivered in New York City or
hospital chart reviews in various countries indicate that Dubai or Singapore. Organizations worldwide should
vii
consider how root cause analysis can be used to help procedure—in this case a plastic surgical clip that was
improve quality. left inside a patient’s skull.7
• In May 2015, the United Kingdom’s National Health
The Current Health Care Environment Service paid £15,000 in damages to the mother of an
Health care continues to experience dramatic change. infant who died in utero due to a medical error.8
Health care organizations are evolving constantly because
of changes in reimbursement, new technology, regulatory The above examples are only a few of the serious patient
requirements, and staffing levels. These modifications cause safety events that have attracted media attention in recent
policies and procedures to change often and, in most cases, years. These events cast a shadow on the public’s trust of
quickly. As health care organizations become more complex, health care. Stakeholders, including patients, justifiably
their systems and processes are increasingly interdependent. ask, “What’s going on?” Failure detection, reduction, and
This interdependence increases the risk of failures and can prevention strategies are receiving new impetus as the
make the recovery from failure more difficult. Clinical and health care community recognizes the value of a proactive
support staff workloads are growing heavier, resulting in approach to reducing risk.
greater stress and fatigue for many health care professionals.
Caregivers are working in new settings and performing new Root cause analysis is one such approach. Historically used
functions, sometimes with minimal training. Consequently, to investigate sentinel events, root cause analysis shows great
maintaining consistency in processes and systems is chal- promise as a proactive tool. Increasingly, health care organi-
lenging, leading to variation. Often, this variation results in zations are using this methodology to investigate close calls
increased risk to patients. (or near misses), no-harm patient safety events, and other
signals of risk. Health care organizations no longer have to
Media reports about patient safety events are occurring with wait until after a sentinel event occurs to perform a root
increasing regularity, including the following examples: cause analysis.
• In September 2013, researchers estimated that the
number of premature deaths associated with preventable When an adverse outcome, a sentinel event, or a cluster
harm to patients in US hospitals was more than 400,000 of less serious incidents or near misses occurs, organiza-
per year. This makes patient safety events the third tions must develop an understanding of the contributing
leading cause of death in the United States. Incidents factors and the interrelationship of those factors. Next, the
resulting in serious patient harm were estimated to be organization must implement an action plan to fortify its
10- to 20-times more common than lethal harm.3 systems against vulnerabilities with the potential to impact
• In November 2014, the journal Pediatrics reported that patients. Resilience is the degree to which a system contin-
an annual average of 63,358 medication errors occur uously prevents, detects, mitigates, or ameliorates hazards
in children younger than age 6 in the United States in or incidents.9
nonhospital settings and that 25% of those errors are in
infants, younger than 12 months old. This means that Purpose of This Book
a medication error affecting a child in the United States Root Cause Analysis in Health Care: Tools and Techniques,
occurs every eight minutes.4 Fifth Edition, is intended to help health care organizations
• In February 2015, the state of Minnesota reported that prevent systems failures by using root cause analysis to do
98 patients in that state were seriously injured, and the following:
another 13 patients died, as a result of patient safety • Identify causes and contributing factors of a sentinel
events during 2014.5 event or a cluster of incidents
• In May 2015, the Jordanian Ministry of Health began • Identify system vulnerabilities that could lead to
investigating an alleged medical error that resulted in a patient harm
Saudi patient becoming comatose.6 • Implement risk reduction strategies that decrease the
• In May 2015, the State of California fined a hospital likelihood of a recurrence of the event or incidents
$100,000 after the unintended retention of a foreign • Determine effective and efficient ways of measuring and
object in a patient’s body following an invasive improving performance
viii
Introduction
Root cause analysis is an effective technique most commonly to meet the unique needs of the team and organization.
used after an error has occurred to identify underlying causes. Appropriate tools for use in each stage of root cause anal-
Failure mode and effects analysis (FMEA) is a proactive ysis are identified in each chapter. A chapter-by-chapter
technique used to prevent process and product problems description of the contents follows.
before they occur.10 Health care organizations should learn
both techniques to reduce the likelihood of adverse events. Chapter 1, “Root Cause Analysis: An Overview,” takes a
holistic look at root cause analysis. It describes variation,
Root Cause Analysis in Health Care: Tools and Techniques, how proximate and root causes differ, when root cause anal-
Fifth Edition, provides health care organizations worldwide ysis can be conducted, and the benefits of root cause anal-
with up-to-date information on The Joint Commission’s ysis. One of the benefits involves effectively meeting Joint
Sentinel Event Policy and safety-related requirements. Commission and Joint Commission International require-
It also describes the Sentinel Event Policy of Joint ments that relate to the management of sentinel events. The
Commission International. The book includes examples chapter also provides guidelines on the characteristics of a
that guide the reader through application of root cause thorough and credible root cause analysis and action plan.
analysis to the investigation of specific types of sentinel
events, such as medication errors, suicide, treatment delay, Chapter 2, “Addressing Sentinel Events in Policy and Strategy,”
and elopement. For ease of access and use by root cause describes the types of adverse events occurring in health
analysis teams, practical checklists and worksheets are care. The Joint Commission’s Sentinel Event Policy and
offered at the end of each chapter. requirements are listed in full, including a description
of reportable and reviewable events. Joint Commission
This publication provides and explains The Joint International’s Sentinel Event Policy also is discussed. The
Commission’s framework for conducting a root cause chapter provides practical guidelines on how an organiza-
analysis. It also helps organizations do the following: tion can develop its own sentinel event policy, including
• Identify the processes that could benefit from root cause the role that an organization’s culture and leadership play in
analysis risk reduction and prevention. It describes the need for root
• Conduct a thorough and credible root cause analysis cause analysis and provides practical guidance on the early
• Interpret analysis results steps involved in responding to an adverse or sentinel event.
• Develop and implement an action plan for improvement
• Assess the effectiveness of risk reduction efforts Chapter 3, “Preparing for Root Cause Analysis,” covers the
• Integrate root cause analysis with other programs early steps involved in performing a root cause analysis. The
first of four hands-on workbook chapters, it describes how
Even without the occurrence of an adverse event, health to organize a root cause analysis team, define the problem,
care organizations should embrace the use of root cause and gather the information and measurement data to study
analysis to minimize the possibility of patient safety events the problem. Details are provided about team composi-
and thereby to improve the care, treatment, and services tion and ground rules. The chapter also covers how to use
provided at their facilities. information gleaned from The Joint Commission’s Sentinel
Event Database and accreditation requirements to identify
Overview of Contents problem areas in need of root cause analysis. The chapter
Root Cause Analysis in Health Care: Tools and Techniques, provides guidance on recording information obtained
Fifth Edition, provides health care organizations with during a root cause analysis, conducting interviews, and
practical, how-to information on conducting a root cause gathering physical and documentary evidence.
analysis. Twenty-one steps are described (in Chapters 3
through 6). Teams conducting a root cause analysis Chapter 4, “Determining Proximate Causes,” provides
might not follow these steps in a sequential order. Often, practical guidance on the next stage of root cause analysis—
numerous steps will occur simultaneously, or the team will determining what happened and the reasons it happened.
return to earlier steps before proceeding to the next step. Organized in a workbook format, the chapter describe s
It is crucial for teams to customize or adapt the process how to further define the event, identify process problems,
ix
determine which care processes are involved with the Chapter 8, “Root Cause Analysis Case Studies from the Field,”
problem, and pinpoint the human, process, equipment, presents root cause analyses that resulted from real-life inci-
environmental, and other factors closest to the problem. dents at health care organizations. In these studies, the tools
The chapter also addresses how to collect and assess data on and techniques used to dig down to the root causes of the
proximate and underlying causes. In addition, the chapter events are identified and explained.
describes the process of designing and implementing
interim changes. Finally, the Glossary provides definitions of key terms used
throughout the book.
Chapter 5, “Identifying Root Causes,” provides practical
guidance, through workbook questions, on identifying A Word About Terminology
or uncovering the root causes—the systems that underlie The terms patient, individual served, and care recipient all
sentinel events—and the interrelationship of the root causes describe the individual, client, consumer, or resident who
to one another and to other health care processes. Systems actually receives health care, treatment, and/or services. The
are explored and described, including human resources, term care includes care, treatment, services, rehabilitation,
information management, environment of care, leader- habilitation, or other programs instituted by an organiza-
ship, communication, and uncontrollable factors. The tion for individuals served.
chapter also addresses how to differentiate root causes and
contributing factors. Acknowledgments
Joint Commission Resources is grateful to the many
Chapter 6, “Designing and Implementing an Action Plan content experts who have contributed to this publication,
for Improvement,” includes practical guidelines on how to including the following:
design and implement an action plan—the improvement • Lisa Buczkowski, RN, MS, associate director, Office of
portion of a root cause analysis. During this stage, an orga- Quality and Patient Safety, The Joint Commission
nization identifies risk reduction strategies and designs and • Maureen Carr, RN, project director, Division of
implements improvement strategies to address underlying Standards and Survey Methods, The Joint Commission
systems problems. Characteristics of an acceptable action • Gerard Castro, PhD, MPH, project director, Office of
plan are provided, as is information on how to assess the Patient Safety, The Joint Commission
effectiveness of improvement efforts. The chapter concludes • Doreen Finn, RN, associate director, Standards
with information on how to effectively communicate the Interpretation Group, The Joint Commission
results in improvement initiatives. • Paul van Ostenberg, senior executive director,
International Accreditation and Standards, Joint
Chapter 7, “Tools and Techniques,” presents the tools and Commission International
techniques used during root cause analysis. Each tool • Ronald Wyatt, MD, medical director, Office of Quality
profile addresses the purpose of the tool, the appropriate and Patient Safety, The Joint Commisison
stage(s) of root cause analysis for the tool’s use, simple steps
for success, and tips for effective use. Twenty-three tools In addition, JCR would like to thank writer James Foster
are profiled: affinity diagrams, brainstorming, capability for his help in revising this book.
charts, change analysis, change management, check sheets,
control charts, failure mode and effects analysis, fishbone
diagrams, flowcharts, Gantt charts, histograms, run charts,
scatter diagrams, SIPOC process maps, stakeholder analysis,
and other tools. Preceding the tool descriptions is a discus-
sion of a performance improvement methodology, Lean Six
Sigma, that incorporates many of these tools.
x
Introduction
7. San Jose Mercury News. State Fines Marin General Hospital

References $100,000 for Leaving Object in Patient’s Skull. Halstead R.
1. Miller MR, Zhan C. Pediatric patient safety in hospitals: A
May 8, 2015. Accessed Oct 14, 2015. http://www.mercurynews
national picture in 2000. Pediatrics.2004 Jun;113(6):1741–1746.
.com/crime-courts/ci_28163048state-fines-marin-general
2. Peters GA, Peters BJ. Medical Error and Patient Safety: Human -hospital-100-000-leaving.
Factors in Medicine. Boca Raton, FL: CRC Press, 2007.
8. STV Group. Health Board Apologises After Unborn Baby Dies
3. James JT. A new evidence-based estimate of patient harms Following Doctor Error. May 19, 2015. Accessed Oct 14, 2015.
associated with hospital care. J Patient Saf. 2013 Sep;9(3):122–128. http://news.stv.tv/stirling-central/1320942-forth-valley-hospital
4. Smith MD, et al. Out-of-hospital medication errors among -apology-after-sarah-mackinlay-of-falkirk-lost-baby/.
young children in the United States, 2002–2012. Pediatrics. 2014 9. World Health Organization (WHO). More Than Words:
Nov;134(11):867–876. Conceptual Framework for the International Classification for
5. Minnesota Department of Health (MDH). Adverse Health Events Patient Safety: Final Technical Report, ver. 1.1. Geneva: WHO,
in Minnesota: 11th Annual Public Report. St. Paul: MDH: 2015. 2009. Accessed Oct 14, 2015. http://www.who.int/patientsafety
Accessed Oct 14, 2015. http://www.health.state.mn.us /taxonomy/icps_full_report.pdf.
/patientsafety/ae/2015ahereport.pdf. 10. The Joint Commission. Failure Mode and Effects Analysis in
6. Arab News. Saudi Brain Dead After ‘Medical Error’ in Jordan Health Care: Proactive Risk Reduction, 3rd ed. Oak Brook, IL:
Hospital. May 8, 2015. Accessed Oct 14, 2015. http://www Joint Commission Resources, 2010.
.arabnews.com/saudi-arabia/news/743946.
xi
xii
CHAPTER
1
Root Cause Analysis: An Overview
Learning Objectives
• Understand the need for comprehensive systematic analysis of sentinel events an other adverse outcomes
• Learn the basics of root cause analysis (RCA), the most common method of comprehensive systematic analysis
• Know how RCA and action plans relate to The Joint Commission’s Sentinel Event Policy
Investigating Patient Safety Events: sentinel event as an unanticipated occurrence involving

The Need for Comprehensive death or major permanent loss of function unrelated to the
Systematic Analysis natural course of the patient’s illness or underlying condi-
The Joint Commission’s Sentinel Event Policy requires tion. A root cause analysis focuses primarily on systems and
accredited health care organizations to conduct a compre- processes, not individual performance. To be successful,
hensive systematic analysis in the wake of a sentinel event. the objective of an RCA must not be to assign individual
Comprehensive systematic analysis seeks to go beyond blame. Rather, through the RCA determine process, a team
individual performance issues to determine how gaps works to understand a process or processes, the causes or
in policies and safety systems may have contributed to potential causes of variation that can lead to error, and
an adverse event and to identify changes to policies and identify process changes that would make variation less
procedures that may prevent similar events from occurring likely to recur.
in the future. The Joint Commission reviews methods of
comprehensive systematic analysis on a case-by-case basis to A root cause is the most fundamental reason (or one of
determine their credibility, thoroughness, and acceptability. several fundamental reasons) a failure, or a situation
The Joint Commission also provides advice and resources in which performance does not meet expectations, has
to institutions to assist them in assessing analytical tools. occurred. In common usage, the word cause suggests
However, RCA is by far the most common method and is responsibility or a factor to blame for a problem. In the
the method preferred by The Joint Commission. context of an RCA, however, the use of the word cause does
not imply an assignment of blame. Instead, the cause refers
What Is Root Cause Analysis? to a relationship or potential relationship between certain
Root cause analysis is a process for identifying the basic factors that enable a sentinel event to occur. The focus in
or causal factor(s) underlying variation in performance.
* Severe temporary harm is critical, potentially life-threatening harm
Variation in performance can (and often does) produce lasting for a limited time with no permanent residual, but requires
unexpected and undesired adverse outcomes, including the transfer to a higher level of care/monitoring for a prolonged period of
occurrence or risk of a sentinel event. The Joint Commission time, transfer to a higher level of care for a life-threatening condition,
or additional major surgery, procedure, or treatment to resolve the
defines sentinel event as a patient safety event (not condition. Adapted from Healthcare Performance Improvement, LLC.
primarily related to the natural course of the patient’s illness The HPI SEC & SSER Patient Safety Measurement System for Healthcare,
or underlying condition) that reaches a patient and results rev. 2. Throop C, Stockmeier C. May 2011. Accessed Oct 15, 2015.
http://hpiresults.com/publications/HPI%20White%20Paper%20-%20
in death, permanent harm, or severe temporary harm.* SEC%20&%20SSER%20Measurement%20System%20REV%202%20
Similarly, Joint Commission International (JCI) defines MAY%202011.pdf.
1
an RCA is on a positive, preventive approach to system Sidebar 1-1.

and process changes following a sentinel event, a near-miss
Root Cause Analysis Case Example
sentinel event, or a cluster of less serious yet potentially
harmful incidents. A 16-year-old patient came to the hospital to deliver
her baby. During the process of her care, an infu-
As shown in Sidebar 1-1, right, root cause analysis can do sion intended exclusively for the epidural route was
more than discover that “A caused B.” The process also connected to the patient’s peripheral intravenous line
can help an organization determine that “if we change A and infused by pump. The patient experienced cardio-
because we had a problem with it, we can reduce the possi- vascular collapse. A cesarean section resulted in the
bility of B recurring or in fact prevent B from occurring in delivery of a healthy infant, but the medical team was
unable to resuscitate the mother.
the first place.”
The media attention surrounding the error accelerated
RCA is a powerful and useful tool that can help health care through the national provider and safety community
organizations around the world reduce errors and move when the nurse was charged with a criminal offense.
These events set in motion intense internal and
quality efforts forward. However, health care organization
external scrutiny of the hospital’s medication and safety
leaders must realize that RCA is not a panacea but one procedures.
tool that should be used in conjunction with others to
improve care. To further understanding about latent systems gaps
and process failure modes, a root cause analysis of the
event was conducted. A team conducted a one-week
When Can a Root Cause Analysis evaluation of the medication use system and the orga-
Be Performed? nization’s current environment, systems and processes,
Historically, root cause analysis has most commonly been staffing patterns, leadership, and culture to help shape
used retrospectively—to probe the reasons for a bad outcome the recommended improvements. For each of the four
or for failures that have already occurred. Root cause anal- proximate causes of the event, performance-shaping
ysis can also be used to probe a near-miss event or pattern factors were identified.
of events or as part of other performance improvement Although the hospital’s organizational learning was
redesign initiatives, such as gaining an understanding of painful, this event offered an opportunity for increasing
variations observed in systematically collected data. The organizational competency and capacity for designing
best RCAs look at the entire process and all support systems and implementing patient safety. Structures and
involved in a specific event to minimize overall risk associ- processes, including safety nets and fail-safe mecha-
nisms, were implemented to promote safer behavioral
ated with that process, as well as the recurrence of the event
choices for providers.
that prompted the root cause analysis.1 The goal of the root
cause analysis is to produce an action plan that identifies the The hospital took a number of clinical steps to improve
strategies the organization intends to implement to reduce the safety of medication administration, including
removing the barriers to scanning medication bar
the risk of similar events occurring in the future.
codes, implementing consistent scanning-compliance
tracking, and providing teamwork training for all nursing
Root cause analysis is also used increasingly by organi- and physician staff practicing in the birth suites.
zations as one step of a proactive risk reduction effort
Source: Smetzer J, et al. Shaping systems for better behavioral
using failure mode and effects analysis (FMEA). FMEA is a choices: Lessons learned from a fatal medication error. Jt Comm J Qual
proactive, prospective approach used to prevent process and Patient Saf. 2010 Apr;36(4):152–163.
product problems before they occur. It provides a look not

only at what problems could occur—the failure modes— Figure 1-1, page 3, shows a frequency/severity matrix one
but also at how severe the effects of those problems could organization uses to help decide when to conduct a root
be. The goal is to prevent poor results, which in health care cause analysis as part of an FMEA. In this matrix, if a
means harm to patients. One step of FMEA involves iden- process failure gets a high score both for severity (or poten-
tifying the root causes of the failure modes.2 At this point in tial severity) of outcome and for frequency, then a root cause
the process, the FMEA team can use the RCA approach. analysis should be done.
2
CHAPTER 1 | Root Cause Analysis: An Overview
RCA in High Reliability Industries and workers are trained accordingly to anticipate, recognize,
Industries that are regarded as highly reliable, such as and either avoid or quickly recover from errors. In contrast,
nuclear power or the aerospace industry, also employ RCA sentinel events in the health care environment occur with
methodology to investigate adverse events. In the nuclear relative frequency and tend to be handled reactively.
power and aerospace industries, sentinel events are rare
because they have been anticipated. These high reliability Domestic and International Requirements
industries have adopted a systems approach, in which errors Both The Joint Commission, which accredits health care
are viewed as an expected part of the workplace, the result organizations in the United States, and Joint Commission
of a chance misalignment of weaknesses in the underlying International (JCI), which accredits health care organi-
system.3 (Imagine a stack of slices of Swiss cheese. Each slice zations in countries other than the United States, have
has holes in different places, thus only if a hole in each slice a Sentinel Event Policy and standards related to sentinel
aligned perfectly with a hole in all the other slices could an events. For example, JCI Quality Improvement and Patient
object pass through the entire stack. The Swiss cheese model Safety (QPS) standards require each accredited organization
shown in Figure 1-2, page 4, represents how an error could to establish which unanticipated events are significant and
possibly penetrate multiple layers of barriers, defenses, and the process for their intense analysis.
safeguards in a system.)
While the determination of what constitutes a signifi-
Consequently, systems, often with significant redundancies, cant event must be consistent with the general definition
have been built to protect against the occurrence of errors, of sentinel event as described in JCI’s policy, accredited
Figure 1-1. Frequency/Severity Matrix for Prioritizing Safety-Related Problems
This matrix helps the organization apply its resources (such as time) to areas where the opportunity to improve
safety is greatest. The larger the number, the more urgent the problem and the more useful a root cause analysis
would be.
Source: US Department of Veterans Affairs National Center for Patient Safety. VHA National Patient Safety Improvement Handbook. Mar 4, 2011, page
B-2. Accessed Oct 15, 2015. http://www.va.gov/vhapublications/ViewPublication.asp?pub_ID=2389.
3
organizations have some latitude in setting more-specific Both JCI– and Joint Commission–accredited organizations
parameters to define unanticipated and major permanent loss are expected to identify and respond appropriately to all
of function. At a minimum, an organization must include sentinel events that occur in the organization or that are
those events that are subject to review under JCI standards, associated with services that the organization provides or
such as unanticipated death related to the natural course of provides for. Appropriate response includes conducting a
the patient’s illness or underlying condition; major perma- timely, thorough, and credible comprehensive systematic
nent loss of function unrelated to the natural course of the analysis; developing an action plan designed to implement
patient’s illness or underlying condition; or wrong-site, improvements to reduce risk; implementing the improve-
wrong-procedure, wrong-patient surgery. ments; and monitoring the effectiveness of those improve-
ments. (See Sidebar 1-2, page 5, for a discussion of
For organizations based in the United States, standards multiple events.)
require the organization to have an organizationwide,
integrated patient safety program within its performance Variation and the Difference Between
improvement activities. Proximate and Root Causes
Whether addressing a sentinel event or a cluster of less
Particularly, thorough and credible comprehensive serious low-harm or near-miss events, root cause analysis in
systematic analyses (for example, root cause analyses) in all environments provides two challenges:
response to sentinel events they are required to conduct. 1. To understand why the event occurred
RCA is the most commonly used form of comprehensive 2. To prevent the same or a similar event from occurring
systematic analysis used by Joint Commission–accredited in the future through prospective process design
organizations to comply with this requirement. or redesign
Figure 1-2. Swiss Cheese Model
The Swiss cheese model shows how an error could penetrate multiple layers of defenses, barriers, and safeguards
in a system.
Source: Reason J.: The Human Contribution: Unsafe Acts, Accidents and Heroic Recoveries. Farnham, England: Ashgate Publishing, 2008.
© Managing the Risks of Organizational Accidents, by James Reason, 1997, Ashgate. Reproduced by permission.
4
Sidebar 1-2. those events or possible events with a significant negative

or potentially negative impact on the patient. The criterion
Investigating Multiple Patient
for a sentinel event is generally death, permanent harm, or
Safety Events severe temporary harm to the patient.
Although root cause analysis (RCA) is associated
frequently with the investigation of a single sentinel
Adverse or sentinel events involve unexpected variation in
event, the methodology also can be used to determine a process. When this variation occurs, the probability of
the cause of multiple occurrences of low-harm events. a serious adverse outcome increases. As mentioned previ-
For example, 40 Danish community pharmacies worked ously, RCA is a process for identifying the basic or causal
together, using RCA, to gain insight into medication factor(s) underlying variation in performance. Variation is
errors, a problem that can result in serious conse- a change in the form, position, state, or qualities of a thing.
quences for patients. Although a sentinel event is the result of an unexpected
The root cause analysis included investigation of variation in a process, variation is inherent in every process.
401 errors, many of which had potential clinical signif- To reduce variation, it is necessary to determine its cause.
icance, even though each error did not necessarily What’s more, variation can be classified by what caused it.
result in harm. Analyzed as a cluster, however, the RCA
resulted in the identification of four common medication
error causes1:
1. Illegible handwritten prescriptions Figure 1-3. The Relationships Between a

2. Similarities in packaging or names, or strength and Sentinel Event and Its Causes
dosage stated in misleading ways
3. Lack of effective control of prescription label and
medicine
4. Lack of concentration caused by interruptions.
Reference:
1. Knudsen P, et al. Preventing medication errors in community
pharmacy: Root-cause analysis of transcription errors. Qual Saf
Health Care. 2007 Aug;16(4):285–290.
To meet these challenges, organizations must under-

stand not only the proximate causes, or active failures (the
apparent, seemingly obvious reasons an error occurred) but
also the underlying causes, or latent conditions (the aspects
of a process than can allow an error to occur), and the
interrelationship of the two. As shown in Figure 1-3, right,
active failures are only the “tip of the iceberg”—that which
is visible or proximal to the patient—while latent condi-
tions lurk unseen “underwater,” posing a hidden poten-
tial danger. Root cause analysis helps organizations delve
beneath the proximate causes to find the underlying causes
of a sentinel event.
Conducting an RCA has significant resource implications.

Active failures, shown as the “tip of the iceberg,”
A team approach, involving a full range of disciplines and
reveal only the proximate causes of a sentinel
departments in the process being studied, is mandatory, as event; latent conditions loom below the waterline,
will be described in Chapter 3. Organizations therefore will representing the underlying causes.
want to conduct root cause analysis principally to explore
5
Understand Common-Cause Variation Special causes in one process are usually the result of
Common-cause variation, although inherent in every common causes in a larger system of which the process is
process, is a consequence of the way a process is designed to a part. For example, mechanical breakdown of a piece of
work. For example, an organization is examining the length equipment used during surgery may indicate a problem
of time required by the emergency department to obtain a with an organization’s preventive maintenance activities.
routine radiology report. The time may vary depending on
how busy the radiology service is or by when the report is Understand the Relationships Between
requested. On a particular day, the radiology department Common and Special Causes
may have received many concurrent requests for reports, In health care, all the clinical and organizational processes
making it difficult for the department to fill one specific and subprocesses associated with an event under review
request. Or the report may have been requested between need to be delineated and evaluated to identify the degree of
midnight and 6:00 a.m. when fewer radiology technologists common-cause and/or special-cause variation. This process
are on duty. Variation in the process of providing radiology will help organizations identify whether variation is due to
reports is inherent, resulting from common causes such as clinical processes or organizational processes or both.
staffing levels and emergency department census.
Any variation in performance, including a sentinel event,
A process that varies only because of common causes is said may be the result of a common cause, a special cause, or
to be stable. The level of performance of a stable process both. In the case of a sentinel event, the direct or proximate
or the range of the common-cause variation in the process special cause could be uncontrollable factors. For example,
can be changed only by redesigning the process. Common- a patient death results from a hospital’s total loss of elec-
cause variation is systemic and endogenous (that is, produced trical power during a storm. This adverse outcome is clearly
from within). The organization needs to determine whether the result of a special cause in the operating room that is
the amount of common-cause variation will be tolerated. uncontrollable by the operating room staff. Staff members
may be able to do little to prevent a future power outage
Special-Cause Variation and more deaths. However, the power outage and resulting
Special-cause variation arises from unusual circumstances or death can also be viewed as the result of a common cause in
events that may be difficult to anticipate and may result in the organization’s system for preparing for and responding
marked variation and an unstable, intermittent, and unpre- to a utility failure and other emergencies. Perhaps the
dictable process. Special-cause variation is not inherently backup generator that failed was located in the basement,
present in systems. It is exogenous (that is, produced from which flooded during the storm, and the organization had
the outside), resulting from factors that are not part of the no contingency plan for such a situation.
system as designed. Mechanical malfunctions, fatigued
employees, and natural disasters such as floods, hurri- When looking at the chain of causation, proximate or
canes, and earthquakes are examples of special causes that direct causes tend to be nearest to the origin of the event.
result in variation. Organizations should strive to identify, For example, proximate causes of a medication error may
mitigate, and/or eliminate special causes wherever possible. include an outdated drug, product mislabeling or misiden-
However, removing a special cause eliminates only that tification, or an improper administration technique. By
current abnormal performance in the process. It does not contrast, root causes are systemic and appear far from the
prevent the same special cause from recurring. For example, origin of the event, often at the foundation of the processes
firing an overly fatigued employee who was involved in a involved in the event. For example, root causes of a medi-
medication error does little to prevent the recurrence of cation error might include manufacturer’s production or
the same error. Instead, organizations should investigate, labeling of two different types or strengths of drugs so that
understand, and address underlying common causes within one looks too much like the other, storage setup that places
their systems and processes such as staffing arrangements, different dosages of the same medication too close together,
employee education, complacency, information manage- an inadequate medication procurement process, communi-
ment, and communication. cation problems, or any number of system, issues that set
people up to make a mistake.
6
Most root causes alone are not sufficient to cause a failure; Maximizing the Value of Root Cause Analysis
rather, the combination of root cause(s) and other contrib- Root cause analysis is designed to answer the following
uting factors sets the stage. For example, flaws in the process three questions:
for communicating changes in the condition of a patient, 1. What happened?
a poorly designed emergency call system, and an inadequate 2. Why did it happen?
assessment process can be root causes of a patient’s fall from 3. What can be done to prevent it from happening again?
bed. Organizations that are successful in effectively identi-
fying all the root causes and their interactions can eliminate The problem, however, is that health care organizations
a plethora of risks when redesigning processes.4 Elimination frequently use root cause analysis to answer these questions
of one root cause reduces the likelihood of that one specific but never determine whether the risk of recurrence of an
adverse outcome occurring again. However, if the organi- adverse event has actually been reduced. Therefore, some
zation misses two or three other root causes, it is possible health care organizations may dedicate resources to root
that they could interact to cause a different but equally cause analysis without knowing whether the investment
adverse outcome. has any payoff. To make root cause analysis more useful,
follow-up activities that measure the implementation of
process changes and improvements in patient outcomes
TIP should become a standardized component of the process.5
Organizations must steer clear of the “myth of one
root cause” and expect to find several possibilities. To ensure that the RCA yields improvement, health care
leaders must address fundamental challenges that can limit
the value of the incident investigation. Consider the strate-
Benefits of Root Cause Analysis gies listed in Sidebar 1-3, page 8.
All health care organizations experience problems of
varying persistence and magnitude. Organizations can The outcome of the root cause analysis is an action plan
improve the efficiency and effectiveness of their operations that the organization intends to implement in order to
and the quality and safety of care through addressing the reduce the risk of similar events occurring in the future.
roots of such problems. Individual accountability for faulty The plan should address responsibility for implementa-
performance should not be the focus of a root cause anal- tion, oversight, pilot testing as appropriate, time lines, and
ysis. (See Chapter 3 for additional discussion of individual strategies for measuring the effectiveness of actions.
accountability.) If a question arises regarding whether
an individual acted appropriately, it should be addressed The Root Cause Analysis and Action Plan:
through the organization’s employee or physician perfor- Doing It Right
mance management system. For the purpose of an RCA, How can an organization ensure that its RCA and action
the focus should be on systems—how to improve systems to plan represent an appropriate response to a particular
prevent the occurrence of sentinel events or problems. This sentinel event? The Joint Commission and JCI provide
approach involves digging into the organization’s systems to guidance to organizations conducting RCAs in their respec-
find new ways to do things. Root cause analysis helps orga- tive sentinel event policies. These policies provide criteria
nizations identify risk or weak points in processes, under- that organizations can use to evaluate their RCA for accept-
lying or systemic causes, and corrective actions. Moreover, ability, thoroughness, and credibility. Organizations can use
information from RCAs shared between and among organi- the tool in Figure 1-4, pages 9–10, to review their RCAs
zations can help prevent future sentinel events. Knowledge based on these criteria.
shared in the health care field can contribute to proactive
improvement efforts and yield results across the health care Crafting an Acceptable Action Plan
delivery system. The Joint Commission and JCI also provide criteria for an
acceptable action plan within their sentinel event policies.
7
Sidebar 1-3. reduction activities. This information may benefit organiza-

tions that are developing their own action plans. Data gath-
Strategies for an Effective RCA
ered by The Joint Commission between January 1995 and
To ensure the effectiveness of their root cause analyses
December 2014 from review of more than 8,876 sentinel
(RCAs), organizations and their leaders should imple- events indicate that nearly 90% of these events fall into the
ment the following strategies: following categories:
• Anesthesia-Related Event
►► Find and resolve latent conditions as well as root
causes. Finding the cause of a problem is a good
• Criminal Event
start, but to be effective, identify the latent condi- • Delay in Treatment
tions that allowed the problem to materialize in the • Dialysis-Related Event
first place. • Elopement
►► Treat the cause rather than try to change people.
• Fall
Many RCA actions are aimed at changing the • Fire
behavior of staff members. Such interventions are • Infant Abduction
rarely effective. Although one person might change • Infant Discharge to Wrong Family
his or her behavior, another person might still • Infection-Related Event
do the same thing as before because the under- • Inpatient Drug Overdose
lying problem with the system has not changed. • Maternal Death
Leaders must understand that processes need to be
• Medical Equipment–Related
changed—and, therefore, more effective systems-
based solutions should be developed.
• Medication Error
• Op/Post-Op Complication
►► Follow through to ensure change. Often health
• Perinatal Death/Injury
care organization leaders initiate an improvement
• Radiation Overdose
program based on the findings of a root cause anal-
ysis, but the initiative loses steam. It is important for
• Restraint-Related Event
leaders to make sure that improvement programs • Self-Inflicted Injury
are fully implemented. • Severe Neonatal Hyperbilirubinemia
• Suicide
Source: The Joint Commission. Patient safety systems. In 2015 • Transfer-Related Event
Comprehensive Accreditation Manual for Hospitals. Oak Brook, IL: 2015.
• Transfusion Error
• Unassigned
According to these policies, an action plan will be consid- • Unintended Retention of a Foreign Body
ered acceptable if it does the following: • Utility System Failure
• Identifies changes that can be implemented to reduce • Ventilator Death
risk or formulates a rationale for not undertaking • Wrong-Patient, Wrong-Site, Wrong-Procedure
such changes
• Identifies, in situations in which improvement actions An organization experiencing a sentinel event in one of
are planned, including the following: these categories is expected to conduct a thorough and
–– Who is responsible for implementation credible root cause analysis, which, at a minimum, investi-
–– When the action will be implemented gates each of the areas identified for that category of event.
(including any pilot testing) This inquiry should determine that there is, or is not,
–– How the effectiveness of the actions will be evaluated opportunity with the associated system(s), process(es), or
–– The point at which alternative actions will be function(s) to redesign or otherwise take action to reduce
considered if improvement targets are not met risk. A root cause analysis submitted in response to a
sentinel event in one of the listed categories is considered
Review of the root cause analyses of sentinel events has unacceptable if it does not, at a minimum, address each of
allowed The Joint Commission to identify patterns for risk the areas specified for that type of event.
8
Figure 1-4. Root Cause Analysis Evaluation Checklist
Root Cause Analysis Evaluation Checklist

Date of Evaluation: Date of Event:
Completed by:
Brief Description of Incident:
Root Cause Analysis Team Participants:
Evaluation Level 1: Acceptability

Follow-Up Action Follow-Up Action
Criteria Criteria Met? (Y/N) Required? Completed (Date)
Progresses from special causes in clinical
processes to common causes in organiza-
tional processes.
Repeatedly digs deeper by asking “Why?”
and then, when answered, asks “Why?”
again, until it no longer makes sense to
ask again.
Identifies changes that could be made in
systems and processes (either through
redesign or development of new systems or
processes) that would reduce the risk of such
events occurring in the future.
Evaluation Level 2: Thoroughness

Determines the factors most directly
associated with the sentinel event and
the process(es) and systems related to its
occurrence, including human factors
Analyzes the underlying systems and
processes through a series of “Why?”
questions to determine where redesign
might reduce risk
Inquires into all areas appropriate to the
specific type of event
Identifies the risk points and their potential
contributions to this type of event
(continued)
9
Figure 1-4. Root Cause Analysis Evaluation Checklist (continued)
Evaluation Level 2: Thoroughness (continued)

Determines potential improvement in
processes or systems that would tend to
decrease the likelihood of such events in the
future, or a determination, after analysis, that
no such improvement opportunities exist
Evaluation Level 3: Credibility

Includes participation by the patient safety
director and by individuals most closely
involved in the processes and systems
under review.
Is internally consistent (that is, the analysis
does not contradict itself or leave obvious
questions unanswered).
Provides an explanation for all findings of
“not applicable” or “no problem.”
Includes consideration of any relevant liter-
ature, guidelines, or evidence-based best
practices.
When conducting a root cause analysis, investigators must Orientation and Training of Staff
ask questions that meet the minimum scope of root cause • Does the hospital offer medication safety training?
analysis for specific types of sentinel events. The following • Did the nurse involved in the error participate in
are sample questions that might be used when investigating medication safety training?
a medication error.
Competency Assessment/Training
Patient Identification Process • Are nurses at the hospital required to demonstrate
• Are specific patient identification processes and protocols competency in medication administration?
in place? • Did the nurse who was involved in the error demonstrate
• Did the nurse verify the patient’s identity? medication administration competency?
• Was the patient identified by a bar-coded wristband or
any other means? Supervision of Staff
• Who was supervising the nurse who was involved
Staffing Levels in the error?
• What are the typical staffing levels on the unit? • How many other nurses was the supervisor
• How many staff members were working on the unit responsible for?
where the error occurred? • Does the supervisor specifically oversee the medication
• How many patients were assigned to the nurse who was administration process?
involved in the error?
10
TIP Sidebar 1-4. Get Support

Use a WWW (Who, What, When): A WWW is a simple
An organization can seek clarification of any questions
tool that helps RCA teams clarify the actions that
about The Joint Commission’s Sentinel Event Policy
need to be taken, who is responsible for completing
and the requirements for a comprehensive systematic
or facilitating those actions, the time frames for when
analysis by visiting http://www.jointcommission.org,
those actions will be completed. The tool consists
which includes detailed information on the Sentinel
of a simple grid, such as in the example below. The
Event Policy and root cause analysis (see Figure 1-5,
first column lists the individual responsible for the
pages 13–16). Note that The Joint Commission will not
action, the second lists the action itself, and the third
give a determination of reviewability at this point but
indicates a due date.
can answer questions and provide support.
WHO WHAT WHEN The Joint Commission established the Sentinel Event
Hotline to respond to inquiries about the Sentinel Event
Policy. For more information visit http://www.joint
commission.org/SentinelEvents/PolicyandProcedures/.
Physical Environment
• Did any environmental factors make it difficult for the
nurse to properly carry out medication administration
duties?
Communication Among Staff Members • What environmental factors (for example, lighting,
• Are there established processes and protocols in place for space considerations) would make it easier for staff
nurses to communicate with physicians and pharmacists members to properly carry out the medication
about medication orders? administration process?
• Did all staff members involved properly follow the
communication protocols? Control of Medications: Storage/Access
• Were the medications in question stored in the accepted
Availability of Information manner?
• Does the hospital routinely supply information about • Were the medications accessed in the accepted manner?
medications?
• Did the staff members involved review all information Labeling of Medications
available to them? • Were the medications in question properly labeled?
• What processes or protocols are in place to verify that
Adequacy of Technological Support the label matched the medication?
• Are there any technologies in place to support the • Are there any protocols in place to ensure that “look-
medication administration process? alike” prescriptions are properly identified on the label?
• If “yes,” were these technologies properly used?
• If “no,” are there technologies available that would Sidebar 1-5, page 12, outlines the high-level key tasks
enhance the medication administration process? involved in performing a thorough and credible root cause
analysis and action plan. Overall, a thorough and credible
Equipment Maintenance/Management root cause analysis should do the following:
• Were all medication distribution systems (for example, • Be clear (understandable information)
medication cabinets) in working order? • Be accurate (validated information and data)
• How often are these systems maintained? • Be precise (objective information and data)
11
Sidebar 1-5.
• Be relevant (focus on issues related or potentially related
Conducting a Root Cause Analysis and
to the sentinel event)
Implementing an Action Plan • Be complete (cover all causes and potential causes)
• Be systematic (methodically conducted)
1. Assign an interdisciplinary team to assess the
sentinel event.
• Possess depth (ask and answer all of the relevant
“Why” questions)
2. Establish a way to communicate progress to senior
leadership. • Possess breadth of scope (cover all possible systemic
3. Create a high-level work plan with target dates, factors wherever they occur)
responsibilities, and measurement strategies.
4. Define all the issues clearly. The Joint Commission’s framework and JCI’s framework for
5. Brainstorm all possible or potential contributing a root cause analysis and action plan are similar and appear
causes and their interrelationships. as Figure 1-5 at the end of this chapter. This framework,
6. Sort and analyze the cause list. to be used extensively in Chapters 3 through 6, provides a
7. For each cause, determine which process(es) and solid foundation for root cause analyses and action plans.
system(s) it is a part of and the interrelationship of The tool selection matrix, found in Chapter 7 as Table 7-1
causes. (page 135), can also be used as a guide to ensure that an
8. Determine whether the causes are special causes, organization considers and selects the most appropriate
common causes, or both. tools and techniques for root cause analysis.
9. Begin designing and implementing changes while
finishing the root cause analysis. References
10. Assess the progress periodically. 1. Croteau RJ, Schyve PM. Proactively error-proofing health care
11. Repeat activities as needed (for example, processes. In Spath PL, editor: Error Reduction in Health Care: A
brainstorming). Systems Approach to Improving Patient Safety. San Francisco: Jossey-
12. Be thorough and credible. Bass, 2000, 179–198.
13. Focus improvements on the larger system(s). 2. The Joint Commission: Failure Mode and Effects Analysis in Health
Care: Proactive Risk Reduction, 3rd ed. Oak Brook, IL: Joint
14. Redesign to eliminate the root cause(s) and the
Commission Resources, 2010.
interrelationship of root causes that can create an
adverse outcome. 3. Reason J. Human error: Models and management. BMJ. 2000
Mar 18;320(7237):768–770.
15. Measure and assess the new design.
4. Knudsen P, et al. Preventing medication errors in community
pharmacy: Root-cause analysis of transcription errors. Qual Saf
Health Care. 2007 Aug;16(4):285–290.
5. Wu AW, Lipshutz AK, Pronovost PJ. Effectiveness and
efficiency of root cause analysis in medicine. JAMA. 2008 Feb
13;299(6):685–687.
12
Figure 1-5. A Framework for a Root Cause Analysis and Action Plan in Response to a Sentinel Event
A Framework for a Root Cause Analysis and Action Plan

in Response to a Sentinel Event
Page 1 of 4
Level of Analysis Questions Findings Root Ask Take
Cause? “Why?” Action
What happened? Sentinel Event What are the details of the
event? (Brief description)
When did the event occur?
(Date, day of week, time)
What area/service was
impacted?
Why did it The process or What are the steps in the
happen? activity in which process, as designed? (A
the event occurred flow diagram may be
helpful here.)
What were the What steps were involved
most proximate in (contributed to) the
factors? event?
13
(Typically “special Human factors What human factors were
cause” variation)
relevant to the outcome?
Equipment factors How did the equipment

performance affect the
outcome?
Controllable What factors directly
environmental affected the outcome?
factors
Uncontrollable Are they truly beyond the
external factors organization’s control?
Other Are there any other factors

that have directly
influenced this outcome?
What other areas or
services are impacted?
(continued)
Figure 1-5. A Framework for a Root Cause Analysis and Action Plan in Response to a Sentinel Event (continued)

in Response to a Sentinel Event (continued)
Page 2 of 4
This template is provided as an aid in organizing the steps in a root cause analysis. Not all possibilities and questions will apply in every
case, and there may be others that will emerge in the course of the analysis. However, all possibilities and questions should be fully
considered in your quest for “root cause” and risk reduction.
As an aid to avoiding “loose ends,” the three columns on the right are provided to be checked off for later reference:
• Root Cause? should be answered “yes” or “no” for each finding. A root cause is typically a finding related to a process or system that has a
potential for redesign to reduce risk. If a particular finding that is relevant to the event is not a root cause, be sure that it is addressed later in
the analysis with a “Why?” question. Each finding that is identified as a root cause should be considered for an action and addressed in the
action plan.
• Ask “Why?” should be checked off whenever it is reasonable to ask why the particular finding occurred (or didn’t occur when it should have) –
in other words, to drill down further. Each item checked in this column should be addressed later in the analysis with a “Why?” question. It is
expected that any significant findings that are not identified as root causes themselves have “roots.”
• Take action? should be checked for any finding that can reasonably be considered for a risk reduction strategy. Each item checked in this
column should be addressed later in the action plan. It will be helpful to write the number of the associated Action Item on page 4 of this tool
in the Take Action? column for each of the findings that requires an action.
14
Why did that Human Resources To what degree are staff
happen? What issues properly qualified and
systems and currently competent for
processes their responsibilities?
underlie those
proximate
factors?
(Common-cause How did actual staffing
variation here may
lead to special-cause
compare with ideal levels?
variation in dependent
processes)
What are the plans for
dealing with contingencies
that would tend to reduce
effective staffing levels?
To what degree is staff
performance in the operant
process(es) addressed?
(continued)

Page 3 of 4
How can orientation and
in-service training be
improved?
Information To what degree is all
management issues necessary information
available when needed?
Accurate? Complete?
Unambiguous?
To what degree is
communication among
participants adequate?
Environmental To what degree was the
management issues physical environment
appropriate for the pro-
15
cesses being carried out?
What systems are in place
to identify environmental
risks?
What emergency and
failure-mode responses have
been planned and tested?
Leadership issues: To what degree is the
- Corporate culture
culture conducive to risk
identification or reduction?
- Encouragement of What are the barriers to
communication
communication of
potential risk factors?
- Clear communication To what degree is the
of priorities
prevention of adverse
outcomes communicated
as a high priority? How?
Uncontrollable What can be done to protect
factors against the effects of these
uncontrollable factors?
(continued)

Page 4 of 4
Action Plan Risk Reduction Strategies Measures of Effectiveness
For each of the findings identified in the analysis Action Item #1:
as needing an action, indicate the planned action
expected, implementation date, and associated
measure of effectiveness. OR . . .
If after consideration of such a finding, a decision Action Item #2:

is made not to implement an associated risk
reduction strategy, indicate the rationale for not
taking action at this time.
Check to be sure that the selected measure will Action Item #3:
provide data that will permit assessment of the
effectiveness of the action.
Consider whether pilot testing of a planned Action Item #4:

improvement should be conducted.
16
Improvements to reduce risk should ultimately be Action Item #5:
implemented in all areas where applicable, not just
where the event occurred. Identify where the
improvements will be implemented.
Action Item #6:
Action Item #7:

Action Item #8:
Cite any books or journal articles that were considered in developing this analysis and action plan:
This framework outlines several questions that may be used to probe for systems problems underlying problematic processes. In each area,
consider whether and how the factors can be improved, as well as the pros and cons of expending resources to make improvements.
CHAPTER
2
Addressing Sentinel Events in
Policy and Strategy
Learning Objectives
• Become familiar with The Joint Commission’s Sentinel Events Policy and that of Joint Commission International
• Understand the differences among types of patient safety events
• Explore the components and value of a culture of safety in reducing variation in performance and the occurrence of
patient safety events
R oot cause analysis (RCA) plays a key role in the

identification and prevention of sentinel events. A
number of questions arise in considering sentinel events,
The Range of Adverse Events in Health Care
The Joint Commission’s Sentinel Event Policy is designed
to improve patient safety in all health care organizations
such as the following: by working with and learning from organizations that
• What is a sentinel event and how does it differ from experience serious adverse events in patient care. The
other events, incidents, or occurrences that take place policy encourages the self-reporting of sentinel events to
routinely in health care organizations? The Joint Commission that enables it to learn about the
• What role does an organization’s culture play in the relative frequencies and underlying cause(s) of sentinel
identification and prevention of adverse events? events and to share “lessons learned” with other health
• What do The Joint Commission and Joint Commission care organizations, thereby reducing the risk of future
International (JCI) require when a sentinel event occurs? sentinel events in other organizations. Accredited organi-
• What types of events require comprehensive systematic zations and those outside the United States are guided by
analysis, such as an RCA? the similar Sentinel Event Policy of JCI. A comparison of
• What issues should be considered as an organization the Joint Commission and JCI policies appears later in
develops its own sentinel event policy? this chapter.
• How should an organization respond after a
sentinel event? As of the 2015 revision of the Sentinel Event Policy,
• Who must be notified after a sentinel event occurs? The Joint Commission defines sentinel event as a patient
• What are the legal and ethical considerations of safety event (not primarily related to the natural course of
disclosure to patients? the patient’s illness or underlying condition) that reaches
a patient and results in any of the following:
Although the answers to these questions tend to be organi- • Death
zation-specific, some general guidelines can be useful. This • Permanent harm
chapter provides such guidelines, but organizations should • Severe temporary harm*
consult additional sources of information as needed.
* Severe temporary harm is critical, potentially life-threatening harm lasting for a limited time with no permanent residual, but requires transfer to a higher
level of care/monitoring for a prolonged period of time, transfer to a higher level of care for a life-threatening condition, or additional major surgery,
procedure, or treatment to resolve the condition. Adapted from Healthcare Performance Improvement, LLC. The HPI SEC & SSER Patient Safety
Measurement System for Healthcare, rev. 2. Throop C, Stockmeier C. May 2011. Accessed Oct 15, 2015. http://hpiresults.com/publications/HPI%20
White%20Paper%20-%20SEC%20&%20SSER%20Measurement%20System%20REV%202%20MAY%202011.pdf.
17
Such events are considered “sentinel” because they signal a Sidebar 2-1.
need for immediate investigation and response. All sentinel
Taxonomy of Patient Safety Events
events must be reviewed by the organization and are subject
to review by The Joint Commission. Accredited organiza- ►► Patient safety event: an event, incident, or condition
tions are expected to identify and respond appropriately to that could have resulted or did result in harm to a
all sentinel events (as defined by The Joint Commission) patient
occurring in the organization or associated with services
►► Adverse event: a patient safety event that resulted in
that the organization provides. An appropriate response harm to a patient
includes all of the following:
►► Sentinel event: a patient safety event (not primarily
• A formalized team response that stabilizes the patient,
related to the natural course of the patient’s illness
discloses the event to the patient and family, and
or underlying condition) that reaches a patient and
provides support for the family as well as staff involved results in death, permanent harm, or severe tempo-
in the event rary harm
• Notification of organization leadership
►► No-harm event: a patient safety event that reaches
• Immediate investigation
the patient but does not cause harm
• Completion of a comprehensive systematic analysis for
identifying the causal and contributory factors ►► Near miss (or close call): a patient safety event that
• Corrective actions did not reach the patient

• Time line for implementation of corrective actions ►► Hazardous (or unsafe) condition: a circumstance
• Systemic improvement (other than a patient’s own disease process or
condition) that increases the probability of an
Sentinel events are one type of patient safety event, which adverse event
is an event, incident, or condition that could have resulted
or did result in harm to a patient. Patient safety events also The patient could have suffered harm if the discrepancy had
include adverse events, no-harm events, close calls, and not been caught and perhaps would even be expected to
hazardous conditions. (See Sidebar 2-1, right, for defini- suffer harm with a similar error.
tions.) A patient safety event can be, but is not necessarily,
the result of a defective system or process design, a system The Joint Commission does not require health care organi-
breakdown, equipment failure, or human error. zations to conduct a comprehensive systematic analysis for a
near miss. However, doing so is a best practice. Using a root
Signals of Risk: Near Misses and cause analysis or other investigation methodology to iden-
No‑Harm Events tify the contributing factors to a near miss can help orga-
Fortunately, the majority of patient safety events cause no nizations address any risk points that could lead to patient
harm. For example, missed medication dosages or dosages harm down the road. This is one example of how RCA can
administered at the wrong time rarely result in death or be used proactively to prevent system failures or process
serious permanent or temporary harm. However, these variations from reaching your patients.
events may signal the presence of a much larger problem.
Organizations should integrate information about such The health care organization determines how it will respond
events as part of their ongoing data collection and analysis. to patient safety events that do not meet the definition of a
sentinel event. For example, adverse events should result in
A near miss also does not result in serious harm, but easily prompt notification of organization leaders, investigation,
could have. For example, suppose a medication was about and corrective actions, in accordance with the organization’s
to be administered to a patient via the incorrect route, such process for responding to patient safety events that do not
as intravenously rather than orally, when the nurse notices meet the definition of a sentinel event. An adverse event
the discrepancy in the order and holds off on giving the may or may not result from an error. (See the box on page
patient the medication so that the issue can be resolved. 19 for a discussion of error.)
18
CHAPTER 2 | Addressing Sentinel Events in Policy and Strategy
The Joint Commission’s Sentinel Event

TIP Policy
Track no-harm events, close calls, and hazardous When developing and implementing a sentinel event policy
conditions and use them as opportunities to learn. of their own, organizations should understand the Joint
(Leadership standards require the scope of the safety
Commission’s Sentinel Event Policy. Figure 2-1, page 21,
program include the full range of safety issues, from
potential or no-harm errors [sometimes referred
provides a graphic representation of the sentinel event
to as near misses, close calls, or good catches] to process flow.
hazardous conditions and sentinel events.)
The information provided here is current as of the time
of this book’s publication. The Sentinel Event Policy
Which Events Does the Joint Commission Review? is contained in the “Sentinel Events” chapter of the
The Sentinel Event Policy was updated extensively effec- Comprehensive Accreditation Manual for each accreditation
tive January 1, 2015. As of that date, all sentinel events are program (Hospital, Home Care, Laboratory, and so forth).
considered reviewable. This means that organizations are The accreditation manuals are updated twice annually,
required to conduct a comprehensive systematic analysis, which may include updates to the Sentinel Event Policy.
such as an RCA. Reporting the sentinel event to The Joint Changes to policy are reported in The Joint Commission
Commission is voluntary, but strongly encouraged. Prior Perspectives® which is accessible via your organization’s Joint
to 2015, The Joint Commission classified some sentinel Commission Connect™ extranet site.
events as “reviewable.” The 2015 revision eliminates
that distinction. Goals of the Sentinel Event Policy
The policy has the following four goals:
Also included in the 2015 revision is a list of patient safety 1. To have a positive impact in improving patient care,
events that are considered sentinel events regardless of treatment, and services and in preventing unintended
whether they result in death, permanent harm, or severe harm
temporary harm. Organizations should respond to these 2. To focus the attention of an organization that has
events the same way they would to any other sentinel experienced a sentinel event on understanding the
event. See Sidebar 2-2, page 20, to review several events factors that contributed to the event (such as underlying
from this list. causes, latent conditions and active failures in defense
Understanding Error
Though not all sentinel events result from errors, many do. These can include both errors of commission or errors of
omission. An error of commission occurs as a result of an action taken—for example, when surgery is performed on
the wrong limb, when a medication is administered by an incorrect route, when an infant is discharged to the wrong
family, or when a transfusion error occurs involving blood crossmatched for another patient. An error of commission
occurs as a result of an action not taken—for example, when a delayed diagnosis results in a patient’s death, when
a medication dose ordered is not given, when a physical therapy treatment is missed, or when a patient suicide is
associated with a lapse in carrying out frequent observation.
Errors of commission and omission may or may not lead to adverse outcomes. For example, suppose a patient
in seclusion is not monitored during the first two hours. The staff corrects the situation by beginning regular
observations as specified in organization policy. The possibility of failure is present, however, and the mere fact
that the staff does not follow organization policy regarding seclusion, and thereby violates acceptable professional
standards, signals the occurrence of a failure requiring study to ensure that it does not happen again. In this case,
the error of omission was insufficient monitoring. If the patient suffers serious physical or psychological harm
during seclusion, the sentinel event is the patient’s adverse outcome. By definition, sentinel events require further
investigation each time they occur.
19
Sidebar 2-2. Examples of Sentinel Events
All instances of the following types of patient safety events are considered sentinel, even if event did not cause death or
severe harm:
►► Suicide of any patient receiving care, treatment, and member, licensed independent practitioner, visitor, or
services in a staffed around-the-clock care setting vendor while on site at the organization
or within 72 hours of discharge, including from the
►► Invasive procedure, including surgery, on the wrong
hospital’s emergency department (ED)
patient, at the wrong site, or that is the wrong (unin-
►► Unanticipated death of a full-term infant tended) procedure
►► Discharge of an infant to the wrong family ►► Unintended retention of a foreign object in a patient
after an invasive procedure, including surgery
►► Abduction of any patient receiving care, treatment, and
services ►► Severe neonatal hyperbilirubinemia (bilirubin > 30
milligrams/deciliter)
►► Any elopement (that is, unauthorized departure) of a
patient from a staffed around-the-clock care setting ►► Prolonged fluoroscopy with cumulative dose > 1,500
(including the ED), leading to death, permanent harm, rads to a single field or any delivery of radiotherapy to
or severe temporary harm to the patient the wrong body region or > 25% above the planned
radiotherapy dose Fire, flame, or unanticipated smoke,
►► Hemolytic transfusion reaction involving administration
heat, or flashes occurring during an episode of patient
of blood or blood products having major blood group
care*
incompatibilities (ABO, Rh, other blood groups)
►► Any intrapartum (related to the birth process)
►► Rape, assault (leading to death, permanent harm, or
maternal death
severe temporary harm), or homicide of any patient
receiving care, treatment, or services while on site at ►► Severe maternal morbidity when it (not primarily
the organization related to the natural course of the patient’s illness or
underlying condition) reaches a patient and results
►► Rape, assault (leading to death, permanent harm,
in any of the following: Permanent harm or severe
or severe temporary harm), or homicide of a staff
temporary harm†
* Fire is defined as a rapid oxidation process, which is a chemical reaction resulting in the evolution of light and heat in varying intensities. A combustion
process that results in smoldering condition (no flame) is still classified as fire. Source: National Fire Protection Association (NFPA). NFPA 901: Standard
Classifications for Incident Reporting and Fire Protection Data. Quincy, MA: NFPA, 2011.
† Severe maternal morbidity is defined, by the American College of Obstetrics and Gynecology, the US Centers for Disease Control and Prevention, and
the Society for Maternal-Fetal Medicine, as a patient safety event that occurs intrapartum through the immediate postpartum period (24 hrs), that requires
the transfusion of four or more units of packed red blood cells and/or admission to the intensive care unit (ICU). Facilities are strongly encouraged to review
all cases of severe maternal morbidity for learning and improvement. Admission to the ICU is defined as admission to a unit that provides 24-hour medical
supervision and is able to provide mechanical ventilation or continuous vasoactive drug support.
systems, or hospital culture), and on changing the How The Joint Commission Becomes Aware
organization’s culture, systems, and processes to reduce of a Sentinel Event
the probability of such an event in the future Each accredited US health care organization is encouraged,
3. To increase the general knowledge about patient safety but not required, to report to The Joint Commission any
events, their contributing factors, and strategies for patient safety event that meets The Joint Commission’s
prevention definition of a sentinel event. Likewise, health care orga-
4. To maintain the confidence of the public and clinicians, nizations outside of the United States are encouraged to
and that patient safety is a priority in accredited report sentinel events to JCI. Alternatively, accreditor may
organizations become aware of a sentinel event by some other means,
such as communication from a patient, family member, or
employee of the organization or via media reports.
20
Figure 2-1. Sentinel Event Process Flow
Process Flow:
1) New sentinel event reported to The Joint Commission (via self-report, complaint, or referral from survey)
2) Create incident report
3) Sentinel event specialist contacts the organization to obtain additional details and make determination of
reviewability
a. Events not meeting reviewable criteria are forwarded to the CRU to request an organization response
b. Events meeting reviewable criteria

i. Alternative type is selected by the organization within 5 days and the due date is confirmed
1. Electronic submission
2. On-site educational session and review by a specially trained Joint Commission staff
member
a. On-site visit by the organization to the Joint Commission Central Office, with sharing
of the documation
b. On-site visit to the organization by one Joint Commission staff member, with sharing of
documation
c. O
n-site visit to the organization by one Joint Commission staff member, without
sharing of documation
3. One-day survey by a surveyor
4) Organization provides the RCA action plan information, within 45 days of having become aware of the
event, via the selected alternative type
5) RCA and action plan are reviewed for thoroughness and credibility
a. Not acceptable
i. The Joint Commission will work with the organization to facilitate development of a thorough
and credible RCA and action plan
ii. Clarifications requested and due within 15 days
b.
Acceptable
i. R equest 4 months of data
ii. Sentinel Event Measures of Success (SE MOS) due in 45 days
6) SE MOS reviewed for sustainability of the improvement efforts
a. Not acceptable
i. The Joint Commission will work with the organization to facilitate improvement efforts
ii. Request 4 additional months of data
b.
Acceptable
i. File closed as complete
ii. Letter posted to the CEO
7) Information de-identified and entered into the Sentinel Event Database
This outline displays the steps in the implementation of Sentinel Event Policy and its procedures.
(CRU, Complaint Response Unit; RCA, root cause analysis)
21
Voluntary Reporting of Sentinel Events to

Sidebar 2-3. Reasons to Call The Joint
The Joint Commission
If an organization wishes to report a sentinel event the
Commission’s Sentinel Event Hotline
organization is required to submit the information through
►► To discuss any hypothetical scenario regarding an
the secure and encrypted extranet (http://www.joint adverse event
commissionconnect.org/accreditation/SentinelEvent/self
►► Interpretation of any aspect of the Sentinel Event
_report.htm).
Policy
►► Education/guidance regarding the root cause anal-
Sentinel event inquiries may be directed anonymously to
ysis process
the Sentinel Event Hotline at 630-792-3700. Sidebar 2-3,
right, presents examples of reasons to call the hotline. ►► Education/guidance regarding the development of
risk reduction strategies and action plans
Each organization will be contacted by a Joint Commission
patient safety specialist to determine whether the event Further, reporting the event enables the addition of the
meets the definition of a sentinel event. After the event is “lessons learned” from the event to be added to The Joint
identified as a sentinel event, the specialist will confirm the Commission’s Sentinel Event Database, thereby contrib-
method for sharing the comprehensive systematic analysis uting to the general knowledge about sentinel events and to
and action plan and will confirm the due date. the reduction of risk for such events in many other hospi-
tals. (See Sidebar 2-4, page 23, for a discussion of govern-
Reasons for Reporting a Sentinel Event to ment reporting requirements.)
The Joint Commission
When an organization reports a sentinel event, The Joint How to Report a Sentinel Event
Commission or JCI is able to help it through the response Reporting a sentinel event can be done online with a
process. Self-reporting a sentinel event gives organizations form accessible on an organization’s Joint Commission
the opportunity to work with someone who may have Connect™ extranet site. Once logged in, place the cursor
more exposure to and more experience in addressing similar over “Continuous Compliance Tools.” A dropdown list will
issues. Joint Commission patient safety specialists work appear. From this list, select “Self Report Sentinel Event.”
with a wide spectrum of health care organizations on a (See a screenshot of this tool in Figure 2-2, page 24.)
regular basis to address many types of sentinel events. We
can lend valuable support to any and every organization Addressing Concerns About Reporting
that experiences a sentinel event. Concerns about potential legal issues can give an organiza-
tion pause before reporting. Often the chief concern is that
Any organization that experiences a sentinel event can a comprehensive systematic analysis will be obtained and
benefit from this review process. In particular, organiza- used in a lawsuit against the organization. Organizations
tions for which such events are very rare may see the most should be aware that it would be very difficult for any
benefit. In these organizations, the staff may be experi- parties to obtain access to that information under any
encing an event for the first time or for the first time in circumstances. In the event of a subpoena, The Joint
several years. The Joint Commission is a valuable collabora- Commission adheres to its Public Information Policy.
tive resource due to our cumulative knowledge of sentinel
events gleaned from working with many other organiza- Legal Concerns over Confidentiality
tions on similar incidents. The basic tenets of the Sentinel Event Policy are that an
organization must perform a comprehensive system-
The rewards of reporting sentinel events extend beyond the atic analysis in response to a sentinel event and that if
organization itself. Reporting sends a message to the public The Joint Commission becomes aware of the event, the
that the organization is doing all it can to prevent recur- organization must share relevant information with The
rence of a sentinel event. Joint Commission. Almost all organizations experiencing
sentinel events appear to be moving quickly to address
22
Sidebar 2-4. State and Federal Reporting Requirements
In addition to reporting sentinel events to The Joint Commission, many hospitals are required to report patient safety
events at the state level. Also many hospitals voluntarily report as Patient Safety Organizations that were established with
the passage of the Patient Safety and Quality Improvement Act of 2005 (http://www.pso.ahrq.gov).
The National Academy for State Health Policy (NASHP) collected information about all state adverse event reporting
systems. As of January 1, 2015, the shaded states were found to have reporting systems.
California New Hampshire

Colorado New Jersey
Connecticut New York
District of Columbia Ohio
Florida Oregon
Georgia Pennsylvania
Illinois Rhode Island
Indiana South Carolina
Kansas South Dakota
Maine Tennessee
Maryland Texas
Massachusetts Utah
Minnesota Vermont
Nevada Washington
NASHP made the following findings:

►► The number of adverse event reporting systems has changed for the first time since 2007, with Texas in January 2015
becoming the 28th state to adopt such a system. There continues to be wide variation in the types of individual events
reported to states. In 2014, 15 states have adopted or adapted the National Quality Forum’s list—a slight increase
from 2007.
►► Reporting systems are now more technologically advanced. Twenty-two systems are now partially or fully electronic,
compared to nine in 2007.
►► Communication of findings to providers and the public continues to be common, with 22 systems publicly reporting
data and 20 sharing system data with facilities. Eight states have increased the frequency of public reporting
since 2007.
►► Formal evaluations of reporting systems are rare (three states), however officials from most (23) systems have
necdotal, survey or other sources indicating an impact on communication among facilities, provider education,
a
internal agency tracking or trending, and/or implementation of facility processes to address quality of care. Nine
states report increased levels of provider and facility transparency and awareness of patient safety as a result of
their reporting systems.
►► System officials partner with provider, patient safety, and state agency representatives to carry out patient safety
initiatives. Despite potential opportunities in delivery system and payment reform, there are few specific examples
of states integrating adverse event reporting systems with statewide quality improvement or other initiatives that
demonstrate the importance of patient safety as a crosscutting statewide priority.
Source: Hanlon C, et al. 2014 Guide to State Adverse Event Reporting Systems. Portland, ME: National Academy for State Health Policy, 2015. Accessed Oct
16, 2015. http://www.nashp.org/sites/default/files/2014_Guide_to_State_Adverse_Event_Reporting_Systems.pdf.
23
the first tenet. However, among other reasons, The Joint

TIP Commission has been told that serious concerns regarding
If you are uncertain whether a patient safety event the potential discoverability of sentinel event–related
meets the definition of a sentinel event, err on the information shared with The Joint Commission have
side of caution by reporting the event. Self-reporting
created a significant barrier to organizations meeting the
does not automatically mean that an incident will be
considered a sentinel event.
second tenet.
If there are questions as to whether the event A number of years ago, in response to these concerns, The
meets the definition of a sentinel event, The Joint Joint Commission, after being advised by an outside group
Commission’s patient safety specialists can help of health care lawyers from various groups, including some
the organization make that determination. The Joint
state hospital association lawyers, offered several alterna-
Commission is there to help the organization through
the process, and patient safety specialists may work
tive ways for a health care organization to report, and The
with the organization to identify potential solutions Joint Commission to review, information regarding the
even if the incident does not meet the definition of a organization’s response to a sentinel event. These alterna-
sentinel event. tives, outlined on page 27, are intended to offer methods
for sharing information with The Joint Commission that
Figure 2-2. Sentinel Event Reporting Tool
Accredited health care organizations can use this tool to self-report a sentinel event via their Joint Commission Connect™
extranet site.
24
some lawyers and organizations may find provide additional such as through media reports or calls from patients.
comfort about any potential exposure of sensitive sentinel Whether the organization voluntarily reports the event or
event–related information. The Joint Commission becomes aware of the event by some
other means, there is no difference in the expected response,
Also, and again, with the advice of outside lawyers, The time frames, or review procedures.
Joint Commission agreed to sign contractual agreements,
the basic form of which is set out in Figure 2-3 on pages If The Joint Commission becomes aware of a sentinel
28–29, that some lawyers and organizations may also event that was not reported to The Joint Commission by
find helpful. These agreements emphasize that The Joint an accredited organization, the chief executive officer of
Commission should not be viewed as an external third the organization, or his or her designee, is contacted, and
party in the limited context of an intensive assessment of a preliminary assessment of the sentinel event is made. If
a sentinel event, and therefore no waiver of confidentiality the occurrences meets the definition of a sentinel event,
protections should occur by sharing sentinel event–related the organization is required to submit or make available an
information with The Joint Commission. acceptable root cause analysis and action plan, or choose an
approved protocol within 45 calendar days of the event or
Essentially, the agreements involve having the health care orga- becoming aware of the event.
nization and The Joint Commission agree to the following:
• The Joint Commission is appointed to the organization’s Disclosable Information
peer review or quality improvement activities. If, during the 45-day analysis period, The Joint
or Commission receives an inquiry about the accreditation
• The Joint Commission is appointed to the organization’s status of an organization that has experienced a sentinel
peer review or quality improvement committee. event, the organization’s accreditation status is reported in
the usual manner without making reference to the sentinel
Finally, The Joint Commission firmly believes that the event. If the inquirer specifically references the sentinel
sharing of information between The Joint Commission and event, The Joint Commission acknowledges that it is aware
accredited organizations should not waive confidentiality of the event and is working with the organization through
protection granted to any particular information by any the sentinel event review process. All RCAs and action plans
particular state law. If requested, The Joint Commission will be considered and treated as confidential by The Joint
would strongly make this point in any court or legislature. Commission in accordance with The Joint Commission’s
Although The Joint Commission cannot guarantee favorable Public Information Policy.
decisions on this point, at the time of the publication of
this book, The Joint Commission knows of no such waiver Required Response to a Reviewable
case in which confidential information was discovered Sentinel Event
or otherwise disclosed because it was shared with The If The Joint Commission becomes aware of a sentinel event
Joint Commission. in an accredited organization, and the occurrence meets the
definition of a sentinel event, the organization is required to
Questions regarding legal protections of sentinel event do the following:
information can be directed to the Office of the General • Prepare a thorough and credible comprehensive
Counsel. Contact information under “Legal Issues Relating systematic analysis and action plan within 45 business
to Accreditation” can be found online at http://www days of the event or of becoming aware of the event.
.jointcommission.org/AboutUs/ContactUs/contact_us Root cause analysis is the most common type of
_directory.htm. comprehensive systematic analysis used by Joint
Commission–accredited organizations.
Sentinel Events That Are Not Reported by • Select and provide information under an approved
the Organization alternative protocol within 45 business days of the
The Joint Commission sometimes becomes aware of a known occurrence of the event.
sentinel event through other means besides self-reporting,
25
Requirements Related to Sentinel Events

Joint Commission standards include the following requirements related to sentinel events and associated elements of
performance (EPs). The organization has an organizationwide, integrated patient safety program within its performance
improvement activities.
►► The leaders implement a organizationwide patient safety program.
►► One or more qualified individuals or an interdisciplinary group manages the safety program.
►► The scope of the safety program includes the full range of safety issues, from potential or no-harm errors (sometimes
referred to as near misses, close calls, or good catches) to hazardous conditions and sentinel events.
►► All departments, programs, and services within the organization participate in the safety program.
►► As part of the safety program, the leaders create procedures for responding to system or process failures.
►► The leaders provide and encourage the use of systems for blame-free internal reporting of a system or process
failure, or the results of a proactive risk assessment.
►► The leaders define patient safety event and communicate this definition throughout the organization.
►► At least every 18 months, the organization selects one high-risk process and conducts a proactive risk assessment.
►► To improve safety and to reduce the risk of medical errors, the organization analyzes and uses information about
system or process failures and the results of proactive risk assessments.
►► The leaders disseminate lessons learned from comprehensive systematic analyses (for example, root cause
analyses), system or process failures, and the results of proactive risk assessments to all staff who provide services
for the specific situation.
►► At least once a year, the leaders provide governance with written reports on the following:
• All system or process failures
• The number and type of sentinel events
• Whether the patients and the families were informed of the event
• All actions taken to improve safety, both proactively and in response to actual occurrences
• For organizations that use Joint Commission accreditation for deemed status purposes:
The determined number of distinct improvement projects to be conducted annually
• All results of the analyses related to the adequacy of staffing
►► The leaders encourage external reporting of significant adverse events, including voluntary reporting programs in
addition to mandatory programs.
►► The organized medical staff has a leadership role in organization performance improvement activities to improve
quality of care, treatment, and services and patient safety.
►► Information used as part of the performance improvement mechanisms, measurement, or assessment includes
the following:
• Sentinel event data.
• Patient safety data.
►► The organization respects the patient’s right to participate in decisions about his or her care, treatment, and services.
►► The organization informs the patient or surrogate decision-maker about unanticipated outcomes of care, treatment,
and services that relate to sentinel events as defined by The Joint Commission. (Refer to the Glossary for a definition
of sentinel event.)
►► The licensed independent practitioner responsible for managing the patient’s care, treatment, and services, or his or
her designee, informs the patient about unanticipated outcomes of care, treatment, and services related to sentinel
events when the patient is not already aware of the occurrence or when further discussion is needed.
26
The Joint Commission will then determine whether the 1. On-site visit by the organization to the Joint
root cause analysis and action plan are acceptable. If the Commission Central Office in Oakbrook Terrace,
determination that an event is sentinel occurs more than Illinois. During this visit, the root cause analysis and
45 business days following the known occurrence of the action plan are reviewed by The Joint Commission
event, the organization will be allowed 15 calendar days in a collaborative way. This collaboration may result
for its response. If the root cause analysis or action plan is in an enhancement of the RCA and action plan with
not acceptable, The Joint Commission will work with the further risk reduction strategies as discussed during
organization to facilitate development of a thorough and the visit. This can also be performed via Web-based
credible root cause analysis and action plan. video conferencing with a patient safety specialist
who is located at The Joint Commission. When the
Submission of Root Cause Analysis and Web-based videoconference is used, the hospital’s
Action Plan participants remain at the hospital.
An organization that experiences a sentinel event is required
to complete two documents: (1) the root cause analysis, 2. On-site visit to the organization by one Joint
which includes enough detail to demonstrate that the anal- Commission staff member, with sharing of
ysis is thorough and credible, and (2) the resulting action documentation. As with option 1, during this visit,
plan that describes the organization’s risk reduction strate- the RCA and action plan are reviewed by The Joint
gies and plan for evaluating their effectiveness. Commission in a collaborative way. This collaboration
may result in an enhancement of the RCA and action
A framework for a root cause analysis and action plan (see plan with further risk reduction strategies as discussed
Figure 1-5, pages 13–16) is available to organizations as during the visit. Please note that no survey activities
an aid in organizing the steps in a root cause analysis and are conducted during this visit.
developing an action plan. It is also available on The Joint
Commission’s website at http://www.jointcommission.org. 3. On-site visit to the organization by one Joint
The root cause analysis and action plan are not to include Commission staff member, without sharing of
the patient’s name or the names of caregivers involved in documentation. During this visit, the organization
the sentinel event. discusses the sentinel event with The Joint Commis
sion but does not show its root cause analysis and
Reporting Options action plan. Please note that no survey activities are
The handling of sensitive documents regarding review- conducted during this visit.
able sentinel events is restricted to specially trained Joint
Commission staff in accordance with procedures designed 4. On-site visit by a specially trained surveyor
to protect the confidentiality of the documents. An organi- arranged to conduct the following:
zation that has experienced a reviewable sentinel event has a. Interview and review of relevant documentation to
several alternate methods for sharing information with The evaluate the following:
Joint Commission. These sharing options include electronic • The process the hospital uses in responding to
submission and on-site interaction, described as follows. sentinel events
• Electronic submission. The organization submits its • The relevant policies and procedures preceding
root cause analysis and action plan via the secure and and following the organization’s review of the
encrypted extranet, The Joint Commission Connect™. specific event, and the implementation thereof,
The organization will be contacted by telephone upon sufficient to permit inferences about the
completion of The Joint Commission’s review. adequacy of the organization’s response to the
sentinel event
• On-site interaction. This option involves a face- to- b. A standards-based survey that traces a patient’s
face dialogue between the organization and The Joint care, treatment, and services and the organization’s
Commission. It may take place in one of four ways: management functions relevant to the sentinel
event under review
27
Figure 2-3. Contractual Agreement Example
(continued)
28
Figure 2-3. Contractual Agreement Example (continued)
This is the basic contractual agreement signed by The Joint Commission and an accredited hospital regarding
information shared with The Joint Commission about the hospital’s response to a sentinel event.
29
The Joint Commission’s Review of the Root Cause Aggregate data relating to root causes and risk reduc-
Analysis and Action Plan tion strategies for sentinel events that occur with signifi-
Joint Commission staff members assess the acceptability of cant frequency form the basis for future error-prevention
an organization’s response to a reviewable sentinel event, advice to health care organizations through Sentinel Event
including the thoroughness and credibility of any root Alerts, National Patient Safety Goals, and other methods
cause analysis information reviewed and the organization’s of information sharing. In handling sentinel event–related
action plan. documents, The Joint Commission destroys original root
cause analysis documents and any copies, or, upon request,
Follow-Up Activity the original documents are returned to the organization.
Through the completion of the root cause analysis, the Handling of these sensitive documents is restricted to
organization is able to identify important contributing specially trained Joint Commission staff in accordance
factors to a sentinel event. An action plan is developed to with procedures designed to protect the confidentiality of
mitigate the possible recurrence of such an event. The Joint the documents.
Commission requires that organizations measure the effec-
tiveness of those improvement activities. This measurement The action plan resulting from the analysis of the sentinel
is most frequently done through a Sentinel Event Measure event is initially retained to serve as the basis for the
of Success (SE MOS). An SE MOS is a numerical or quan- follow-up activity. After the action plan has been imple-
tifiable measure usually related to an audit that determines mented to the satisfaction of The Joint Commission,
if a planned action was effective and sustained. (In some as determined through follow-up activities, The Joint
instances, The Joint Commission will consider other forms Commission destroys the action plan.
of mutually agreed-upon follow-up activities to evaluate the
effectiveness of an action plan. ) Joint Commission International’s
Sentinel Event Policy
Upon submission of SE MOS data indicating sustain- The JCI Sentinel Event Policy, applicable to accredited
ability of the improvement efforts, the file will be closed organizations outside the United States, has a great deal of
as complete. The information will be de-identified and similarity to that of The Joint Commission.1 (For a compar-
entered into the Sentinel Event Database, as discussed in ison of the two sentinel event policies, see Table 2-1 on pages
the following section. 32–33) The JCI Sentinel Event Policy has four similar goals:
1. To have a positive impact in improving patient care,
The Sentinel Event Database treatment, and services and preventing sentinel events
To achieve the third goal of the Sentinel Event Policy— 2. To focus the attention of an organization that has
to increase the general knowledge about sentinel events, experienced a sentinel event on understanding the
their causes, and strategies for prevention—The Joint causes that underlie that event and on changing the
Commission collects and analyzes data from the review organization’s systems and processes to reduce the
of sentinel events, root cause analyses, action plans, and probability of such an event in the future
follow-up activities. These data and information form 3. To increase general knowledge about sentinel events,
the content of The Joint Commission’s Sentinel Event their causes, and strategies for prevention
Database. 4. To maintain the confidence of the public and
internationally accredited organizations in the
In response to concerns about potential increased legal accreditation process
exposure for accredited organizations through the sharing
of such information with The Joint Commission, The Joint Reporting of sentinel events to JCI is voluntary but encour-
Commission has committed to the development and main- aged. As stated in the Sentinel Event Policy, JCI’s definition
tenance of this Sentinel Event Database in a fashion that of a reviewable sentinel event takes into account a wide
excludes organization, caregiver, and patient identifiers. array of occurrences applicable to a wide variety of health
care organizations. The following sentinel events are subject
30
to review by JCI and include any occurrence that meets the Sidebar 2-5.
following criteria:
National Mandatory Reporting Systems
• The event has resulted in an unanticipated death or
major permanent loss of function.
for Selected Countries
or
Czech Republic: Health care professionals are
• The event resulted from wrong-site, wrong-patient, required by the government to report such events as
wrong-procedure surgery. adverse drug reactions, medical equipment failures,
nosocomial (health care–associated) infections, and
If, either through voluntary reporting by the accredited transfusion reactions. Information is aggregated at
organization or through other means (such as communica- levels ranging from hospital and medical specialization
tions from a patient, a patient’s family member, employees, through regional and national. Reports are not acces-
or the media), JCI becomes aware of the occurrence of a sible to the public.
reviewable sentinel event, then the event becomes subject Denmark: Under the Act on Patient Safety in the
to JCI’s Sentinel Event Policy. In that case, the organiza- Danish Health Care System that took effect January 1,
tion is expected to prepare a thorough and credible action 2004, health care professionals who become aware of
plan within 45 calendar days of the event or of becoming an adverse event resulting from a patient’s treatment at
or stay in the hospital are required to report the event
aware of the event and then submit its root cause analysis
to a national database. After analysis by the county in
and action plan to JCI, which evaluates them and deter- which the event occurred, the report is de-identified and
mines whether they are acceptable. All root cause analyses forwarded to the National Board of Health, which issues
and action plans are considered and treated as confidential periodic alerts and publishes an annual report.
by JCI.
Ireland: Under the Clinical Indemnity Scheme (CIS)
law established in 2002, all covered enterprises are
A root cause analysis will be considered acceptable by JCI if required to report adverse events and near misses
it has the following characteristics: related to clinical care. The report data are aggre-
• The analysis focuses primarily on systems and processes, gated through a secure Web-based system to detect
not individual performance. trends. The CIS maintains a website, disseminates a
• The analysis progresses from specific causes in the clinical quarterly newsletter, and conducts seminars to share
care process to common causes in the organizational lessons learned.
process. Japan: The Ministry of Health, Labour and Welfare
• The analysis repeatedly digs deeper. does not directly administer its own reporting system,
• The analysis identifies changes that could be made but it requires that hospitals have internal reporting
in systems and processes (either through redesign or systems. Teaching hospitals submit mandatory reports
of patient injuries, equipment failures, and near misses
development of new systems or processes) that would
to the Japan Council for Quality Health Care imple-
reduce the risk of such events occurring in the future. mented a private national reporting system in 2004.
Reporting by other types of hospitals is voluntary. The
Although reporting a sentinel event to JCI is voluntary, council aggregates data and distributes summary
an organization may be required to report sentinel events reports to health care providers and the public.
to a governmental or private agency within its country.
Source: World Health Organization (WHO). WHO Draft Guidelines for
Sidebar 2-5, below, describes some of these national Adverse Event Reporting and Learning Systems: From Information to
reporting systems. Action. Geneva: WHO Press, 2005.
31
Table 2-1. Comparison of The Joint Commission and Joint Commission International
Sentinel Event Policies
The Joint Commission Sentinel Joint Commission International (JCI)

Event Policy Sentinel Event Policy
Definition of A patient safety event (not primarily related to An unanticipated occurrence involving death or
Sentinel Event the natural course of the patient’s illness or serious physical or psychological injury.
underlying condition) that reaches a patient and
results in any of the following:
• Death
• Permanent harm
• Severe temporary harm
Other Information Such events are called “sentinel” because they Such events are called “sentinel” because they
About Sentinel signal the need for immediate investigation and signal a need for immediate investigation and
Events response. The terms sentinel event and medical response. The terms sentinel event and medical
error are not synonymous; not all sentinel events error are not synonymous; not all sentinel events
occur because of an error and not all errors result occur because of an error and not all errors result
in sentinel events. in sentinel events. A sentinel event may occur due
to wrong-site, wrong-procedure, wrong-patient
surgery.
Goals of the 1. To have a positive impact in improving 1. To have a positive impact in improving patient
Sentinel Event patient care, treatment, and services and in care, treatment, and services and preventing
Policy preventing unintended harm sentinel events
2. To focus the attention of a hospital that has 2. To focus the attention of an organization
experienced a sentinel event on understanding that has experienced a sentinel event on
the factors that contributed to the event (such understanding the causes that underlie that
as underlying causes, latent conditions and event and on changing the organization’s
active failures in defense systems, or hospital systems and processes to reduce the
culture), and on changing the hospital’s probability of such an event in the future
culture, systems, and processes to reduce the 3. To increase general knowledge about sentinel
probability of such an event in the future events, their causes, and strategies for
3. To increase the general knowledge about prevention
patient safety events, their contributing factors, 4. To maintain the confidence of the public and
and strategies for prevention internationally accredited organizations in the
4. To maintain the confidence of the public, accreditation process
clinicians, and hospitals that patient safety is a
priority in accredited hospitals
Organization- Standards require each accredited organization Standards require each accredited organization
specific Definition of to define patient safety event and communicate to establish which unanticipated events are
Sentinel Event this definition throughout the organization. At significant and the process for their intense
a minimum, an organization’s definition must analysis.
encompass sentinel events as defined by The While the determination of what constitutes a
Joint Commission. An accredited organization significant event must be consistent with the
is encouraged to include in its definition events, general definition of sentinel event as described
incidents, and conditions in which no or only in this policy, accredited organizations have some
minor harm occurred to a patient. The hospital latitude in setting more specific parameters to
determines how it will respond to patient safety define unanticipated and major permanent loss
events that do not meet the Joint Commission of function. At a minimum, an organization must
definition of a sentinel event. include those events that are subject to review
under and Sentinel Event Policy.
(continued)
32
Table 2-1. Comparison of The Joint Commission and Joint Commission International
Sentinel Event Policies (continued)
The Joint Commission Sentinel Joint Commission International (JCI)

Event Policy Sentinel Event Policy
Self-Reporting a Encouraged but not required Encouraged but not required

Sentinel Event
Reviewable Sentinel All sentinel events must be reviewed by the Any event that has resulted in an unanticipated
Events organization and are subject to review by The death or major permanent loss of function or
Joint Commission. from wrong-site, wrong-patient, wrong-procedure
surgery
Expectations for Prepare a thorough and credible root cause Identify and respond appropriately to all sentinel
an Organization’s analysis and action plan within 45 business days events. Appropriate response includes conducting
Response to a of the event or of becoming aware of the event. a timely, thorough, and credible root cause
Sentinel Event analysis; developing an action plan designed
to implement improvements to reduce risk;
implementing the improvements; and monitoring
the effectiveness of the improvements. When JCI
becomes aware of a reviewable sentinel event
in an accredited organization, the organization
is expected to prepare a thorough and credible
action plan within 45 calendar days of the event
or of becoming aware of the event and to submit
to JCI its root cause analysis and action plan
or otherwise provide for JCI evaluation of its
response to the sentinel event.
Confidentiality All root cause analyses and action plans will be All root cause analyses and action plans will be
considered and treated as confidential by The considered and treated as confidential by JCI.
Joint Commission in accordance with the Joint
Commission’s Public Information Policy.
Follow-Up to Organizations are required to measure the Upon acceptance of the organization’s root
Review effectiveness of the improvement activities cause analysis and action plan, JCI will assign an
outlined in the action plan accepted by The Joint appropriate follow-up activity, typically a written
Commission. This measurement is done through progress report due in four months.
a Sentinel Event Measure of Success (SE
MOS) submission. If the SE MOS data indicate
sustainability of improvement efforts, the file will
be closed as complete.
Related Joint Commission International the implementation of the organization’s quality and
Standards patient safety program.
Quality Improvement and Patient Safety (QPS) standards: • The individual(s) with oversight for the quality program
require that qualified individuals guide the implementation selects and supports qualified staff for the program
of the organization’s program for quality improvement and and supports those staff with quality and patient safety
patient safety and manage the activities needed to carry out responsibilities throughout the organization.
an effective program of continuous quality improvement • The quality program provides support and coordination
and patient safety within the organization. to department/service leaders for like measures across
the organization and for the organization’s priorities
Additional requirements include the following: for improvement.
• An individual(s) who is experienced in the methods –– The quality program implements a training program
and processes of improvement is selected to guide for all staff that is consistent with staff’s roles in the
33
quality improvement and patient safety program. receive and act on reports of the quality and patient
–– The quality program is responsible for the regular safety program.
communication of quality issues to all staff. –– Those responsible for governance annually approve
• The hospital uses a defined process for identifying and the hospital’s program for quality and patient safety.
managing sentinel events. –– Those responsible for governance at least quarterly
• Hospital leadership has established a definition of a receive and act on reports of the quality and patient
sentinel event that at least includes a) through f ) found safety program, including reports of adverse and
in the intent. sentinel events.
• The hospital completes a root cause analysis of all –– Minutes reflect actions taken and any follow-up on
sentinel events in a time period specified by hospital those actions.
leadership that does not exceed 45 days from the date of • Hospital leadership communicates quality improvement
the event or when made aware of the event. and patient safety information to governance and
• Hospital leadership takes action on the results of the root hospital staff on a regular basis.
cause analysis. –– Hospital leadership reports on the quality and patient
• Those responsible for governance approve the hospital’s safety program quarterly to governance.
program for quality and patient safety and regularly –– Hospital leadership reports to governance include, at
Why a Defined Process?

A sentinel event is an unanticipated occurrence involving death or serious physical or psychological injury.
Serious physical injury specifically includes loss of limb or function. Such events are called sentinel because they
signal the need for immediate investigation and response. Each hospital establishes an operational definition of
a sentinel event that includes at least
a) an unanticipated death, including, but not limited to,
• death that is unrelated to the natural course of the patient’s illness or underlying condition (for example,
death from a postoperative infection or a hospital-acquired pulmonary embolism);
• death of a full-term infant; and
• suicide;
b) major permanent loss of function unrelated to the patient’s natural course of illness or underlying condition;
c) wrong-site, wrong-procedure, wrong-patient surgery;
d) transmission of a chronic or fatal disease or illness as a result of infusing blood or blood products or trans-
planting contaminated organs or tissues;
e) infant abduction or an infant sent home with the wrong parents; and
f) rape, workplace violence such as assault (leading to death or permanent loss of function); or homicide
(willful killing) of a patient, staff member, practitioner, medical student, trainee, visitor, or vendor while on
hospital property.
The hospital’s definition of a sentinel event includes a) through f) above and may include other events as required
by laws or regulations or viewed by the hospital as appropriate to add to its list of sentinel events. All events that
meet the definition of sentinel event must be assessed by performing a credible root cause analysis. Accurate
details of the event are essential to a credible root cause analysis, thus the root cause analysis needs to be
performed as soon after the event as possible. The analysis and action plan is completed within 45 days of the
event or becoming aware of the event. The goal of performing a root cause analysis is for the hospital to better
understand the origins of the event. When the root cause analysis reveals that systems improvements or other
actions can prevent or reduce the risk of such sentinel events recurring, the hospital redesigns the processes and
takes whatever other actions are appropriate to do so.
It is important to note that the terms sentinel event and medical error are not synonymous. Not all errors result in
a sentinel event, nor does a sentinel event occur only as a result of an error. Identifying an incident as a sentinel
event is not an indicator of legal liability.
34
least once every six months, the number and type of • All unusual events, even though they may result in only
sentinel events and root causes, whether the patients minor adverse outcomes
and families were informed of the sentinel event, • All events that must be reported to an external agency
actions taken to improve safety in response to sentinel • Events with potential for an adverse public, economic, or
events, and if the improvements were sustained. regulatory impact
–– Information on the quality improvement and patient
safety program is regularly communicated to staff, An organization’s definition should, at a minimum, include
including progress on meeting the International those events that meet the definition of a sentinel event
Patient Safety Goals. under The Joint Commission’s or JCI’s Sentinel Event
• The hospital integrates the human subjects research Policy. The definition should also apply organizationwide
program into the quality and patient safety program of and should appear in writing in an organization plan or
the hospital. policy. Through a collaborative process, organization leaders
• The research program is a component of the hospital’s as well as medical, nursing, and administrative staff should
processes to report and act on sentinel events, adverse develop the definitions or categories of events that warrant
events of other types, and the processes to learn from root cause analysis.
near misses.
• The research program is included in the hospital’s In developing the organization’s sentinel event policy,
programs for hazardous material management, medical leaders may also address the process for reporting a sentinel
technology management, and medication management. event to leadership, how organizations should handle near
• The evaluation of staff participating in the research misses, and the process for the ongoing management of
program is incorporated into the ongoing monitoring sentinel events and prevention efforts. Leaders may also
processes of professional performance. wish to identify the following:
• The individual responsible for receiving initial
Developing Your Own Sentinel Event Policy notification of a sentinel event
Both US–based and international health care organizations • The individual responsible for assessing whether or
should consider developing their own sentinel event policy. not the event warrants an in-depth root cause analysis
The first step in developing an organization’s sentinel event based on the organization’s definition of a sentinel
policy* is to determine which events warrant root cause event (this may be the same individual—for example,
analysis. The Joint Commission expects accredited organi- a physician, risk manager, quality assurance coordinator,
zations to identify and respond appropriately to all sentinel or program manager)
events, as defined by The Joint Commission, occurring in • How this individual communicates the need for
the organization or associated with services that the orga- in-depth investigation and necessary information to a
nization provides or provides for. Doing so helps ensure team of individuals responsible for performing the root
improvement of the organization’s processes. As outlined cause analysis
earlier, appropriate response includes a thorough and cred- • The individual responsible for facilitating and overseeing
ible RCA, implementation of improvements to reduce risk, a team-based root cause analysis process
and monitoring of the effectiveness of those improvements.
Addressing Discoverability and Disclosure
An organization may wish to define a sentinel event as a Leaders should also address confidentiality, discoverability,
serious event involving staff and visitors as well as patients. and disclosure. Information obtained during the inves
Or an organization may wish to include the following: tigation of sentinel events through RCA or other techniques
is often highly sensitive. The organization’s sentinel event
* Please note that although The Joint Commission requires organizations to policy should address how confidentiality will be protected.
define the term sentinel event and communicate its definition throughout The policy should also address the procedure for obtaining
the organization, The Joint Commission does not require organizations to legal consultation to protect relevant documents such as
develop a sentinel event policy. The information provided here is intended
to provide advice on how an organization might develop a sentinel event meeting minutes, reports, and conversations from discovery
policy if it wishes to do so. in the event of a future lawsuit. The policy should be clear
35
on whether the state in which the organization operates Sidebar 2-6.

protects the details of a sentinel event investigation from
Steps in Developing an Organization’s
discovery under the organization’s quality management,
peer review, or risk management programs.
Sentinel Event Policy
1. Define the term sentinel event, setting specific
Following the development of the organization’s sentinel parameters for what constitutes unexpected,
event policy, leaders should ensure that all staff and physi- serious, and the risk thereof. Remember that the
cians are knowledgeable about the organization’s sentinel general definition must be consistent with The Joint
event policies and procedures. In-service programs and new Commission’s definition.
staff and physician orientation should address the organi- 2. Include the definition of sentinel event in writing in
zation’s sentinel event policy on a regular and continuing an organization plan or policy.
basis, including regular updates concerning sentinel events 3. Determine which events warrant root cause
published by The Joint Commission and additional issues analysis using a collaborative process.
that may be identified through the organization’s compli- 4. Determine the process for reporting a sentinel
ance with the National Patient Safety Goals. Sidebar 2-6, event to leadership.
right, outlines the steps described in this section. 5. Determine the process for reporting the event to
external agencies.
Leadership, Culture, and Sentinel Events 6. Determine how near misses are to be handled.
For the sentinel event policy to be effective, the culture of 7. Determine how the ongoing management of
an organization and the role of its leaders need to support sentinel events and prevention efforts.
the prevention and identification of sentinel events. 8. Address how confidentiality of information related
to sentinel events will be protected.
9. Address the procedure for obtaining legal
Leaders must be deeply committed to patient safety and to
consultation to protect relevant documents.
ensuring that members of their organization truly embody
10. E
ducate all staff and physicians about the policy
their mission, vision, and values. They play a critical role
and procedures; ensure ongoing education.
in fostering an organization culture in which sentinel event
11. Review the policy annually and revise information,
reporting, root cause analyses, and proactive risk reduction
the process for reporting sentinel events, confiden-
are encouraged. Reporting helps an organization start the tiality issues, legal issues, and relevant staff educa-
process of both identifying root causes and developing and tion, as appropriate.
implementing risk reduction strategies. Understanding that
continuous improvement is essential to an organization’s
success, leaders must have the authority and willingness to sentinel events has been to identify the individual(s) most
allocate resources for root cause analyses and improvement closely associated with the event and take some form of
initiatives. They must ensure that the processes for identi- punitive action. “Who did it?” has too often been the first
fying and managing sentinel events are defined and imple- question asked, rather than “Why did it happen?”
mented. They must be willing and able to set an example
for the organization and empower staff to identify and The organization that routinely asks “Why?” rather than
bring about necessary change. Effective leaders empower “Who?” will, over time, learn more about the quality and
staff throughout the organization to acquire and apply the safety of the care it is providing—as well as its sentinel
knowledge and skills needed to continuously improve events, near misses, and hazardous conditions—and
processes and services. will better understand the relationship between its staff
members and the processes, systems, and environment in
An organization’s response to a sentinel event speaks which they function.2
volumes about the culture of that organization. Its response
can also significantly influence the likelihood of similar Through commitment to performance improvement,
events occurring in the future. Historically, the response to patient safety, and proactive risk reduction, leaders build an
36
organization culture that values change, creativity, team- Preservation of Evidence

work, and communication. Teams provide much of the To learn from a failure and understand why it occurred,
impetus for performance improvement. Communication it is critical to know exactly what occurred. Preserving the
and information flow throughout the organization to foster evidence is essential to this process. Immediate steps should
a barrier-free learning environment. be taken to secure any biological specimens, medications,
equipment, medical or other records, and any other mate-
Early Response Strategies rial that might be relevant to investigating the failure.3 In
An organization has just experienced a sentinel event medication-related events, syringes of recently used medi-
leading to a serious adverse outcome. What must be done? cations and bottles of medications should be preserved and
Following the occurrence of a sentinel event, staff members sequestered. Because such evidence may be discarded as a
must simultaneously take a number of actions. An organi- part of routine operations, such as when empty vials are
zation’s sentinel event policy should outline early response thrown into trash cans, it must be obtained and preserved
strategies. These strategies include the following: promptly. Protocols established by the health care organiza-
• Providing prompt and appropriate care for the tion should specify the steps to be taken to preserve relevant
affected patient(s) evidence following a sentinel event.
• Containing the risk of an immediate recurrence of
the event Event Investigation
• Preserving the evidence Documentation and appropriate communication and
disclosure to relevant parties must also be considered imme-
Appropriate Care diately following the occurrence of a sentinel event.
The prompt and proper care of a patient who has been
affected by a sentinel event should be the providers’ and Documentation
staff members’ first concern following the event. Care could Proper medical record documentation of errors or sentinel
involve, as appropriate, stabilizing the patient, arranging events is critical for the continuity of care. Documentation
for his or her transportation to a health care facility for tips appear in Sidebar 2-7, page 38. A thorough incident-
surgery or testing, providing medications, taking actions reporting form can be very helpful during the early stages
to prevent further harm, and reversing the harm that has of event investigation and during Steps 2 through 4 of root
occurred, if possible. When appropriate, physicians should cause analysis (discussed in Chapters 3 and 4). Health care
obtain medical consultation related to the adverse event and organizations use a variety of occurrence or incident-re-
arrange to receive necessary follow-up information. porting tools and generally have a policy and procedure
covering their use. Forms or questionnaires may be general
Risk Containment in nature, covering all types of adverse events, or be specific
Following a sentinel event, the organization must respond to event types.
by immediately containing the risk of the event occur-
ring again. For example, if a patient suffered a stroke Communication and Disclosure
after receiving an incorrect drug that looked very similar With the occurrence of a sentinel event, personnel involved
to the correct drug and was stored in the same area, are in the incident should promptly notify those responsible for
other patients at risk for the same type of incident? If so, error reporting and investigation within the organization.
the organization must take immediate action to safeguard Supervisors, quality and risk management professionals,
patients from a repetition of the unwanted occurrence. In and administrators should be informed. These individuals
the example given, one of the two drugs would immedi- can determine how best to notify other parties, including
ately be moved to a separate storage area farther away from the press and external agencies such as federal, state, and
the other. Risk management texts, articles appearing in the local authorities. Legal counsel should be sought early
literature, and associations such as the American Society in the process. Counsel can provide guidance on how to
for Healthcare Risk Management (ASHRM) can provide discuss the situation with the patient and family, how to
detailed guidance. Contact ASHRM by phone at prevent disclosure of potentially defamatory information,
312-422-3980 or go online at http://www.ashrm.org. and how to handle media relations.4
37
Sidebar 2-7. Conferring with other members of the care team following
an adverse event enables the provider to clarify factual
Tips for Documenting Adverse Events
details and the proper sequence of what occurred. It can
in the Medical Record also help identify what needs to be done in response.
►► Assign the most involved and knowledgeable
member(s) to record factual statements of the event
Good communication between providers and patients is
in the patient’s record and any follow-up needed or instrumental in achieving positive care outcomes. Yet health
done as a result of the incident. care professionals often do not tell patients or families
►► Avoid writing in the record any information unrelated about their mistakes. Fear of malpractice litigation and
to the care of the patient (for example, “incident the myth of perfect performance reinforce poor provider
report filed” or “legal office notified”). communication of errors to patients and their families.
There is little doubt that the current malpractice crisis is
Source: Keyes C. Responding to an adverse event. Forum. 1997 Apr;
18(1):2–3. Updated on CRICO/RMF website. Accessed Oct 17, 2017. a deterrent to the openness required for quality improve-
https://www.rmf.harvard.edu/~/media/Files/_Global/KC/Forums/1997 ment.6 However, errors not communicated to patients,
/ForumApr1997.pdf#page=2. Forum is published by Risk Management
Foundation of the Harvard Medical Institutions. Reproduced by families, fellow staff members, and organizations are errors
permission. that do not contribute to systems improvement.
One recommendation states that organizations maintain Disclosing mistakes to patients and their families is difficult
two lists of key people to contact following a sentinel event: at best. Yet legal and ethical experts generally advise practi-
(1) key individuals within the organization and (2) indi- tioners to disclose mistakes to patients and their families in
viduals outside the organization.5 Both lists should be kept as open, honest, and forthright a manner as possible. One
up-to-date with current telephone numbers and should be suggestion is that physicians have an ethical duty to disclose
accessible to managers, supervisors, and members of a crisis errors when the adverse event does one of the following7:
management team. A sample sentinel event notification • Has a perceptible effect on the patient that was not
checklist appears as Figure 2-4, page 39. discussed in advance as a known risk
• Necessitates a change in the patient’s care
Responding to media queries through organization proto- • Potentially poses an important risk to the patient’s
cols helps avert complications related to patient confiden- future health
tiality, legal discovery, and heat-of-the-moment coverage.5 • Involves providing a treatment or procedure without the
Notification requirements should be reflected in organiza- patient’s consent
tion policies and procedures. They should include policies
for communication with the patient and family. This idea maintains that disclosure of a mistake may foster
learning by compelling the physician to acknowledge it
A provider’s communication and disclosure with relevant truthfully and that the physician-patient relationship can
parties following the occurrence of an event that led to or be enhanced by honesty.8 Disclosing a mistake might
could have led to patient injury is critical. Relevant parties even reduce the risk of litigation if the patient appreciates
include the following: the physician’s honesty and fallibility as a fellow human
• Patients and families affected by the event being. One study reports that the risk of litigation nearly
• Colleagues who could provide clarification, support, and doubles when patients are not informed by their physicians
the opportunity to learn from the error of moderately serious mistakes.9 Physician guidelines in
• The health care organization’s and individual provider’s disclosing medical mistakes to patients are offered by yet
liability insurers another report.10 Sidebar 2-8, page 40, outlines practical
• Appropriate organization staff, including risk managers issues a physician may encounter in disclosing an error to
or quality assurance representatives a patient or his or her family. Please note that these are
• Others who could provide emotional support or guidelines about issues to consider, not Joint Commission
problem-solving help requirements. Organizations should be aware that the
disclosure of an error or event requires individualized
38
Figure 2-4. Sentinel Event Notification Checklist

Sentinel Event
Occurrence: Date and Time:
Contact Person:
Chief Executive Officer

Name: Office Phone Number:
After-hours Phone Number: Name/Phone of Designated Backup:
Notified by: Date and Time:
Results of Contact:
Chief Nursing Officer

After-hours Phone Number: Notified by:
Date and Time: Results of Contact:
Medical Staff Director

Risk Manager
Legal Counsel
Public Relations Director

Chair, Board of Directors

This checklist can be used as a guide to properly notify the relevant officers following the occurrence of a sentinel event.
Fill in the appropriate names and phone numbers and keep the information in a location readily accessible to managers
and supervisors. The list should be periodically reviewed and updated.
Source: Adapted from Spath P. Avoid panic by planning for sentinel events. Hosp Peer Rev. 1998 Jun;23(6):112–117. Reproduced by permission.
39
Sidebar 2-8. Practical Issues for Physicians in Disclosing Medical Mistakes to Patients
Definition of a Medical Mistake ►► Express personal regret and apologize for the mistake.
A medical mistake is a commission or an omission ►► Elicit questions or concerns from the patient and
with potentially negative consequences for the patient address them.
that would have been judged incorrect by skilled and ►► Ask if there is anyone else in the family to whom he or
knowledgeable peers at the time it occurred. she should speak.
Deciding Whether to Disclose a Mistake
Consequences of Disclosure
In general, a physician has an obligation to disclose clear
Physicians are most often concerned about the potentially
mistakes that cause significant harm that is remediable,
harmful consequences of disclosing a mistake—partic-
mitigable, or compensable. In cases in which disclosing
ularly the risk of a lawsuit. Serious mistakes may come
a mistake seems controversial, the decision should not
to light even if the physician does not disclose them. Any
be left to the individual physician’s judgment. Obtaining
perception that the physician tried to cover up a mistake
a second opinion is important to represent what a reason-
might make patients angry and more litigious. The risks
able physician would do and be willing to defend in public.
inherent in disclosing a mistake may be minimized if the
This second opinion is best obtained from an institution’s
following things happen:
ethics committee or quality review board rather than from
►► Patients appreciate the physician’s honesty.
informal consultation with peers.
►► Patients appreciate that physicians are fallible.
Timing of Disclosure ►► Disclosure is prompt and open.
Disclosure should be made as soon as possible after the
►► Disclosure is made in a manner that diffuses patient
mistake but at a time when the patient is physically and
anger.
emotionally stable.
►► Sincere apologies are made.
Who Should Disclose the Mistake? ►► Charges for associated care are forgone.
When a mistake is made by a physician in training, ►► A prompt and fair settlement is made out of court.
responsibility is shared with the attending physician. It
may be most appropriate for the attending and house Disclosure of Mistakes Made by Other Physicians
officer to disclose the mistake to the patient together. A physician may encounter situations in which he or
When a mistake is made by a practicing physician, he or she recognizes that a colleague physician has made
she should disclose the mistake to the patient. When the a mistake. That colleague may choose to disclose the
mistake results from the system of medical care delivery, mistake or not. The physician recognizing the mistake has
it may be appropriate to involve an institutional repre- the following options:
sentative in the disclosure, such as an administrator, risk ►► Wait for the other physician to disclose the mistake.
manager, or quality assurance representative. ►► Advise the other physician to disclose the mistake.
What to Say? ►► Simultaneously advise quality assurance or risk
Disclosure is often difficult, for technical as well as management.
emotional reasons. The facts of the case may be too ►► Arrange a joint meeting to discuss the mistake.
complicated to be explained easily and may not be known ►► Tell the patient directly of the error.
precisely. The physician may be tempted to frame the
disclosure in a way that obscures that a mistake was The physician may be reluctant to disclose a colleague’s
made. In telling the patient about an error, the physician error due to the following:
should do the following: Lack of definitive information
►► Treat it as an instance of breaking bad news to the ►► Fear of hurting the colleague’s feelings
patient. ►► Fear of straining a professional relationship
►► Begin by stating simply that he or she regrets that he or
Fear of a libel suit
she has made a mistake or error.
►► The sense that he or she could easily have made
►► Describe the decisions that were made, including those
the same error
in which the patient participated.
►► Social norms against “tattling” on peers
►► Describe the course of events, using nontechnical
language.
Source: Adapted from McPhee SJ, et al. Practical issues in disclosing
►► State the nature of the mistake, consequences, and medical mistakes to patients. Paper presented at the Examining Errors in
corrective action taken or to be undertaken. Health Care Conference, Rancho Mirage, CA Oct 13–15, 1996.
40
handling. Risk management or legal counsel should be cause analysis or a near miss to demonstrate its process
involved in helping to guide communication with the for responding to a sentinel event.
patient and his or her family.
Surveyors also review the effectiveness and sustainability
Survey Process of organization improvements in systems and processes in
In conducting an accreditation survey, Joint Commission response to sentinel events previously evaluated under the
surveyors seek to evaluate the organization’s compliance Joint Commission’s Sentinel Event Policy.
with the applicable requirements discussed in Chapter 1
and to score those requirements based on performance Onward with Root Cause Analysis
throughout the organization over time (for example, the Having developed and implemented a sentinel event policy,
preceding 12 months for a full accreditation survey). Under the organization is now ready to start performing root cause
the sentinel event requirements, accredited organizations analyses and developing an action plan. The next four chap-
are expected to identify and respond appropriately to all ters describe in a step-by-step, workbook format, how to
sentinel events occurring in the organization or associated perform a root cause analysis and develop, implement, and
with services that the organization provides or provides for. assess an improvement-driven action plan.
Surveyors are instructed not to seek out specific sentinel References

events beyond those already known to The Joint Commis 1. Joint Commission International. Understanding and Preventing
sion. The intent is to evaluate compliance with the relevant Sentinel and Adverse Events in Your Health Care Organization.
Oak Brook, IL: Joint Commission Resources, 2008.
Leadership and Performance Improvement requirements—
that is, how the organization responds to sentinel events 2. Croteau RJ. Sentinel events, root cause analysis and the trustee.
Trustee. 2003 Mar;56(3):33–34.
when they occur. However, if a surveyor becomes aware
3. Perper JA. Life-threatening and fatal therapeutic misadventures.
of a sentinel event while on site, he or she will take the
In Bogner MS, editor: Human Error in Medicine. Hillsdale, NJ:
following steps:
Lawrence Erlbaum Associates, 1994, 27–52.
1. Inform the CEO or the CEO’s designee that the event
4. Spath P. Avoid panic by planning for sentinel events. Hosp Peer
has been identified. Rev. 1998 Jun;23(6):112–117.
2. Inform the CEO or the CEO’s designee that the event 5. Keyes C. Responding to an adverse event. Forum. 1997
will be reported to the Joint Commission Sentinel Apr;18(1):2–3.
Event Unit for further review and follow-up under the 6. Blumenthal D. Making medical errors into “medical treasures.”
provisions of the Sentinel Event Policy. JAMA. 1994 Dec 21;272(23):1867–1868.
7. Cantor MD, et al. Disclosing adverse events to patients. Jt Comm
During the on-site survey, the surveyor(s) will assess the J Qual Patient Saf. 2005 Jan;31(1):5–12.
organization’s compliance with sentinel event–related 8. Wu AW, McPhee SJ. Education and training: Needs and
standards in the following ways: approaches for handling mistakes in medical training. Paper
• Review the organization’s process for responding to a presented at the Examining Errors in Health Care Conference,
sentinel event. Rancho Mirage, CA, Oct. 13–15, 1996.
• Interview the organization’s leaders and staff about 9. Witman A, Hardin S. Patients’ responses to physicians’ mistakes.
their expectations and responsibilities for identifying, Forum. 1997 Apr;18(1):4–5.
reporting, and responding to sentinel events. 10. McPhee SJ, et al. Practical issues in disclosing medical mistakes
to patients. Paper presented at the Examining Errors in Health
• Ask for and review an example of a root cause analysis
Care Conference, Rancho Mirage, CA, Oct. 13–15, 1996.
that has been conducted in the past year to assess the
adequacy of the organization’s process for responding
to a sentinel event. Additional examples may be
reviewed if needed to more fully assess the organization’s
understanding of, and ability to conduct, root cause
analyses. In selecting an example, the organization may
choose a previously Joint Commission–reviewed root
41
42
CHAPTER
3
Preparing for Root Cause Analysis
Learning Objectives Root Cause Analysis: Step-by-Step

• Prepare for root cause analysis by recruiting
and organizing a team and identifying roles and STEP 1:
55 Organize a Team
responsibilities for each member STEP 2:
55 Define the Problem
• Identify and define the problem to be STEP 3:
55 Study the Problem
investigated, including identifying the systems STEP 4:
Determine What Happened
involved STEP 5:
Identify Contributing Process Factors
• Learn about the various types of evidence to be STEP 6:
Identify Other Contributing Factors
collected and how they may be used to illuminate STEP 7:
Measure—Collect and Assess Data
the problem on Proximate and Underlying Causes
STEP 8:
Design and Implement Immediate
Changes
STEP 9:
Identify Which Systems Are
Now suppose a sentinel event or a near miss has occurred Involved—The Root Causes
in an organization. Leadership has been informed and, STEP 10:
Prune the List of Root Causes
in the case of an actual sentinel event, the organization STEP 11:
Confirm Root Causes and Consider
has completed preliminary response procedures, including Their Interrelationships
ensuring patient safety, risk containment, and preservation STEP 12:
Explore and Identify Risk Reduction
of evidence (as described in Chapter 2). The appropriate Strategies
staff members have documented the event and ensured STEP 13:
Formulate Improvement Actions
communication with appropriate stakeholders. What STEP 14:
Evaluate Proposed Improvement
happens next? Actions
STEP 15:
Design Improvements
This chapter describes how to prepare for a root cause STEP 16:
Ensure Acceptability of the Action
Plan
analysis. It covers strategies for organizing a team, defining
STEP 17:
Implement the Improvement Plan
the problem, and studying the problem. To help illustrate
STEP 18:
Develop Measures of Effectiveness
the steps of root cause analysis, sentinel events involving
and Ensure Their Success
suicide, elopement, treatment delay, and medication error
STEP 19:
Evaluate Implementation of
are described as examples throughout this and subse- Improvement Efforts
quent chapters. A description of each of these sentinel STEP 20: Take Additional Action

events appears in Sidebar 3-1, page 46. Checklists appear
STEP 21:
Communicate the Results
throughout the chapter. Worksheets are grouped together at
the end of the chapter. Use this chapter and the following
three chapters as a workbook—fill in the blanks. Where the A series of steps comprise the root cause analysis process.
icon appears, a corresponding full description of that These are listed in the box, above. The process begins with
tool appears in Chapter 7. the selection of a multidisciplinary team.
43
STEP 1 Organize a Team from all the involved parties while the incident is still fresh
The first step involved in conducting a root cause analysis is in their minds and then examine what they have collected
to convene a team to assess the sentinel event or potential to identify obvious follow-up questions.
sentinel event. The team should be organized immediately
after the event occurs to ensure that the root cause analysis When the team is assembled, work groups and a detailed
process is not delayed. The team should collect information work plan should be created. The work plan must include
Sidebar 3-1. Sentinel Event Examples
Suicide Several weeks later, the woman calls the physician’s

A 20-year-old male patient is admitted for observation office again, noting that she can feel a hard lump in her
to the behavioral health care unit of a general hospital. breast and mentioning that it hurts. The physician tells
He has a well-documented history of depression. On his her to come into the office right away. Upon review of her
second day in the unit, he attends a particularly clamorous record, the physician finds the results and orders another
group session. Following the session, between 10:00 a.m. mammogram with needle localization right away. The test
and 11:00 a.m., he commits suicide in his bathroom by reveals a change in the size of a nodule from a previous
hanging himself with bedding sheets from the shower mammogram. This result requires an immediate biopsy,
head. A registered nurse finds him at 11:05 a.m., calls a which is positive. Subsequent surgery reveals that the
code, and starts unsuccessful resuscitation efforts. cancer has metastasized.
Elopement Medication Error

Nursing staff in a long term care facility note that an 80- A 60-year-old male patient receiving home care services
year-old woman with a history of progressive dementia is complains about a headache to his home health nurse
unusually irritable and restless. She is pacing, talking in on each of the nurse’s three visits during a one-week
a loud voice, and complaining about a number of issues. period. The man indicates that he is tired of “bothering”
The nurses on duty are unable to appease her or to deter- his primary care physician about various symptoms. At
mine the cause of her distress. Staff members frequently the conclusion of the third visit, the nurse offers to discuss
remind the woman to move away from the exit door. In the man’s complaint with his primary care physician upon
the evening, staff members discover that the woman is no return to the agency.
longer on the unit nor in the building. The woman left the
When the nurse discusses the headache with the man’s
facility without warm clothing on during a cold evening with
physician, the physician instructs the nurse to call the local
subzero temperatures. She is found dead the following
pharmacy with the following prescription: “Fioricet Tabs.
morning in a wooded area near the facility. Cause of death
#30 Sig: 1-2 tabs q 4-6 hours prn headache Refill x3.” In
was exposure.
error, the nurse telephones the pharmacy and provides
the following prescription: “Fiorinal Tabs. #30 Sig: 1-2 tabs
Treatment Delay
q 4-6 hours prn headache Refill x3.” The man has a long
A 60-year-old female patient goes to an ambulatory health
history of peptic ulcer disease, which resulted in several
care organization to receive her annual physical from
hospitalizations for gastrointestinal bleeding.
her longtime physician. The physician orders a mammo-
gram, which she schedules for two weeks later. A week The man takes the Fiorinal, which contains 325 milligrams
after the mammogram, a nurse in the physician’s office of aspirin per tablet. In contrast, the intended medica-
calls the woman and informs her that additional mammo- tion—Fioricet—contains 325 milligrams of acetaminophen
gram views are required but does not convey any sense per tablet. The man completes the entire first prescription
of urgency and then files the x-ray reports in the woman’s and 15 tablets of the first refill. At this point, he goes to the
medical record rather than in the proper file for tests emergency department with acute abdominal pain, blood
requiring follow-up. When the woman calls the physician’s in the stool, and a hemoglobin count of 4.9. He is immedi-
office to ascertain whether the repeat mammogram has ately admitted to the intensive care unit. Within hours, he
been ordered, another office employee tells the woman needs life support. After several units of blood and a four-
not to worry about it if she did not get a call from the week hospital stay, the man recovers and is able to return
physician directly. to his home.
44
CHAPTER 3 | Preparing for Root Cause Analysis
target dates for accomplishing specific objectives. Target care environment, a team should be interdisciplinary.
dates help guide the analysis process and provide a struc- Unlike committees that hold recurring meetings for an
tured game plan for the team moving forward.1 ongoing purpose, a team generally meets for shorter
periods to address a specific issue. After a particular project
Team leaders, however, sometimes experience difficulty is completed, the team often disbands—with a sense of
when trying to keep team members focused. A study accomplishment.
reported in the Australian Health Review found that team
leaders can struggle to maintain focus primarily because Why Use a Team?
teams lack an understanding of the root cause analysis A team approach brings increased creativity, knowledge,
process. Another significant challenge for team leaders in and experience to solving a problem. Just as a patient’s
the study was organizing a team within seven days after the care is provided by a team, the analysis of that care should
incident occurred.2 be carried out by a team that includes representatives
of all the professional disciplines involved in that care.
Leaders must lay the groundwork by creating an environ- Multidisciplinary teams distribute leadership and decision
ment conducive to root cause analysis and improvement making to all levels of an organization. Teams in health care
through team initiatives. Often, leaders need to assure staff organizations provide a powerful way to integrate services
that organization improvement through the identifica- across the continuum of care.3 They also provide a powerful
tion and reduction of risks, rather than the assignment of and often successful way to effect systemwide improvement.
blame, is the objective. Guilt, remorse, fear, and anxiety are
common emotions felt by staff following the occurrence Who Should Work on the Team?
of a sentinel event. These emotions must be addressed and A team may be established on an ad hoc basis, or, if the
discussed at the earliest stages of team formation. Leaders relevant disciplines or services are represented, the core
must put staff members at ease so that staff members of an appropriate team may already exist in the form of a
can contribute to risk reduction. Leaders further lay the targeted performance improvement team or some other
groundwork by empowering the team to make changes type of team. The selection of team members is critical. The
or recommendations for changes, providing the resources team should include staff members at all levels closest to the
(including time to do the work), and ensuring a defined issues involved—those with fundamental knowledge of the
structure and process for moving forward. See Checklist particular process involved. These individuals are likely to
3-1, below. be those with the most to gain from improvement initia-
tives. If the organization has a quality improvement depart-
ment, it is helpful to include a representative from that
Checklist 3-1.
department who was not directly involved with the event to
Essential Elements for a Team Go-Ahead
act as a facilitator.
While developing a team and selecting team members,
ensure that the following three elements are present in The team may include the individual(s) directly involved
the organization’s leadership: in the sentinel event if the organization chooses to include
the person(s). The decision about whether to include the
33 Awareness of and support from top leadership
individual(s) directly involved in the sentinel event can
33 Leadership commitment to provide necessary be made on a case-by-case basis. For instance, if an indi-
resources, including time
vidual is emotionally traumatized by the event, it may
33 An empowered team with authority and make sense to instead invite another person of the same
responsibility to recommend and implement
process changes discipline with comparable process knowledge to join the
team. Later, during the action planning stage when systems
and processes are being redesigned, it may make sense to
What Is a Team? bring the individual(s) directly involved with the sentinel
The word team implies a group that is dynamic and event onto the team to experience the feeling of making a
working together toward a well-defined goal. In the health positive contribution to the change process. The team also
45
should include an individual with some distance from the sentinel event relates to information management, perhaps
process but who possesses excellent analytic skills. In addi- a member of the local chapter of an information manage-
tion, the team should include at least one individual with ment association or organization could be invited.
decision-making authority as well as individuals critical to
the implementation of anticipated changes. Team members Team composition may need to change as the team moves
should bring to the table a diverse mix of knowledge bases in and out of areas within the organization that affect or
and should be well-versed in and committed to perfor- are affected by the issues being analyzed. An organiza-
mance improvement. See Checklist 3-2, below. tion should allow for and expect this change to happen.
However, the core team members should remain as stable
as possible throughout the process, at least in terms of lead-
Checklist 3-2. Team Composition
ership and areas or functions represented. Realistically, the
While drawing up a tentative list of team members, selection of all team members cannot take place until the
check to ensure that the team includes the following: broad aim of the improvement initiatives to be generated
by the root cause analysis and improvement action plan is
33 Individuals closest to the event or issues involved
identified.4
33 Individuals critical to implementation of potential
changes
For example, to investigate the root cause of a medical gas
33 A leader with a broad knowledge base, who is utility disruption that led to the death of a patient when
respected and credible
his oxygen supply was inadvertently cut off, the core team
33 A person with decision-making authority
should include physicians, representatives from clinical
33 Individuals with diverse knowledge bases staff (nursing and/or respiratory therapy), management,
administration, and information management; someone
Physician participation on performance improvement teams who works primarily with medical gases and the utility
varies. Improvement initiatives resulting from root cause management program; and if vendors are a primary part of
analyses often address some aspect of clinical care. In such the analysis, someone from the purchasing or contracting
cases, medical staff involvement is essential. However, at services department. Possible team members could include
times physician involvement is not essential (for example, the following:
in a situation in which a pharmacy stocked an incorrect
drug in a medication drawer and a nurse administered the The core team investigating the suicide in a behavioral
drug without noticing the discrepancy). Leaders should health care unit might include the following individuals:
consider carefully whether a physician needs to be involved • Any staff member, such as a psychiatrist, who was
in the case at hand. Leaders also must understand the attending the patient at the time of the sentinel event
barriers to physician involvement and take steps to over- and can discuss the specific circumstances of the event
come those barriers, which include skepticism about the • A nurse from the behavioral health care unit
purpose of root cause analyses and improvement teams, • An occupational therapist, physical therapist, or
concerns about relevance and effectiveness of teams, and recreation therapist (who has clinical skills and
lack of time and incentive to participate. In cases in which knowledge but would not necessarily spend much time
physician participation is vital, the physician must be on the behavioral health care unit)
convinced to take part. • A social worker on the unit
• A medical staff leader who understands processes and has
Be creative when considering possible team members. the authority to change medical staff policy
Might a former patient, family member, or other commu- • The manager of the behavioral care health unit
nity member be able to provide a unique perspective and • A representative from quality improvement or risk
valuable input? For example, perhaps the town’s retired management (who will act as the facilitator)
pharmacist or a former patient who experienced a near- • An administrative representative at the level of vice
miss sentinel event could be invited to join the root cause president (such as nursing, patient care, or an associate
analysis team. If one of the suspected root causes of a vice president) who can make changes
46
• A safety engineer through the root cause analysis process while encouraging
• A safety consultant (on an ad hoc basis) open communication and broad participation. The leader
may function as a facilitator, or a separate team member
The core team investigating the elopement of a resident can be assigned to play the facilitator role. This individual
from a nursing care center might include the following should be skilled at being objective and moving the team
individuals: along. It is best if the leader and facilitator are not stake-
• The director of nursing holders in the processes and systems being evaluated.
• A unit nurse (regular care provider)
• A nursing assistant (who regularly cared for the resident)
• The medical director
TIP
• The safety director or person responsible for the safety Core teams limited in size to no more than nine
program individuals tend to perform with greater efficiency.
• The individual responsible for performance Experts needed at various points can be added
improvement (facilitator of the group) as ad hoc team members and attend only the
• A social service worker relevant meetings.
• A unit activity staff member
The core team investigating the treatment delay in an Ad hoc members who can provide administrative support,
ambulatory health care organization might include the additional insight, and resources should be identified as
following individuals: well. Use Worksheet 3-1 at the end of this chapter, page 60,
• The director of the ambulatory health care organization to indicate proposed team members.
• A staff physician
• The medical director Ground Rules
• The appointments scheduler At the first team meeting, the leader should establish
• A staff nurse ground rules that will help the team avoid distractions
• The office manager and detours on the route to improvement. The following
• The manager of the laboratory from which the results ground rules provide a framework that will allow the team
were delayed to function smoothly:
• The director or manager of quality or performance • Team mission: The leader should establish the group’s
improvement mission or focus as one of systems improvement rather
than individual fault finding (see Sidebar 3-2, page 48,
The core team investigating the medication error in a home on creating a “no-blame” environment).
care organization might include the following individuals: • Sentinel event policy: The leader should provide
• A home health nurse copies of the organization’s sentinel event policy and
• A nursing supervisor procedures, enabling all team members to become
• An agency director or administrator familiar with expectations.
• A member of the pharmacy supplier’s staff • Decision making: The group must decide what kind
• A local pharmacist (on an ad hoc basis) of consensus or majority is needed for a decision,
• The medical director recognizing that decisions belong to the entire team.
• The quality or performance improvement coordinator • Attendance: Attendance is crucial. Constant late
• An information technology or management staff arrivals and absences can sabotage the team’s efforts. Set
member, as available guidelines for attendance.
• Meeting schedule: For high attendance and steady
The team should have a leader who is knowledgeable, inter- progress, the team should agree on a regular time, day,
ested, and skilled at group consensus building and applying and place for meetings. These matters should be revisited
the tools of root cause analysis. This person guides the team at various times during the team’s life.
47
• Time line: The leader should present a time line at the Sidebar 3-2.
initial meeting to keep the group on track as it moves
Creating a “No-Blame” Environment
toward its quality improvement goals.
• Opportunity to speak: By agreeing at the outset to give To create a “no-blame” environment at meetings,
all members an opportunity to contribute and to be leaders should do the following:
heard with respect, the team will focus its attention on
►► Act as a role model by demonstrating a “let’s learn
the important area of open communication.
from this” type of response when a mistake is
• Disagreements: The team must agree to disagree. It made.
must acknowledge and accept that members will openly
►► To show that everyone makes errors, give an
debate differences in viewpoint. Discussions may
example of a mistake that the leader personally
overflow outside the meeting room, but members should
has made.
feel free to say in a meeting what they say in the hallway.
• Assignments: The team should agree to complete assign ►► Provide examples or positive responses to errors
ments within the particular time limits so that delayed and some of the lessons learned from errors.
work from an individual does not delay the group. ►► Thank staff members for speaking up and offering
• General rules: The team should discuss all other rules suggestions.
that members believe are important. These rules can ►► Reward staff members for suggestions that are
include whether senior management staff can drop in; shown to result in a measurable, positive impact.
how electronic devices such as cell phones, pagers, and
►► Communicate directly and frequently
the like should be handled to minimize distraction; what
the break frequency should be; and so forth.
See Sidebar 3-3, right, for techniques team leaders or Sidebar 3-3.
facilitators can use to ensure high-quality group discussion.
Leadership Techniques for Promoting
How Can the Group Leader Facilitate
High-Quality Group Discussion
Discussions?
The following techniques for team leaders or facilitators
The goal of root cause analysis is to look at a sentinel or can help ensure high-quality group discussion:
adverse event from many angles. However, teams should
►► Use small groups to report ideas.
complete their root cause analysis in a limited amount of
time. To expedite a root cause analysis, the team may have ►► Offer quiet time for thinking.
to create smaller work groups and coordinate those efforts ►► Ask each person to offer an idea before allowing
through detailed work plans and target dates. For this comments or second turns at speaking.
reason, it is likely more important that a team leader have
►► Keep the discussion focused on observations
expertise as a process facilitator than be an expert in any
rather than opinions, evaluations, or judgments.
particular clinical area. The goal of a leader is to track all
the “moving parts” of a root cause analysis and collate the ►► Keep the discussion moving forward within the
team’s findings into a coherent explanation and action plan. allotted time frame.
►► Pull the group together if the discussion fragments
When starting a root cause analysis, the biggest challenge into multiple conversations.
for a group leader often is ensuring that the team is providing ►► Encourage input from quiet members.
an honest, objective assessment of what actually happened.
►► Seek consensus or group decisions.
Team members from the department where the sentinel
event occurred often will be defensive about the incident
and might try, consciously or otherwise, to minimize or also may try to blame others for the incident in an attempt
cover up the role of their department. Some team members to deflect possible criticism away from their coworkers.
48
A group leader must devise ways to get team members as possible is to help focus the team’s analysis and improve-
to talk honestly about the event and provide the details ment efforts. If the team defines and understands the
necessary for an effective root cause analysis. An initial step problem clearly, much time and effort can be saved and
is to emphasize and maintain the confidentiality of the much frustration avoided.
team’s work. Some organizations require team members
to sign a confidentiality agreement stipulating that infor- Often, when using root cause analysis, the problem is
mation shared within the team is not transmitted or defined as a single or sentinel event that caused death or
disclosed outside of the team’s established communication serious injury. However, root cause analysis also can be
mechanisms.5 used to address a series of adverse events that do not in and
of themselves carry the significance of a sentinel event but
Group leaders also can set the stage for open, honest discus- taken in aggregate do pose a serious problem to the health
sions by being aware of group dynamics and keeping team care organization.
members focused on the task at hand. Group dynamic
issues that can detract a team from its goals include fear,
dominant personalities, lack of participation, the use of
TIP
blaming language, and predetermination of the correct A well-defined problem statement describes what
is wrong and focuses on the outcome, not why the
solutions or changes.
outcome occurred.
The group leader should set the ground rules for discussion
and establish that certain behavior will not be tolerated. Whether the root cause analysis is addressing a single
Do not tolerate the use of blaming language. Do not permit incident or a series of events, it is essential to clearly define
one person or group to dominate the discussion. Group the problem.
leaders who ensure that confidentiality is established and
who reinforce the rules of discussion will find that partic- In response to a sentinel event, the team might ask, “What
ipation will increase over time when team members are actually happened or what alerted the staff to a near miss?”
reassured that the goal is to address systems issues and not Initially, the problem or event can be defined simply, such
to blame individuals. as the following:
• Surgery was performed on the incorrect site.
One way the group leader can help team members relax • Patient committed suicide by hanging.
and feel more comfortable about expressing themselves is • Patient could have died following overdose of drug.
to start each meeting with an “ice breaker”—that is, a brief, • Abductor tried to leave the unit with an infant that did
informal activity designed to help participants become not belong to him or her.
more familiar with one another. The ice breaker may be a
brief game or puzzle or simply having each person in turn These simple statements focus on what happened or the
answer a general, open-ended, perhaps even whimsical outcome, not on why it happened. During later steps in
question unrelated to the business at hand, for example, the root cause analysis process, the team will focus on
“What was your favorite food when you were a child?” or the sequence of events, on the whys, and on contributing
“If you could take a free vacation anywhere in the world, factors.
where would you go?” A good ice breaker sets a positive,
inclusive tone for the meeting. For near misses or improvement opportunities, the
preceding problem statements could be restated as follows:
STEP 2 Define the Problem • Surgery was almost performed on the incorrect site.
One of the first steps taken by the root cause analysis team • Patient attempted to commit suicide by hanging.
is to define the problem—that is, to describe, as accurately • Patient almost received overdose of drug.
as possible, what happened or what nearly happened. The • Abductor was discovered by nurse before leaving
purpose of defining the problem as clearly and specifically the unit.
49
Use Worksheet 3-2 at the end of this chapter, page 61, to processes. Complex systems may be dynamic, with constant
define the problem. change and tight time pressure and constraints. The tight
coupling of process steps can increase the risk of failure.
Particularly in the event of a near miss, multiple contrib- Tightly coupled systems or processes do not provide much
uting factors may be present. Which factor should be slack or the opportunity for recovery. Sequences do not
selected first for analysis? Each team needs to develop vary, and delays in one step throw off the entire process.
ranking criteria to help meet this challenge. One option is Variable input and process steps that are nonstandardized
to rank contributing factors by their cost impact, organi- can also increase the risk of process failure. So can processes
zation priority, consequence or severity, safety impact, or carried out in a hierarchical rather than team structure.
real or potential hazard.6 Contributing factors should be
addressed one at a time. The highest-ranked factor should For example, medication ordering is frequently cited as a
be tackled and solved before initiating work on lower- risk-prone system due to organization hierarchies. Nurses
ranked factors. (This topic will be explored in greater depth and pharmacists may be reluctant to question physicians
in Chapter 6.) writing the orders. Some organization cultures may create
a hierarchical rather than team structure for the entire
Help with Problem Definition medication use process. Similarly, verification of surgical
The information disseminated from the Sentinel Event sites by surgical team members can suffer from hierarchical
Database of The Joint Commission can be helpful in an pressures. Nurses may be reluctant to question physicians.
organization’s identification of a problem or area for anal- The hierarchical mind-set can affect everyday scenarios as
ysis. The purpose of this database is to increase general well, as in a situation in which a housekeeper hesitates to
knowledge about sentinel events, their causes, and strategies question a respiratory therapist observed touching patients
for prevention. The Joint Commission collects and analyzes without first performing hand hygiene. Language barriers
data from the review of actual sentinel events, root cause coupled with a hierarchical culture can present a particu-
analyses, action plans, and follow-up activities in all types larly dangerous scenario.
of health care organizations. The hope is that by sharing the
lessons learned with other health care organizations, the risk Identify High-Risk Processes
of future sentinel events will be reduced. Most frequently, sentinel events result from multiple system
failures. They also frequently occur at the point at which
Organizations can learn about sentinel events that occur one system overlaps or hands off to another. An organiza-
with significant frequency, their root causes, and possible tion should be tracking high-risk, high-volume, and prob-
risk reduction strategies through The Joint Commission’s lem-prone processes as part of its performance improve-
publication Sentinel Event Alert (see Sidebar 3-4, on page ment efforts. High-risk, high-volume, problem-prone
51) and the official newsletter The Joint Commission areas vary by organization and are integrally related to the
Perspectives® received by every accredited organization. care, treatment, and services provided. For example, to
Sentinel events reported in the national media, such as reduce the risk of infant abductions, a large maternity unit
the medication error described in the Introduction, can should focus on its infant-parent identification process. To
also serve as a source of ideas for problem analysis. All reduce the risk of patient suicide, a behavioral health care
organizations can use the areas or problems outlined here unit should focus on its process for suicide risk assess-
as a starting point in the identification of a problem area ment. Starting places to find such processes include the
for analysis. list of frequently occurring sentinel events published by
The Joint Commission; an organization’s risk management
Identify Complex Systems data, morbidity and mortality data, and performance data
Identification of failure-prone systems yields problem areas (including sentinel event indicators and aggregate data
requiring focus through root cause analysis. A number indicators); and information about problematic processes
of factors increase the risk of system failures, including generated by field-specific or professional organizations.
complexity—a high number of steps and handoffs in work
50
Sidebar 3-4. Sentinel Event Alert
The following topics have been covered in the Joint Commission’s Sentinel Event Alert publication:
►► Preventing falls and fall-related injuries in health care ►► Preventing, and managing the impact of, anesthesia
facilities (Sep. 2015) awareness (Oct. 2004)
►► Safe use of health information technology (Mar. 2015) ►► Revised guidance to help prevent kernicterus
(Aug. 2004)
►► Managing risk during transition to new ISO tubing
connector standards (Aug. 2014) ►► Preventing infant death and injury during delivery
(Jul. 2004)
►► Preventing infection from the misuse of vials
(Jun. 2014) ►► Preventing surgical fires (Jun. 2003)
►► Preventing unintended retained foreign objects ►► Infection control–related sentinel events (Jan. 2003)
(Oct. 2013) ►► Bed rail–related entrapment deaths (Sep. 2002)
►► Exposure to Creutzfeldt-Jakob Disease (Sep. 2013) ►► Delays in treatment (Jun. 2002)
►► Medical device alarm safety in hospitals (Apr. 2013) ►► Preventing ventilator-related deaths and injuries
►► Safe use of opioids in hospitals (Aug. 2012) (Feb. 2002)
►► Prevention of wrong-site surgery (Dec. 2001,
►► Health care worker fatigue and patient safety
(Dec. 2011) Aug. 1998)
►► Medication errors related to potentially dangerous
►► Radiation risks of diagnostic imaging (Sep. 2011)
abbreviations (Sep. 2001)
►► A follow-up report on preventing suicide: Focus on
►► Preventing needlestick and sharps injuries (Aug. 2001)
medical/surgical units and the emergency department
(Nov. 2010) ►► Medical gas mix-ups (Jul. 2001)
►► Preventing violence in the health care setting ►► Exposure to Creutzfeldt-Jakob disease (Jun. 2001)
(Jun. 2010) ►► Look-alike, sound-alike drug names (May 2001)
►► Preventing maternal death (Jan. 2010) ►► Kernicterus threatens healthy newborns (Apr. 2001)
►► Leadership committed to safety (Aug. 2009) ►► Fires in the home care setting (Mar. 2001)
►► Safely implementing health information and converging ►► Mix-up leads to a medication error (Feb. 2001)
technologies (Dec. 2008)
►► Infusion pumps (Nov. 2000)
►► Preventing errors relating to commonly used anti-
►► Fatal falls (Jul. 2000)
coagulants (Sep. 2008)
►► Making an impact on health care (Apr. 2000)
►► Behaviors that undermine a culture of safety (Jul. 2008)
►► Operative and postoperative complications (Feb. 2000)
►► Preventing pediatric medication errors (Apr. 2008)
►► High-alert medications and patient safety (Nov. 1999)
►► Preventing accidents and injuries in the MRI suite
(Feb. 2008) ►► Blood transfusion errors (Aug. 1999)
►► Preventing adverse events caused by emergency ►► Infant abductions (Apr. 1999)

electrical power system failures (Sep. 2006) ►► Preventing restraint deaths (Nov. 1998)
►► Tubing misconnections—a persistent and potentially ►► Inpatient suicides (Nov. 1998)

deadly occurrence (Apr. 2006)
►► Examples of voluntarily reportable sentinel events
►► Using medication reconciliation to prevent errors (May 1998)
(Jan. 2006)
Current and past issues of Sentinel Event Alert are
►► Preventing vincristine administration errors (Jul. 2005) posted on The Joint Commission’s website at http://www
►► Patient controlled analgesia by proxy (Dec. 2004) .jointcommission.org/sentinel_event.aspx.
51
Developing a Preliminary Work Plan and

Reporting Mechanism Gantt chart, stakeholder analysis
Brainstorming A team’s outline of the reporting mechanism aims to ensure

that the right people receive the right information at the
After a team has defined the problem, it can develop a right time. At the beginning of the process, the team leader
preliminary work plan for investigating the sentinel event or facilitator should establish a means of communicating
through root cause analysis. The plan should outline the team progress and findings to senior leadership. Keeping
overall strategy, key steps, individuals responsible for each senior leaders informed on a regular basis is critical to
step, target dates, and reporting mechanisms. It is essential management support of the root cause analysis initiative
at this point to define the scope of the plan so that team and implementation of its recommendations. Although it is
members can provide a specific answer to the question, difficult to provide guidelines on how a regular basis should
“How will we know if we are successful?” be defined because the time frame varies widely depending
on circumstances, communication with senior leaders
The creation of a detailed work plan is critical to the process should increase in frequency with the following:
and to securing management support. A plan outlining target • Serious adverse outcomes
dates for accomplishing specific objectives provides a tool • Repeated adverse events
against which to guide and measure the team’s progress. • Events requiring solutions from multiple parts of the
organization
The full work plan should include target dates for major • Possible solutions requiring the investment of significant
milestones and key activities in the root cause analysis amounts of money
process. These dates can mirror the steps of the root cause • Media involvement in the case and its solutions
analysis and action plan itself, including the following:
• Defining the event and identifying the proximate and Frequency of communication also varies according to the
underlying causes actions required in the short term to prevent recurrence of
• Collecting and assessing data about proximate and the event. Communication frequency should be weighed
underlying causes against the speed with which information emerges from the
• Designing and implementing immediate changes investigation. If information is emerging rapidly, the team
• Identifying the root causes leader should give thought to the most productive timing
• Planning improvement for communication. A description of reporting consider-
• Testing, implementing, and measuring the success of ations for each of the sample root cause analysis investiga-
improvements tions follows.
For a sentinel event that is reviewable by The Joint The reporting mechanism for a suicide investigation should
Commission or Joint Commission International, the root ensure that the psychiatrist on the team is providing his or
cause analysis must be completed within 45 days of the her colleagues with regular updates on the team’s progress
occurrence of the sentinel event. at clinical department meetings. Such updates prepare the
medical staff for recommended policy changes. Similarly,
Checklist 3-3, page 53, indicates the key steps to include at executive staff meetings, the vice president on the team
in a work plan for a root cause analysis. Each activity is should be providing the CEO, chief operating officer, and
described further in later chapters. Use Worksheet 3-3 other leaders with regular updates on the team’s progress.
at the end of this chapter, page 62, to outline key steps, Communication should be frequent to foster leadership
individuals responsible, and target dates for a root cause acceptance of future recommendations, particularly those
analysis. Also outline the reporting mechanisms and use involving significant resources. Safety should be a standing
the checklist portions to double-check overall strategy and agenda item at board meetings as well.
report quality.
52
Checklist 3-3. Key Steps in Root Cause Analysis and Improvement Planning
The work plan can include the following key activities with target dates for each major milestone:
33Organize a team. 33Formulate actions for improvement (considering

Completion date: actions, measures, responsible party, desired
completion date, and so forth).
33Define the problem. 33Consider the impact of the improvement action.
Completion date:
33Design improvements.
33Choose area(s) for analysis. Completion date:
33Develop a plan. 33Implement action plan.
33Study the problem. Completion date:
Completion date:
33Measure effectiveness.
33Gather information. Completion date:
33Determine what happened and why (proximate
causes). 33Develop measures of effectiveness.
Completion date: 33Ensure success of measurement.
33Evaluate implementation efforts.
33Identify process problem(s). Completion date:
33Determine which patient care processes are involved.
33Determine factors closest to the event. 33Communicate results.
33Extract measurement data. Completion date:
33Identify root causes.
Completion date:
Note: In preparing a root cause analysis in response
33Determine which systems are involved. to a sentinel event remember that the analysis must be
33Design and implement an action plan for improvement. completed no more than 45 days after the event’s occur-
33Identify risk reduction strategies. rence or becoming aware of the event.
Completion date:
The reporting mechanism for the elopement investiga- integration of any new processes or technology that may be
tion should ensure that the safety director is providing recommended to enhance safe medication ordering.
the facility management and operations staff with regular
updates on the team’s progress. Such updates prepare the STEP 3 Study the Problem
staff for any recommended building alterations to enhance The team is now ready to start studying the problem. Doing
the safety of the care environment. so involves collecting information surrounding the event
or near miss. Time is of the essence because key facts can
The reporting mechanism for the treatment delay investi- be forgotten in a matter of days. The individual(s) closest
gation in an ambulatory health care organization should to the event or near miss may have already collected some
ensure that the office manager is keeping the physician information that the team can use as a starting point. Often
or medical director informed of the team’s progress on a a written statement provided by individuals involved in the
regular basis. This regular communication prepares the event and prepared as near the time of the event as possible
medical director for any changes that might be warranted can be useful throughout the root cause analysis process. At
with respect to staffing levels and staff orientation, training, times, the individual(s) closest to the event may withhold
and ongoing competence assessment. critical information due to fear of blame. The team should
consider how to minimize such fear (see Sidebar 3-5, on
The reporting mechanism for the medication error investi- page 56, which provides sample statements for enhancing
gation should ensure that the information technology staff comfort). It may be necessary in some instances to obtain
member is keeping his or her colleagues informed of the an individual’s written statement and then proceed without
team’s progress. This information flow facilitates the smooth his or her contribution in the early stages of the analysis.
53
Sidebar 3-5.
Sample Strategies for Enhancing Comfort
When collecting information about a sentinel event or near miss, health care leaders can use statements such as
those presented here to help the staff member(s) involved feel more comfortable about offering information by doing
the following:
►► Emphasizing that the goal is to learn from the event, ►► Stating that the goal of the inquiry is to look at orga-
not to blame or punish anyone: “Our objective is to nizational processes, not at individual behaviors: “We
learn how to deliver better, safer care at our hospital. really don’t want to know what you did or what your
We’re not here to punish anyone.” coworkers did during this event because your individual
responses to the situation are not the thing that matters
►► Providing examples of mistakes made in the past, how
to us. It’s the overall process that we are looking at.”
the organization learned from the mistakes, and how
patients benefitted from the subsequent improvements: ►► Asking for suggestions on how processes can be
“Last year, when nurses shared details on their medi- changed: “Everyone has an opinion, and we truly value
cation administration practices, we found three areas yours. We realize that in our hectic day-to-day work
in which we could make process improvements. As a lives, you don’t always get a chance to share opinions.
result, our drug error rates dropped significantly.” Well, here is your chance: Tell us what needs to be
changed.”
►► Encouraging staff members to speak freely without
fear of retribution from management or peers: “Please ►► Reinforcing the idea that frontline staff are the only
go ahead and talk freely about the incident. We’re not true process experts—they have the best knowledge
interested in assigning blame, we’re only interested in and vantage point to help come up with effective
helping patients. Isn’t that really why we’re all here in improvements: “You are the real expert. You’re on the
the first place?” front line, so you know what really goes on when you
are delivering care. We’re trying to understand every-
►► Rewarding staff members for providing insight into
thing you go through, so we truly want you to share
the event: “We want to gather insight into what works
your expertise.”
and what doesn’t work with our current processes.
Therefore, we are offering a $25 bookstore gift card
to anyone who can point to a process that needs
improvement.”
Early on, the team should give thought to how information particularly helpful for events such as patient falls, in which
should be recorded. Some methods are more suitable than the team wants to get a sense of the layout of the room
others. For example, audio recording or video recording where a fall occurred. In addition, these various methods of
an interview with someone intimately involved with the recording information aid in reporting the team’s prog-
event is likely to increase that individual’s defensiveness. ress. See Sidebar 3-6, on page 57, for ways of recording
Note taking is an effective way to record interviews. Video information.
recordings, drawings, and/or photographs are effective
media to record physical evidence. For instance, if an orga- In all cases, the team should seek guidance from the orga-
nization experienced an accidental death when an indi- nization’s legal counsel regarding protection of information
vidual served was strangled after slipping through guard from discovery through its inclusion in peer review and
rails on a bed, a video image or photo of the bed with guard other means. The team also should seek guidance from
rails in place provides evidence of the position of the device the organization’s risk management department or legal
following the event. The team should not rely on anyone’s counsel along with representatives of the health informa-
memory. Instead, complete notes, audio recordings, video tion management department concerning patient confi
recordings, photographs, and drawings ensure accuracy and dentiality and the information collected during the root
thoroughness of information collection. Drawings can be cause analysis.
54
The team must ensure focus of data collection. Collecting reassessment for elopement risk; and information regarding
a huge amount of data, much of which might not be how data about individuals at risk for elopement are inte-
related in any way, is both unproductive and confusing. grated into initial and ongoing care plans.
To focus collection efforts, examine the problem statement
and collect data along potential lines of inquiry. The list of Information to be collected in a treatment delay investiga-
questions in Sidebar 3-7, on page 57, can provide guidance tion might include data about how test results are processed
for potential lines of inquiry. For example, if a problem within the organization and how staff members are trained
statement such as “Patient was given wrong medication” in these processes initially and on a continuing basis. Data
suggests an equipment maintenance or human resources related to competence assessment testing and staffing levels
problem, then collect data relevant to medication distribu- would also be valuable.
tion systems, training, and so forth. A sampling of infor-
mation that might be collected relevant to the suicide, Information to be collected in a medication error inves-
elopement, treatment delay, and medication error examples tigation might include data on how medication orders
described in Sidebar 3-1, page 46, follows: are transmitted to local pharmacies and the percentage of
queries and errors due to illegible physician handwriting,
Information to be collected in a suicide investigation might misinterpretation of physician handwriting, telephone or
include an environment of care inventory of all fixtures in verbal orders, and order transcription. Data related to the
the behavioral health care units and data on which fixtures frequency of nurse communication about a new medication
are breakaway compliant and which are not—that is, those and patient education efforts would also be valuable.
that are or are not capable of breaking automatically in
response to a predetermined external force (for example, the Although information or data collection occurs through-
weight of an individual). out the root cause analysis process, the team may also want
to gather three key types of information at this early stage7:
Information to be collected in an elopement investiga- 1. Witness statements and observations from those closest
tion might include an environment of care inventory of to the event or near miss as well as those indirectly
unattended or unmonitored exits; availability and func- involved
tionality of wander-prevention technology such as electric 2. Physical evidence related to the event or near miss
bracelets and wired exits; data related to the thoroughness 3. Documentary evidence
and frequency of initial resident assessment and ongoing
Each type of information is further addressed in the
following sections.
Sidebar 3-6.
Ways to Record Information Witness Statements and Observations
Interviews with staff members can provide a wealth of
The following media can be used to record information information during a number of stages of root cause
obtained during a root cause analysis: analysis. Shortly after an event or near miss, interviews
• Written notes with staff members directly involved can probe for what
• Photographs or drawings happened or almost happened and why (proximate causes).
• Audio recording Interviews with staff members indirectly involved can
• Video recording
explore possible root causes. Later in the process, inter-
It is important that the team obtain legal counsel on views can provide insight into possible improvement
how to protect the organization from being required to initiatives and implementation strategies. The team should
disclose the information in a potential lawsuit as well as try to strike a balance between individual interviews and
the requirements for consent to tape or film individuals.
group interviews. Individual interviews are important for
The facilitator should collect all notes and materials to
keep in the root cause analysis file to maintain existing
obtaining as accurate an account of the situation as possible
protections from legal discovery. from each person, without the influence of others in the
room. However, additional group interviews afterward can
55
encourage greater interplay of ideas and reinforcement of to Sidebar 3-7 for examples of ways to pose such questions
the process improvement message. and for examples of question types that can help ensure that
the questions sound natural.
Conducting interviews is both an art and a science. Some
people do it well; some do not. The team should carefully A two-step probing technique, using an open-ended
consider who is best suited to interview each subject and exploratory question and then a follow-up question asking
the best possible timing and sequence of interviews to be “why,” can yield valuable information. For example, the
conducted. The goal of the interviews is to identify facts, interviewer might first ask, “What can you tell me about
possible systemic causes, and improvement opportunities— the administration of medication in the unit?” After the
not to place blame. The team should identify all likely inter- interviewee responds, the interviewer might follow up
view candidates at each stage and be aware that people tend with, “Why do you think that is the case?” This technique
to forget information or remember it incorrectly, rationalize should be used judiciously and reserved for important
situations, and perceive situations differently. areas because its repetition could make the interviewee
feel “drilled.”
Four discrete stages of what normally appears to be a
continuous interview process are preparing for the inter-
view, opening the interview, conducting the interview, and
closing the interview. Consider the following descriptions8:
TIP
Overcome interviewee defensiveness by doing the
Preparing for the interview. The interviewer plans the following”
interview. Planning involves reviewing previously collected ■■ Restate the focus and purpose of the interview
information, developing carefully worded interview ques- and reiterate that information obtained will be used
tions that are open-ended (see Sidebar 3-7, page 57), sched- to help prevent future occurrences of an adverse
uling the interview, determining how information will be event rather than to assign blame.
recorded and documented, preparing to answer questions ■■ Send positive, supportive messages through
that the interviewee is likely to ask, identifying material statements such as “What you have said is so
that should be available as a reference during the interview, helpful” and “I understand, and you have obviously
and establishing the physical setting. Carefully worded given this a lot of good thought.”
responses to such questions as “Why do you want to talk ■■ Gently ask about a defensive reaction and probe
with me?” and “What will you do with the information I why the interviewee feels threatened. (Take great
care here to ensure that this inquiry will not do
provide?” can go a long way toward reducing the interview-
more harm than good.)
ee’s defensiveness, and so can a neutral setting where privacy
is ensured and interruptions avoided.
Opening the interview. The interviewer should greet Throughout the interview, the interviewer should listen
the interviewee, exchange informal pleasantries, state the well, avoid interrupting the interviewee, avoid talking exces-
purpose of the interview, and answer the interviewee’s ques- sively, ask purposeful questions, and summarize throughout
tions. The goal is to establish rapport, put the interviewee at the interview to confirm a proper understanding of what
ease, establish credibility, and get the interviewee involved the interviewee has related. The interviewer should also be
in the interview process as quickly as possible. The inter- aware of his or her own body language, as well as of the
viewer should indicate at this stage the amount of time the interviewee’s body language and other nonverbal cues.
interview is expected to take.
Closing the interview. The interviewer should check to
Conducting the interview. The interviewer poses the open- ensure that he or she has obtained all necessary informa-
ended questions that were developed during the preparation; ask the interviewee if he or she has any questions or
tion stage. Open-ended questions elicit information by concerns; summarize the complete interview to ensure that
encouraging more than a “yes” or “no” response. Refer back the information accurately reflects the interviewee’s words;
56
process and do not reflect any suspicion of the

Sidebar 3-7.
information provided in initial interviews.
Types of Open-Ended Questions • Interviewees should be encouraged to contact the team
leader with any concerns or additional information.
Open-ended questions ensure that the interviewee
provides more than a simple “yes” or “no” answer.
Group interviews can be more efficient than individual inter-
Three types of open-ended questions that can be used
views, but disadvantages should be considered and weighed
effectively to gain the depth and breadth of information
with care. Disadvantages include dominance by more vocal
needed during interviews follow.
members of the group and the emergence of groupthink,
Exploratory Questions which can stifle individual accounts of an event.
Exploratory questions can be used effectively to begin
the discussion or move to a new topic. They encourage
When in-person interviews are not possible, telephone
the interviewee to provide both comprehensive and
in-depth information. Examples include the following: interviews can provide an alternative. However, the tele-
• “What can you tell me about . . . ?” phone has some serious limitations. It is much more diffi-
• “What can you recall about . . . ?” cult to establish and maintain rapport when eye contact is
not part of the interview, and, of course, nonverbal cues are
Follow-Up Questions
Often, it may be necessary to clarify or amplify informa- much harder to detect.
tion provided by the interviewee. Follow-up questions
can help. Examples include the following: Written responses from an observer to specific questions
• “What do you mean by . . . ?” raised by the team are another alternative (see Worksheet
• “Can you tell me more about . . . ?” 3-4 at the end of this chapter, pages 63–65). However, this
• “What is . . . ?” method is less likely than either in-person or telephone
• “How did this come about?”
interviews to elicit in-depth information. A matter as
Comment Questions seemingly trivial as how much space is provided for answers
Comment questions (or statements) encourage elabo- on the form can have a significant impact on the quantity
ration and express interest while not sounding so much of information provided. In addition, when the observer
like a question. Examples include the following:
must put something in writing, his or her concern about
• “Could you please describe that further?”
the privacy and confidentiality of the information may
• “I would like to know more about that.”
increase defensiveness, thereby preventing full disclosure
Source: Minnesota Department of Public Health. Root Cause Analysis
Toolkit. Apr 7, 2014. Accessed Oct 17, 2015. http://www.health.state and honesty.
.mn.us/patientsafety/toolkit/.
Physical Evidence
and thank the interviewee, expressing appreciation for his Interviews with personnel closest to the event can help the
or her time, honesty, and assistance. team identify relevant physical evidence, including equip-
ment, materials, and devices. Physical evidence related to
After the interviewee leaves the interview area, the inter- the event or near miss should be gathered at an early stage.
viewer should document any further observations and Preserving the evidence immediately following the event or
identify follow-up items. Conclusions and results should near miss can be essential to understanding why an event
be communicated to the root cause analysis team as occurred or almost occurred. In many instances, physical
appropriate. evidence inadvertently (or deliberately) may be taken,
misplaced, destroyed, moved, or altered in some way.
Remember the following points when gathering information
from caregivers6: Physical evidence for a sentinel event involving a medica-
• People’s memories and their willingness to help can be tion error, for example, might include the drug vial, syringe,
affected by the way questions are asked. prescription, IV drip, filter straws, and medication storage
• Interviewees should be informed that follow-up area. Physical evidence for a suicide in a 24-hour care
interviews are a normal part of the root cause analysis setting might include breakaway bars and fixtures in the
57
shower or elsewhere, a window, a ceiling, and other sites. For a suicide, documentary evidence could include the
Physical evidence for a wrong-site surgery might include following:
the x-rays, the operative arena, surgical instruments, and • The patient’s history and physical on admission
so forth. • Staff observation notes
• Attendance logs for unit activities
The evidence should be thoroughly inspected by a knowl- • Policies and procedures for patient observation
edgeable team member, ad hoc member, or consultant. • An inventory of items in the patient’s possession on
Perhaps equipment was not fully assembled or parts admission
were missing. Observations from the inspection should • The patient’s psychosocial assessment
be documented. All physical evidence should be labeled • All physician and nursing notes prior to the incident
with information on the source, location, date and time
collected, basic content, and name of the individual For a medication error involving the administration of the
collecting it and then secured in a separate area, if feasible. wrong medication and the subsequent death of a patient,
If not, such as with a large piece of equipment, the item documentary evidence could include the following:
should be tagged to indicate that it failed and that its use is • The patient’s medical record
prohibited, pending investigation results. • Trending data on medication errors
• Procedures for medication allergy interaction checking
Documentary Evidence • Pharmacy lot number logs
Documentary evidence includes all material in paper or • Pharmacy recall procedure
electronic format that is relevant to the event or near miss. • Maintenance logs for equipment repair
It could include the following: • Downtime logs for computer software
• Patient records, physician orders, medication profiles, • Equipment procedure logs for mixing of solutions
laboratory test results, and all other documents used to • The error report to the US Pharmacopoeia and state
record patient status and care licensing agency (for Joint Commission customers)
• Policies and procedures, correspondence, and meeting or to the local, regional, or national agency (for Joint
minutes Commission International customers)
• Human resources–related documents such as • Lab test results of drug samples
performance evaluations, competence assessments, and • Interdepartmental and interorganizational memos or
physician profiles reports regarding the event
• Indicator data used to measure performance
• Maintenance information such as work orders, For a mechanical error involving the shutoff of oxygen and
equipment logs, instructions for use, vendor manuals, the subsequent death of a patient, documentary evidence
and testing and inspection records could include the following:
• Procedures for informing patient care areas about
All such evidence should be examined, secured, and labeled downtime of mechanical or life-support systems
appropriately. • Construction and technical documents and drawings of
the medical gas distribution system
Documentary evidence varies considerably, based on the • Inspection, performance measurement, and testing
actual sentinel event or possible sentinel event. Examples of policies and procedures
documentary evidence for various error types follow. This • Policies and procedures for shutoff of utility systems
information is a starting place in considering the kind of • Utility system performance measurement data
documentary evidence needed for any organization’s root • Documents related to the utility systems planning process
cause analysis. • Management competence assessment programs
• Technical staff training, retraining, and competence
assessment programs in utilities systems processes
• Incident and emergency reporting procedures
• Maintenance procedures and logs
58
Literature Review improvement strategies recommended by other health care

At this point and throughout the root cause analysis, organizations following a sentinel event or near miss.
a thorough review of the professional literature is an
important component of the process. Literature searches Tools to Use
can yield helpful information about the event at hand The team should begin to consider performance measure-
and other organizations’ experiences with a similar event. ment tools that might be helpful in the next step of the
The literature can help identify possible root causes and root cause analysis—determining what happened and why.
improvement strategies. Appropriate professional or The tool list that follows includes the tools that will be used
industry associations and societies can provide a good in the Chapter 4 search for proximate causes.
starting point in the review process. Obtain a variety of
information on the subject. A review of other organiza- Flowchart, control chart,
tions’ practices and experiences can help avoid mistakes brainstorming, fishbone diagram,
and inspire creative thinking. Information that might be change analysis, Gantt chart
obtained to investigate causes and improvement strategies
for the sample sentinel events shown in Sidebar 3-1, page
44, includes the following possibilities. References
1. Joint Commission International. Understanding and Preventing
For a suicide event, the team might obtain policies, proce- Sentinel and Adverse Events in Your Health Care Organization.
Oak Brook, IL: Joint Commission Resources, 2008.
dures, and forms for suicide risk assessment from other
organizations. A team member might conduct an online 2. Middleton S, Walker C, Chester R. Implementing root cause
analysis in an area health series: Views of the participants. Aust
literature search to obtain protocols for suicide risk assess-
Health Rev. 2005 Nov;29(4):422–428.
ment from relevant professional journals.
3. Ogrinc GS, et al. Fundamentals of Health Care Improvement:
A Guide to Improving Your Patients’ Care, 2nd ed. Oak Brook, IL:
For an elopement event, the team might obtain resident Joint Commission Resources, 2012.
assessment policies, procedures, and forms from other long 4. Shaqdan K, et al. Root-cause analysis and health failure mode
term care organizations and specifically information related and effect analysis: Two leading techniques in health care quality
to how they assess at-risk-for-elopement status. The facility assessment. J Am Coll Radiol. 2014 Jun;11(6):572–579.
or safety manager might obtain information from other 5. Hoffman C, et al. Canadian Root Cause Analysis Framework: A Tool
organizations related to systems used to ensure appropriate for Identifying and Addressing the Root Causes of Critical Incidents
security in the environment of care, such as wander-preven- in Healthcare. Edmonton, AB: Canadian Patient Safety Institute,
tion technology. Assessment protocols from relevant profes- 2006.
sional journals might be helpful as well. 6. Boyd M. A method for prioritizing interventions following root
cause analysis (RCA): Lessons from philosophy. J Eval Clin Pract.
2015 Jun;21(3):461-469.
For a treatment delay event, the team might obtain
7. Andersen B, Fagerhaug T, Beltz M. Root Cause Analysis and
policies, procedures, and protocols for communicating
Improvement in the Healthcare Sector: A Step-by-Step Guide.
abnormal test findings from the professional literature and
Milwaukee: ASQ Quality Press, 2010.
peer ambulatory health care organizations. Training and
8. Minnesota Department of Health. Root Cause Analysis Toolkit.
competence assessment literature could also provide insight (Updated: Apr 7, 2014.) Accessed Oct 17, 2015. http://www
for improvement strategies. Professional organizations .health.state.mn.us/patientsafety/toolkit/.
might be a source of information on criteria for calling in
additional specialists.
For a medication error event, the team might contact other

home care organizations to obtain information about the
policies and protocols used to ensure a safe medication
use process. An online literature review could provide
59
Worksheet 3-1. Composing the Team
Fill in the team leader, facilitator, and team members, including ad hoc members who serve on an as-needed basis.
Ensure interdisciplinary representation by including information such as job titles, degrees, and responsibilities.
Core Team Members
1. Leader
2. Facilitator
3.
4.
5.
6.
7.
8.
9.
10.
Ad Hoc Members
60
Worksheet 3-2. Defining the Problem
Use this space to formulate a simple, one-sentence definition of the event or near miss.
A sentinel event occurred: What happened?

A sentinel event occurred: What happened?
61
Worksheet 3-3. Preliminary Planning
Use this form to guide initial planning.
Overall strategy
 Does it include the team’s aim?  Is it objective?  Is it measurable?

Key steps/initiatives Individual responsible Target date

Reporting mechanisms
Who receives copies of the reports?

Are the reports or other output from the team
 Informative?  Accurate?  Timely?
62
Worksheet 3-4. Gathering Information
Use this worksheet as a place to start gathering written information from individuals who cannot be interviewed in
person or by telephone. Be aware that the amount of space you provide for each question often determines the
amount of information provided by the respondent. If a detailed answer to a certain question is desired, be sure to
leave plenty of space and provide a prompt such as “Please provide as much information as possible.”
What conditions existed prior to the event?
What procedures or processes were being conducted prior to and during the event?
Who was present and involved in the event?
(continued)
63
Worksheet 3-4. Gathering Information (continued)
What indicated that a problem was occurring?
How did you respond?

How did others in the area respond?
What procedures or processes might have been associated with the event?
(continued)
64
Worksheet 3-4. Gathering Information (continued)
What might have caused the event?
How might the event be prevented in the future?
Any other comments or thoughts?
65
66
CHAPTER
4
Determining Proximate Causes

• Learn to identify contributing process factors
STEP 1:
Organize a Team
• Identify what needs to be measured and learn to
STEP 2:
Define the Problem
apply a variety effective measurements
STEP 3:
Study the Problem
• Explore the tools with which a team can search STEP 4: Determine What Happened
55
for proximate or direct causes.
STEP 5:
55 Identify Contributing Process Factors
STEP 6:
55 Identify Other Contributing Factors
STEP 7:
55 Measure—Collect and Assess Data
Prior to this point in the process, the team has created on Proximate and Underlying Causes
a very simple, one-sentence definition of what happened or
STEP 8:
55 Design and Implement Immediate
could have happened. The next step involves creating a more Changes
detailed description or definition of the event. STEP 9:
Involved—The Root Causes
STEP 4 Determine What Happened STEP 10:
This description provides the when, where, and how details STEP 11:
of the event. It should include the following: Their Interrelationships
• Briefly describe of what happened STEP 12:
• Where and when did the event occurred (place, date, Strategies
day of week, and time) STEP 13:
• What areas or services were affected by the event STEP 14:
Actions
(See Worksheet 4-1 at the end of the chapter, page 77, STEP 15:
Design Improvements
for questions to ask in developing a detailed definition of STEP 16:
Plan
an event.)
STEP 17:
For example, with a sentinel event involving a medication STEP 18:
error, a more detailed definition of the event might state, and Ensure Their Success
“A 60-year-old man receiving home care services complains STEP 19:
about a headache to his home health nurse on each of the Improvement Efforts
nurse’s three visits during a one-week period. The man STEP 20: Take Additional Action

indicates that he is tired of bothering his primary care
STEP 21:
physician about various symptoms.
At the conclusion of the third visit, the nurse offers to Fioricet, which results in gastrointestinal bleeding, a trip
discuss the man’s complaint with his primary care physi- to the emergency department, and a four-week hospital
cian upon return to the agency. The discussion results in stay. The prescription was delivered on Friday, January 1,
the physician prescribing a medication. A transcription and the patient was admitted to the hospital on Monday,
error results in the patient receiving Fiorinal, instead of January 4.”
67
The relevant areas for this sentinel event might be nursing, In most cases, identifying the proximate causes is simple; in
medicine, home care, and pharmacy. other cases, it might take some digging. For example, in the
case of the patient found dead by her bed, proximate causes
Developing a flowchart is often helpful to determine the could include “failure to monitor patient,” “bed alarm not
sequence of events. These tools can help the team retain working,” “call light not working,” “patient not properly
focus on seeking the facts of the event. Ensure that the oriented to use of call light,” “incorrect sedation dispensed,”
source of each piece of information is noted on the tool so or “incorrect administration of sedation.”
the source can be consulted if more information is needed.
Underlying causes in the health care environment may relate
In creating this more detailed definition of an event, team to the provision of care or to other processes. Hence, iden-
members should stick to the known facts and not specu- tification of the patient care processes or activities involved
late, prior to completing the root cause analysis, on what in the sentinel event or potential sentinel event will help the
is not yet known or why things happened. For example, team identify contributing causes. At this point, asking and
perhaps the team’s problem statement at this point reads, answering the following three questions will assist the team:
“80-year-old female found in room beside her bed, lying 1. Which processes were involved in the event or almost led
on floor, dead. Event occurred sometime between 0200 and to an event?
0330, Thursday, March 4.” Did the patient die because of 2. What are the steps and linkages between the steps in the
a fall? Or did the patient fall after or while dying? A root process (a) as designed, (b) as routinely performed, and
cause analysis will help the team identify the cause of death. (c) as occurred with the sentinel event?
In this example, relevant areas or services affected by the 3. Which steps and linkages were involved in, or
event might include nursing (to investigate monitoring contributed to, the event?
systems and medication administration), biomedical (to
investigate the type of bed, alarm, and call light system), A variety of tools can help ensure a thorough response to
staffing office (to investigate whether the type of staff— these questions. A flowchart is a useful way to visualize
regular or float—affected the event), education (to investi- the response to the questions “What are the steps in the
gate orientation and training provided to staff and patient), process?” and “What actually happens?” Brainstorming
pharmacy (to investigate patient medications), medical can be used to identify processes and supplement the list of
staff protocols (to investigate medications ordered for the process steps to ensure that all relevant steps are included.
patient), and policies and procedures (related to fall risk Fishbone diagrams and change analysis are useful tech-
assessment, interventions, patient education, and so on). niques in analyzing a response to the question “Which steps
and linkages were involved in or contributed to the event?”
Flowchart
Flowchart, brainstorming, fishbone
diagram, change analysis
STEP 5 Identify Contributing
Process Factors The team can then probe further by asking three more
Root cause analysis involves repeatedly asking “Why?” to questions:
identify the underlying root causes of an event or possible 1. What is currently done to prevent failure at this step or
event. At this point, the team asks the first of a series of its link with the next step?
“why” questions. The goal of the first “why” question is to 2. What is currently done to protect against a bad outcome
identify the proximate causes of the event. Proximate causes, if there is failure at this step or linkage?
or direct causes, are the most apparent or immediate reasons 3. What other areas or services are affected?
for an event. They involve factors closest to the origin of an
event, and they can generally be gleaned by asking, “Why Comparing the flowchart of the process as designed and
did the event happen?” specified in written policies and procedures to the flow-
charts of the process as routinely performed and as occurred
68
CHAPTER 4 | Determining Proximate Causes
with the sentinel event can alert the team to staff actions reporting of hazardous conditions; and electronic commu-
that circumvent policies and procedures either knowingly nication problems.
or unknowingly. Staff members who perform the processes
in question should routinely compare their actual actions Two other common factors in procedure-related failures are
to the prescribed policies and procedures to detect any flawed processes and failure to follow a defined process. It is
discrepancies. (Use Worksheet 4-2 at the end of the chapter, important to distinguish between the two when analyzing
page 78, as a summary of questions to raise and tools to root causes and developing improvement actions.
consider using.)
A flawed process is a process that can become a root cause of
STEP 6 Identify Other Contributing Factors errors, even when people are following the process correctly.
In health care environments, proximate causes tend to fall When a flawed process is identified as the cause of an error,
into a number of distinct categories beyond and in addition then the process must be analyzed and changed to eliminate
to process factors. They include the following: the possibility of more errors.
• Human factors
• Equipment factors Failure to follow a defined process occurs when staff members
• Information-related factors fail to follow an established process and then that mistake
• Controllable or uncontrollable environmental factors in turn causes an error. When failure to follow a defined
process is identified as a factor, there must be an inquiry as
To identify additional proximate causes of an event to why the process wasn’t followed: Is it a bad policy or is
involving human factors, the team might ask: it that staff members were not informed about the policy?
• What human factors were relevant to the outcome? Are shortcuts tolerated in the organization?
• How did the equipment performance affect the
outcome? Continuing the example of a patient suicide as described
• What information-related factors were relevant to in Sidebar 3-1, in identifying the proximate causes,
the outcome? a team might conclude that proximate causes include
• What factors directly affected the outcome? the following:
Were such factors within or truly beyond the • Human factors such as failure to follow policies on
organization’s control?” precaution orders or failure to conduct appropriate staff
education or training
Finally, the team might ask, “Are there any other factors • Assessment process factors such as a faulty initial assess
that have directly influenced this outcome?” Figures 4-1 ment process that did not include identification of a
through 4-3, pages 70–71, identify common factors asso- history of suicide attempts or an immediate psychiatric
ciated with failures related to procedure, and equipment. consultation
(Use Worksheet 4-3 at the end of the chapter, pages 79–80, • Process or human factors such as a faulty history and
to identify factors closest to the event.) physical assessment that did not identify patient suicide
risk factors
Communication-related errors often set in motion the • Equipment factors such as a nonfunctional paging system
cascade of events that result in procedure-related failures. that delayed communication with the patient’s physician
By far, the majority of these errors involve two types of Brainstorming to identify all possible or potential
communication: (1) communication of relevant patient contributing causes may be a useful technique for
information among staff members and (2) communication teams at this stage of the root cause analysis. Following
between physicians and other physicians or staff members. traditional brainstorming ground rules—such as not
Other communication-related errors include communica- labeling anything a bad idea and ensuring that team
tion between staff and the patient or family; oral commu- members do not express reactions or provide commen
nication problems, such as incomplete change-of-shift tary as ideas are expressed—is critical to success. The
reports; problems with administration, such as delayed focus must be on improving patient outcomes rather than
69
Figure 4-1. Common Factors Associated with Procedure-Related Failures
Source: Mobley RK. Root Cause Failure Analysis. Boston: Newnes, 1999, p. 40. Reproduced by permission.
Figure 4-2. Common Factors Associated with Training-Related Failures
70
Figure 4-3. Common Factors Associated with Equipment-Related Failures
individual performance. Affinity diagrams can be used to STEP 7 M

easure—Collect and Assess
help sort or organize the causes or potential causes into Data on Proximate and
natural, related groupings. Fishbone diagrams can help Underlying Causes
highlight the numerous factors involved in an event. To advance further toward discovering root causes, the
team must explore in depth proximate and underlying
Brainstorming, affinity diagram, causes. This exploration involves measurement—collecting
fishbone diagram and assessing relevant data. Although this exploration is
presented here as Step 7, data collection and analysis initia-
While asking questions to uncover causes, the team leader tives may occur throughout root cause analysis and need
should keep team members focused on processes, not not be sequential or follow Step 6 and precede Step 8.
people. One individual’s actions generally will not be a root
cause. The team must focus on the systems within which Measurement is the process of collecting and aggregating
people are operating. That is the level at which most root data. The process helps assess the level of performance,
causes are found. By repeatedly asking “Why?” the team can determine whether improvement actions are necessary, and
continue working until it feels it has exhausted all possible ascertain whether improvement has occurred.
questions and causes. The importance of this stage cannot
be overstated. It provides the initial substance for the root Data gathering must begin as soon as possible after the
cause analysis without which a team cannot proceed. event occurs to prevent loss or alteration of the data. Data
from people are the most easily altered or destroyed and
After sorting and analyzing the cause list, the team may need to be made a priority. Other forms of data are more
start to determine which process or system each cause is a stable; however, physical data need to be identified quickly
part of and whether the cause is a special or common cause to prevent their inadvertent loss or destruction.
in that process or system. This process, described fully in
the next chapter, helps unearth systems-based root causes.
71
Purposes of Measurement Sidebar 4-1.

Organizations collect data to monitor the stability of existing
Choosing What to Measure
processes, identify opportunities for improvement, identify
changes that lead to improvement, and maintain changes. Choosing what to measure is extremely important.
An organization may wish to start by defining the broad
The first purpose of measurement is to provide a base- processes or systems most likely to underlie proxi-
line when little objective evidence exists about a process. mate causes. For example, if a team is investigating a
For example, a health care organization may want to medication error involving the process used to commu-
learn more about the current level of staff competence. nicate an order to the pharmacy and the process used
A dementia long term care or psychiatric special care unit by pharmacy staff to check the dosage ordered, the
team may decide to measure the following for a defined
may want to know more about the effectiveness of the
period of time:
bed alarm systems to prevent patient falls and elopement.
Specific indicators for a particular outcome or a particular ►► Time elapsed between when an order is written by
step in a process may be used for ongoing data collec- medical staff and when the pharmacy receives the
order by fax, pickup, or phone
tion. When assessed, these data can help management
and staff determine whether a process is ineffective and ►► Number of calls to prescribers for clarification of
needs more intensive analysis. Data about costs, including an order
costs of faulty or ineffective processes, may also be of ►► Time elapsed between when the dosage is
significant interest to leaders and can be part of ongoing checked by pharmacy staff and when medication
performance measurement. is dispensed
►► Time elapsed between when medication is
The second purpose of measurement is to gain more infor- dispensed and when medication is administered
mation about a process chosen for assessment and improve-
ment. For example, a performance rate varies significantly
from the previous year, from shift to shift, or from the processes (for example, the preparation and administration
statistical average. Records may indicate that the staff of medications) are in control. When a process has been
members on duty during weekend hours do not complete stabilized at an acceptable level of performance, it may be
suicide risk assessments at intake. Or perhaps the data indi- measured periodically to verify that the improvement has
cate that patient observations are incomplete or infrequent been sustained. Measurement to monitor improvement
when specific personnel are present. Perhaps patient assess- actions is described fully in Chapter 6.
ment methods or other care planning factors are suspected
as root causes of a sentinel event. Such findings may cause Choosing What to Measure
a health care organization to focus on a given process to Choosing what to measure is absolutely critical at all stages
determine opportunities for improvement. of root cause analysis—in probing for root causes and
in assessing whether a recommended change or action
The target for further study should be time limited and can represents an actual improvement (see Sidebar 4-1, above).
test a specific population, a specific diagnosis, a specific Measurement requires indicators that are stable, consis-
service provided, or an organization management issue. tent, understandable, easy to use, and reliable. Indicators,
Detailed measurement is then necessary to gather data or performance measures, are devices or tools for quantifying
about exactly how the process performs and about factors the level of performance that actually occurs. They are valid
affecting that performance. if they identify events that merit review, and they are reli-
able if they accurately and completely identify occurrences
A third purpose of measurement is to determine the effec- (see Sidebar 4-2, on page 73).
tiveness of improvement actions. For example, a nursing
unit that begins to use a new piece of equipment needs Process and Outcome Measures
to establish a baseline performance rate and continue to A process measure is an intermediate indicator of the
measure use. Measurement can also demonstrate that key success of an intervention. An outcome measure is a specific,
72
Sidebar 4-2. and aggregate data indicators. A special type of single-event,

outcome indicator is a sentinel event. A sentinel event indi-
Definition of an Indicator or Measure
cator identifies an individual event or phenomenon that is
An indicator (or measure) can be described in the
significant enough to trigger further investigation each time
following ways: it occurs. Such indicators are well known to risk managers,
who help ensure that each event is promptly evaluated to
►► Quantitative—It is expressed in units of
prevent future occurrences.
measurement.
►► Valid—It identifies events that merit review. Although sentinel event indicators are useful to ensure
►► Reliable—It accurately and completely identifies patient safety, they are less useful than aggregate data
occurrences. indicators in measuring the overall level of performance in
►► A measure of a process or outcome—It involves
an organization. An aggregate data indicator quantifies a
a goal-directed series of activities or the results of process or outcome related to many cases. Unlike a sentinel
performance. event, the events typically measured by an aggregate data
indicator may occur frequently. Aggregate data indica-
tors are divided into two groups: rate-based indicators and
measurable indicator of the end result of an action. For continuous variable indicators.
example, monitoring the number of speeding tickets a
person receives is a process measure of that person’s safety as Rate-Based Indicators
a driver. The number of car accidents that person has is an Rate-based indicators express the proportion of the number
outcome measure of how safe a driver that person actually is. of occurrences to the entire group within which the occur-
rence could take place, as in the following examples:
Outcome measures are often misleading in health care.
Patients receiving cesarean sections
For example, a root cause analysis could be used to change a
All patients who deliver
process that theoretically led to a sentinel event. The organi-
zation can measure outcomes, and the sentinel event might
Total number of elopements
not occur again for many years. However, because sentinel
Patients at risk for elopement (wandering and confused)
events occur only rarely, it is difficult to make a direct
correlation between the outcome measure and the change.
Patients with central line catheter infections
This is a limitation of the use of outcome measures to assess
All patients with central line access devices
the effectiveness of process changes intended to reduce
the incidence of rare events. For example, if a wrong-site
Patient falls associated with adverse drug reactions
surgery typically occurs in each hospital only once every four
All patient falls
to five years, can a hospital that has gone without
a wrong-site surgery for several years claim that the root The rate can also express a ratio comparing the occurrences
cause analysis has led to improved outcomes? identified with a different but related phenomenon, as in
the following example:
Examples of an outcome indicator are “catheter-associated
Patients with central line infections
sepsis for patients with a central venous access device” or
Central line days
“percentage of patients at risk for falls who actually expe-
rience falls while in the health care organization.” An Continuous Variable Indicators
example of a process indicator is “patients older than 65 This type of aggregate data indicator measures performance
having medication monitoring for drugs that can decrease along a continuous scale. For example, a continuous vari-
renal function.” able indicator might show the precise weight in kilograms
of a patient receiving total parenteral nutrition (TPN). Or
In addition to categorizing indicators as measures of process it might record the number of written pharmacist recom-
or outcome, they can be classified as single-event indicators mendations accepted by the attending physicians. While
73
Checklist 4-1. Sidebar 4-3.

Criteria to Ensure Appropriate Data Key Questions About Measurement
Collection
Throughout the root cause analysis, the team should
Choosing what to measure is critical, and so is ask the following questions concerning measurement:
ensuring that the data collected are appropriate to the ►► What will be measured?
desired measurement. This checklist includes criteria to (This defines what is critical to determining root
help the team ensure that the data collected are appro- causes.)
priate for monitoring performance.
►► Why will it be measured?
33 The measure can identify the events, patterns, or (This verifies the criticality of what will be
trends of performance it was intended to identify. measured.)
33 The measure has a documented numerator and ►► What can be gained from such measurement?
has a denominator statement or description of the
(This describes the incentives of measurement.)
population to which the measure is applicable.
►► Who will perform the measurement?
33 The measure has defined data elements and
allowable values. ►► How frequent is the measuring?
33 The measure can detect changes in performance ►► How will the data be used when measurement is
over time.
completed?
33 The measure allows for comparison over time ►► Is the measure or measurement reliable?
within the organization or between the organization
and other entities. ►► Is measurement a onetime event or a periodic or
continuous process?
33 The data intended for collection are available.
►► What resources are needed for measurement?
33 The sources of the data and methods of collecting
the data are defined What is available?
33 Results can be reported in a way that is useful to ►► Do the measures consider dimensions
the organization and other interested parties. of performance?
a rate-based indicator might relate the number of patients STEP 8 D

esign and Implement
approaching goal weights to the total number of patients on Immediate Changes
TPN, a continuous variable indicator measures the patient’s Even at this early stage, when the team has identified only
average weight change (that is, the patient’s weight one proximate causes, some quick or immediate fixes may be
month minus the patient’s weight the previous month). appropriate. For example, in the case of an organization
that experienced the suicide of a patient who was not iden-
See Checklist 4-1, above, for criteria that will help the team tified as being at risk for suicide, the organization could
ensure that the measure or indicator selected is actually immediately evaluate its current risk assessment tool to
appropriate for monitoring performance. See Sidebar 4-3, learn whether it meets current standards of practice. Or the
right, for key questions the team should ask about measure- organization could address environment of care issues such
ment throughout the root cause analysis process. as nonbreakaway shower heads and bed linens.
Additional information on measurement, including Teams conducting root cause analysis need not wait until
how to measure the effectiveness of improvement initia- they finish their analysis to start designing and imple-
tives and ensure the success of measurement, appears in menting changes. During the process of asking “Why?”
Chapter 6. potential interventions emerge. Immediate changes may
not only be appropriate but necessary. First, they may
be needed to reduce an immediate risk. For example, an
74
unsecured window may need to be repaired and secured the medical staff to discuss revisions to requirements for
right away; an intoxicated employee should be immediately histories and physicals.
removed from the environment; and broken or malfunc-
tioning equipment should be promptly taken out of the Or the organization could start conducting mandatory
area of care, treatment, or service, and secured. in-service training for all staff on suicide risk assessment.
Or it could place all patients with psychiatric or substance
Second, immediate changes may also uncover additional abuse diagnoses on suicide precautions. A Gantt chart
causes that were previously masked but are critical to the used by one organization to outline the key steps and time
search for the root cause(s). Finally, immediate changes frames for a plan to eliminate proximate causes that led to a
can be part of a performance improvement cycle such patient suicide appears as Figure 4-4, page 78.
as Plan-Do-Study-Act (PDSA) to test process redesign
before implementing it organizationwide. For instance, an Gantt chart
organization may wish to test the use of new bathroom
hardware in one room before changing hardware
throughout the facility. In the case of an organization that experienced a wrong-
site surgery, the organization could require a second staff
In addition, the organization could evaluate the assess- member to observe operating room team procedures to
ment tool for suicide risk and the process used to check assess compliance with the Universal Protocol for Preventing
for contraband, and it could then initiate meetings with Wrong Site, Wrong Procedure, Wrong Person Surgery™.
75
Figure 4-4. Gantt Chart
3rd Qtr 2009 4th Qtr 2009 1st Qtr 2010 2nd Qtr 2010
Person(s) Responsible
TASK Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun
Establish review team
Expand security round

to include helipad
Revise suicide precaution

policy
Staff training: Suicide

precaution policy
Staff training: Assess-

ment and interventions
for patients with mental
health needs
Staff training: Patients

leaving until unattended
Team meetings to assess

Plan-Do-Study-Act
(PDSA)
Medical record review of

all suicide precaution
charts
Concurrent assessment
of patients or suicide
precautions
This Gantt chart details the target date and person(s) responsible for each task in the process of identifying and
eliminating proximate causes of a patient suicide.
76
Worksheet 4-1. Further Defining What Happened
In developing a more detailed definition of what happened, the team should consider the following three questions:
1. What are the details of the event? (Write a brief, two- or three-sentence description.)

2. When and where did the event occur (place, date, day of week, and time)?

3. What area(s) or service(s) was affected?
77
Worksheet 4-2. Further Defining What Happened
To determine proximate causes of a sentinel event or possible sentinel event involving patient care or organization
processes, the team can ask the following questions and consider using the following tools to aid in answering each
question.
1. Which processes were involved or could have been involved in the event or near miss?
Suggested Tools:  Brainstorming  Fishbone diagram

2. What are the steps in the process, as designed?
Suggested Tools:  Flowchart  Brainstorming

3. What are the steps in the process, as it occurred?
Suggested Tools:  Flowchart

4. Which steps were (or could have been) involved in, or contributed to, the event or near miss?
Suggested Tools:  Fishbone diagram  Change analysis  Failure mode and effects analysis (FMEA)

5. What is currently done to prevent failure at this step?
Suggested Tools:  Flowchart

6. What other areas or services are affected?
Suggested Tools:  FMEA  Brainstorming
78
Worksheet 4-3. Identifying Factors Close to the Event
Use this worksheet to identify factors closest to the event or possible event.
Human factors included or could include
Equipment factors included or could include
Controllable environmental factors included or could include
(continued)
79
Worksheet 4-3. Identifying Factors Close to the Event (continued)
Uncontrollable environmental factors included or could include

Other factors included or could include
80
CHAPTER
5
Identifying Root Causes

• Identify which systems underlie the proximate
cause(s) of the event STEP 1:
Organize a Team
STEP 2:
Define the Problem
• Find the root causes within those systems that
contributed to the event STEP 3:
Study the Problem
STEP 4:
Determine What Happened
• Consider how the root causes in different
STEP 5:
systems relate to and affect each other
STEP 6:
STEP 7:
on Proximate and Underlying Causes
The probing continues. At this point, the team has a detailed
STEP 8:
description of the event or near miss, a description of the Changes
patient care processes involved, and a list of proximate STEP 9:
55 Identify Which Systems Are
causes and other factors that might have caused or contrib- Involved—The Root Causes
uted to the problem or could do so in the future. The team STEP 10:
55 Prune the List of Root Causes
has also started to collect data on proximate causes. STEP 11:
55 Confirm Root Causes and Consider
Their Interrelationships
STEP 9 Identify Which Systems Are STEP 12:
Involved—The Root Causes Strategies
Now the team again asks, “Why did that proximate cause STEP 13:
happen? Which systems and processes underlie proximate STEP 14:
factors?” The goal of asking questions at this stage is to Actions
identify the underlying causes for the proximate causes. For STEP 15:
Design Improvements
example, in the case of the elderly patient found dead on STEP 16:
Plan
the floor by her bed, questions might include the following:
STEP 17:
• Why was the patient not monitored for an hour to an
STEP 18:
hour and a half?
• Why was a new graduate nurse assigned to this patient’s
STEP 19:
care? Did the nurse have the assistance of ancillary staff? Improvement Efforts
• How much orientation had the nurse completed? STEP 20: Take Additional Action

• Why was the patient given a sedative?
STEP 21:
• Why was the call light not by the patient’s hand?
• Was the patient assessed to be at risk for falling?
• Were interventions established? If so, why were they As in all stages of the process, it is critical to keep the team
not done? focused on probing for system or common-cause prob-
lems rather than focusing on human errors. Teams often
81
have trouble at this stage of the root cause analysis. The

tendency is to stop short after identifying proximate causes
TIP
and not to probe deeper. The probing must continue until CLARIFY ALL ISSUES. The team must clearly
a reason underlying a cause can no longer be identified. The define the issues regarding the patient safety event
and ensure that team members share a common
resulting cause, then, is a root cause.
understanding of the issues. No matter how obvious
the issues may seem, individual team members may
Underlying causes may involve special-cause variation, not understand them in the same way. For example,
common-cause variation, or both. Being “special” or a staff member may not consider identification of the
“common” is not an inherent characteristic of the cause surgical site by all operating room team members
itself. Rather, it describes the relationship of the cause to a to be worth the time involved. Or, if a patient suffers
specific process or system. It is possible for the same cause a burn during surgery, surgical team members may
to be a special cause in one process and a common cause in not agree about whether the burn could have been
affected by the proximity of the oxygen cannula to the
another. A flowchart of the process(es) at this stage may be
cautery or how the cannula was handled during the
very helpful. cauterizing procedure.
Flowchart
They include processes for the following:
• Human resources
For a special cause in a process, teams should search for • Information management
the common cause in the system of which the process is • Environmental management
a part. Keep asking, “Why did the special-cause variation • Leadership—embracing organization culture,
occur?” to identify one or more common-cause variations encouragement of communication, and clear
in the supporting systems that may represent root causes. communication of priorities
There should be a review of the processes and subprocesses
that compose the system. This examination should include In addition, factors beyond an organization’s control should
a review of existing policies and procedures by the process be considered a separate category. Organizations must exer-
owners in comparison with the actual practice. Any varia- cise caution in assigning factors to this category, however.
tions should be evaluated for the extent of common-cause Although a causative factor may be beyond an organiza-
variation and the presence of special-cause variation. tion’s control, the protection of patients from the effects
of the uncontrollable factor is within the organization’s
For example, a special cause is created when one group of control in most cases and should be addressed as a risk
surgeons and their assistants do not follow hospital proce- reduction strategy.
dures for hand hygiene, resulting in a sentinel event. This
special cause might be part of a common-cause variation Concrete questions about each function listed above can
in a larger system: the hospital’s inconsistent education in help team members reach the essence of the problem—the
sterile techniques and hand washing. Use Worksheet 5-1 systems that lie behind or beneath problematic processes. At
at the end of this chapter, page 89, to organize the team’s this stage, questions can be worded in the following form: “To
probe for underlying causes. what degree does . . . ?” Follow-up questions for each could
be “Can this be improved, and if so, how?” See Sidebar 5-1,
A logical starting point in the team’s effort to determine page 83, for a fuller itemization of possible questions.
the systems involved with the event or near miss is listing
and categorizing the possible causal factors. Common or Other questions may emerge in the course of an anal-
root causes of a sentinel event in a health care organization ysis. The team should fully consider all questions. One
can be categorized according to the important organiza- team investigating a patient suicide found that systems
tion functions or processes performed by the organization. involving human resources, information management,
82
CHAPTER 5 | Identifying Root Causes
and environmental management issues were root causes of Patient safety events can be very complex and involve
the sentinel event: multiple causes. Understanding causes is essential if the
• In the human resources area, age-specific staff organization is to create lasting improvements. Certain
competence had not been assessed adequately, and staff tools can be particularly helpful in systematically looking
needed additional training in management of suicidal at an event to determine its causes. The tools are designed
patients. to help root cause analysis team members understand
• In the information management area, information processes and factors that contribute to both good and
about the patient’s past admission was not available. problematic performance. Groups can also use the tools to
Communication delays resulted in failure to implement study a process, without requiring a statistical background.
appropriate preventive actions. They may be used singly or in combination to show the
• In the environmental management area, the team relationship between processes and factors, reach conclu-
found that access to the appropriate unit for the patient sions, and systematically analyze causes.
was denied.
Sidebar 5-1. Root Cause Analysis Questions
The following questions may be used to probe for systems Questions concerning environmental management issues
problems that underlie problematic processes. may include the following:
►► How appropriate is the physical environment for the
Questions concerning human resources issues may
processes being carried out? Can it be improved,
include the following:
and if so, how?
►► To what degree were staff members involved in this
event properly qualified and currently competent for ►► To what degree are systems in place to identify envi-
their responsibilities? Can these qualifications be ronmental risks? Can these systems be improved,
improved, and if so, how? and if so, how?
►► How did actual staffing at the time of the event ►► What emergency and failure-mode responses have
compare with ideal levels? Can the staffing level be been planned and tested? Can these responses be
improved, and if so, how? improved, and if so, how?
►► What are the plans for dealing with contingencies Questions concerning leadership issues may include the
that tend to reduce effective staffing levels? Can following:
these plans be improved, and if so, how? ►► How conducive is the culture to risk identification and
reduction? Can this culture be improved, and if so,
►► To what degree is staff performance in the operant
how?
processes addressed? Can such staff performance
be improved, and if so, how? ►► What are the barriers to communication of potential
risk factors? Can these barriers be reduced or elimi-
►► How can orientation and in-service training be
nated, and if so, how?
improved?
►► To what degree is the prevention of adverse out-
Questions concerning information management issues
comes communicated as a high priority? How is such
may include the following:
prevention communicated? Can this communication
►► To what degree was all necessary information
be improved, and if so, how?
available when needed in the case of this event?
What are the barriers to information availability and Questions concerning uncontrollable factors may include
access? To what degree is the information accurate, the following:
complete, and unambiguous? Can these factors be ►► What can be done to protect against the effects of
improved, and if so, how? uncontrollable factors?
►► How adequate is the communication of information
among participants? Can such communication be
improved, and if so, how?
83
on how a task is to be performed prior to accomplishing

Flowchart, fishbone diagram the task
• Change management: The process whereby the
hardware or software associated with a particular
Fishbone diagrams are particularly helpful in categorizing operation, technique, or system is modified
and visualizing multiple system or process problems that • Resource management: The process whereby personnel
have contributed to a sentinel event or near miss. The stan- and material are allocated for a particular task or
dard categories coming off the main “spine” of the diagram objective
include people, procedures, equipment or materials, envi- • Managerial methods: An administrative technique used
ronment, and policies. Such categories as communication, to control or direct work-related plan activities, which
education, leadership, and culture may also be appropriate. includes the process whereby staff and material are
Subcauses branch off each major category. Checklist 5-1 on allocated for a particular task objective
page 85 provides a handy way to help ensure that the team
has considered selected systems-based issues. Equipment Factors
• Design configuration and analysis: The design layout
Another proposed classification system for causal factors is of a system or subsystem needed to support plan
geared more toward a manufacturing environment. However, operation and maintenance
it may be helpful to review this and other classification • Equipment condition: The failure mechanism of the
systems to ensure that the team has identified all possible equipment that is the physical cause of the failure
causal factors. This causal factor category list follows.1 • Environmental conditions: The physical conditions of
the equipment area
Human Factors • Equipment specification, manufacture, and
• Verbal communication: The spoken presentation or construction: The process that includes the manufacture
exchange of information and installation of equipment in a plant
• Written procedures and documents: The written • Maintenance and testing: The process of maintaining
presentation or exchange of information components and systems in optimal condition
• Human-machine interface: The design of equipment • Plant and system operation: The actual performance
used to communicate information from the plan to of the equipment or component when performing its
a person intended function
• Environmental conditions: The physical conditions of
a work area External Factor
• Work schedule: Factors that contribute to the ability • External: Human or nonhuman influence outside the
of a worker to perform his or her assigned task in an usual control of the company
effective manner
• Work practices: Methods workers use to ensure safe and STEP 10 Prune the List of Root Causes
timely completion of tasks The team’s list of causal factors may be lengthy. Regardless
• Work organization and planning: The work-related of the list’s length or the technique used, the team should
tasks including planning, identifying the scope of, analyze each cause or factor using reasoning skills based on
assigning responsible individuals to, and scheduling the logic. Asking two questions helps clarify whether each cause
task to be performed or problem listed is actually a true root cause:
• Supervisory methods: Techniques used to directly 1. If we fix this problem, will the problem recur in the
control work-related tasks; in particular, a method used future?
to direct workers in the accomplishment of tasks 2. If this problem is a root cause, how does it explain what
• Training and qualification: How a training program happened or what could have happened?
is developed and the process of presenting information
84
Checklist 5-1. Problematic Systems or Processes
Use this checklist to identify and rank problematic systems or processes. Use a 1 to indicate a problem that is a primary
factor and a 2 to indicate a problem that is a contributing factor.
Human Resources Issues Information Management Issues
Qualifications of staff Availability of information
Defined Accuracy of information
Verified Thoroughness of information
Reviewed and updated on a regular basis Clarity of information
Qualifications of physicians Communication of information between
Defined relevant individuals/participants
Verified Environmental Management Issues
Reviewed and updated on a regular basis Physical environment
Qualifications of agency staff Appropriateness to processes
Defined being carried out
Verified Lighting
Reviewed and updated on a regular basis Temperature control
Training of staff Noise control
Adequacy of training program content Size/design of space
Receipt of necessary training Exposure to infection risks
Competence/proficiency testing following training Cleanliness
Training of physicians Systems to identify environmental risks
Adequacy of training program content Quality control activities
Receipt of necessary training Adequacy of procedures and techniques
Competence/proficiency testing following training Inspections
Training of agency staff Planned, tested, and implemented emergency
Adequacy of training program content and failure-mode responses
Receipt of necessary training
Competence/proficiency testing following training Leadership and Communication Issues
Competence of staff Data use
Initially verified Use in decision making
Reviewed and verified on a regular basis Use to identify changes in the internal and
Competence of physicians external environments
Initially verified Planning
Reviewed and verified on a regular basis For achievement of short-term and long-term
Competence of agency staff goals
Initially verified To meet challenge of external changes
Reviewed and verified on a regular basis Design of services and work processes
Supervision of staff Creation of communication channels
Adequate for new employees Performance improvement
Adequate for high-risk activities Introduction of innovation
Current staffing levels Communication
Based on a reasonable patient acuity Present, as appropriate
measure Appropriate method
Based on reasonable workloads Understood
Current scheduling practices Timely
Overtime expectations Adequate
Time for work activities Management of change
Time between shifts for shift changes Staffing
Sufficient number and mix of staff members
Competent to perform job responsibilities
85
Using three criteria to determine whether each cause is

Checklist 5-2.
a root cause or a contributing (or proximate) cause, the
Differentiating Root Causes following three questions may then be asked:
and Contributing Causes 1. Is it likely that the problem would have occurred if the
cause had not been present?
To differentiate root causes from contributing causes,
2. Is the problem likely to recur due to the same causal
ask the following questions of each of the causes on
factor if the cause is corrected or eliminated?
the team’s list. If the answer is “no” to each of the three
questions, the cause is a root cause. If the answer is 3. Is it likely that similar conditions will recur if the cause is
“yes” to any one of the three questions, the cause is a corrected or eliminated?
contributing cause.
If the answer to each question is “no,” then the problem is
Cause #1 a root cause. If the answer to any of the questions is “yes,”
then the problem is a contributing/proximate cause. It
1. Would the problem have occurred if Cause #1 had
not been present?
may be helpful to develop a checklist with these questions
built in. A sample checklist with the questions appears as
 No = root cause  Yes = contributing cause
Checklist 5-2, left.
2. Will the problem recur due to the same causal factor
if Cause #1 is corrected or eliminated?
STEP 11 Confirm Root Causes and
 No = root cause  Yes = contributing cause Consider Their Interrelationships
3. Will similar conditions recur if Cause #1 is corrected The team will very likely identify more than one root cause
or eliminated? for a sentinel event or near miss. Even in those very rare
 No = root cause  Yes = contributing cause instances when a sentinel event results from the intentional
act of an individual, more than one root cause is likely
Cause #2 (for example, personnel screening, communication, and so
1. Would the problem have occurred if Cause #2 had forth). Sentinel events in industry tend to have two to four
not been present? root causes, and these root causes tend to be interrelated.
 No = root cause  Yes = contributing cause To date, The Joint Commission’s Sentinel Event Database
indicates four to six root causes identified by participating
organizations for each sentinel event.
if Cause #2 is corrected or eliminated?
 No = root cause  Yes = contributing cause For example, organizations that experienced suicides in a
3. Will similar conditions recur if Cause #2 is corrected 24-hour care setting reviewed by The Joint Commission
or eliminated? identified the following root causes2:
 No = root cause  Yes = contributing cause • The environment of care, such as the presence of
nonbreakaway bars, rods, or safety rails; lack of testing of
Cause #3 breakaway hardware; and inadequate security
1. Would the problem have occurred if Cause #3 had • Patient assessment methods, such as incomplete
not been present? suicide risk assessment at intake, absent or incomplete
 No = root cause  Yes = contributing cause reassessment, and incomplete examination of patients
(for example, failure to identify contraband)
if Cause #3 is corrected or eliminated? • Staff-related factors, such as insufficient orientation or
training, incomplete competency review or credentialing,
 No = root cause  Yes = contributing cause
and inadequate staffing levels
3. Will similar conditions recur if Cause #3 is corrected • Communication issues and information-related
or eliminated?
factors, such as incomplete communication among
 No = root cause  Yes = contributing cause caregivers and information being unavailable
when needed
86
Organizations that experienced treatment delays

TIP reviewed by The Joint Commission identified the following
Organizations that have conducted a root cause root causes2:
analysis should consider having their findings and • Inadequate communication among caregivers
process reviewed by objective sources. The Joint
• Insufficient patient assessment
Commission uses a criteria-based review process to
assess the quality of data derived from a root cause
• Continuum of care issues
analysis. The process assesses the thoroughness and • Staff-related factors, such as insufficient orientation and
credibility of a root cause analysis submitted to The training and insufficient competency and credentialing
Joint Commission. processes
• Inadequate availability of information
Organizations that experienced infant abductions Organizations that experienced patient falls reviewed by
reviewed by The Joint Commission identified the following The Joint Commission identified the following root causes2:
root causes2: • Insufficient staff orientation and training
• Physical environment factors, such as no line of sight to • Inadequate caregiver communication
entry points as well as unmonitored elevator or stairwell • Inadequate assessment and reassessment
access to postpartum and nursery areas • Unsafe environment of care
• Security equipment factors, such as security equipment • Inadequate care planning and provision
not being available, operational, or used as intended
• Staff-related factors, such as insufficient orientation Although this information may provide insight into areas to
or training, competency and credentialing issues, and explore, organizations should not rely exclusively on these
insufficient staffing levels lists but should also uncover their own unique root causes.
• Communication issues
The identification of all root causes is essential to
Organizations that experienced medication errors preventing a failure or near miss. Why? Because the
reviewed by The Joint Commission identified the following interaction of the root causes is likely to be at the root of
root causes2: the problem. If an organization eliminates only one root
• Communication issues cause, it has reduced the likelihood of that one very specific
• Storage and access issues adverse outcome occurring again. But if the organization
• Staff-related factors, such as insufficient orientation misses two other root causes, it is possible that those root
and training, competency and credentialing issues, and causes could interact in another way to cause a different but
insufficient staffing levels equally adverse outcome.
• Lack of procedural compliance
The root causes collectively represent latent conditions—
Organizations that experienced wrong-site surgery conditions that exist as a consequence of management
reviewed by The Joint Commission identified the following and organizational processes and which can be identi-
root causes2: fied and corrected before they contribute to mishaps.
• Miscommunication by operating room teams The combination of root causes sets the stage for sentinel
• Insufficient orientation and training of staff events. Effective identification of all root causes and an
• Lack of procedural compliance understanding of their interaction can aid organizations
• Lack of available information in changing processes to eliminate a whole family of risks,
• Distraction not just a single risk.
• Leadership issues
If a team identifies more than four root causes, a number
of the causes may be defined too specifically. In this case,
the team may wish to determine whether one or more of
the root causes could logically be combined with another to
87
reflect more basic, system-oriented causes. The team then Health care employees often struggle with the root cause
should verify each of the remaining root causes. Doing so analysis process because it requires them to scrutinize not
involves cross-checking for accuracy and consistency all only one another but also one another’s errors. A recent
facts, tools, and techniques used to analyze information. study of root cause analysis meetings found that even when
Any inconsistencies and discrepancies should be resolved. the analysis focused on near-miss situations that did not
result in death, staff members were extremely wary about
How does an organization know whether and when it has how they positioned themselves in relation to the issues
identified all true root causes of a sentinel event? and staff involved. The “talk” of root cause analysis work is
difficult in all situations.4
Most root cause analysis teams ultimately reach a point
where they ask themselves, “When can we stop asking A team should report its root cause findings to the leaders
‘Why?’” This question is best answered by considering of its organization. Leaders must be informed, as should the
whether an identified cause is actionable in a way that will individuals likely to be affected by changes emerging from
likely prevent recurrences or otherwise protect patients the findings, during the next stage of the root cause anal-
from recurrences. If the answer is “yes,” then it might be ysis. Chapter 6 provides information on communicating
acceptable to stop there, but it is by no means necessary to the results of the team’s efforts.
stop there. Even root causes can have deeper root causes—
they usually do. References
1. Ammerman M. The Root Cause Analysis Handbook: A Simplified
Organizations often struggle to identify root causes because Approach to Identifying, Correcting, and Reporting Workplace Errors.
New York: Productivity Press, 1998.
they are reluctant to confront sensitive, politically charged
issues such as organizational culture, resources, pressure 2. The Joint Commission. Sentinel Event Data: Root Causes by
Event Type: 2004–2014. Apr 24, 2015. Accessed Oct 18, 2015.
to produce or to move patients quickly through surgery,
http://www.jointcommission.org/assets/1/18/Root_Causes_by
and lack of leadership or support. Employees typically are _Event_Type_2004-2014.pdf.
reluctant to address the problems they perceive because
3. Stecker MS. Root cause analysis. J Vasc Interv Radiol. 2007
they fear that their candor could cause repercussions within Jan;18(1 Pt 1):5–8.
the organization. As a result, for root cause analysis to be 4. Iedema RA, et al. Turning the medical gaze in upon itself: Root
successful, it must be seen as confidential and nonpunitive. cause analysis and the investigation of clinical error. Soc Sci Med.
There must also be timely, relevant feedback to personnel 2006 Apr;62(7):1605–1615.
reporting the adverse events to help them see that their
input is meaningful.3
88
Worksheet 5-1. Probing for Underlying Causes
The team might find it helpful to use a worksheet when probing for the underlying causes of proximate causes.
Proximate Causes Underlying Causes
1.
2.
3.
4.
5.
89
90
CHAPTER
6
Designing and Implementing
an Action Plan for Improvement
• Identify risk reduction strategies
STEP 1:
Organize a Team
• Set priorities and objectives for improvement in
STEP 2:
Define the Problem
areas identified as at the root of the problem
STEP 3:
Study the Problem
• Develop, test, implement, and measure the STEP 4: Determine What Happened

effectiveness of improvement efforts
STEP 5:
STEP 6:
STEP 7:
The team asks, “So what are we going to do with the prob- on Proximate and Underlying Causes
lematic systems now that we have identified them?” When
STEP 8:
the team has a solid hypothesis about one or more root Changes
causes, the next step is to explore and identify risk reduction STEP 9:
strategies to help ensure that system flaws are corrected. Involved—The Root Causes
STEP 10:
STEP 12 E
xplore and Identify Risk STEP 11:
Reduction Strategies Their Interrelationships
Organizations that have root cause analysis experience STEP 12:
55 Explore and Identify Risk Reduction
should consider conducting data analysis of multiple root Strategies
cause analyses. Sentinel and adverse events are not isolated STEP 13:
55 Formulate Improvement Actions
occurrences—the underlying root causes often have been STEP 14:
55 Evaluate Proposed Improvement
identified in past root cause analyses as contributing or root Actions
causes of other, similar occurrences. Even dissimilar events STEP 15:
55 Design Improvements
often will be found, through root cause analysis, to have STEP 16:
55 Ensure Acceptability of the Action
Plan
certain root causes in common. This kind of convergence
STEP 17:
55 Implement the Improvement Plan
to common root causes by separate root cause analyses of
STEP 18:
55 Develop Measures of Effectiveness
dissimilar events is a good demonstration that root cause
analyses are delving deeply enough. The questions root
STEP 19:
55 Evaluate Implementation of
cause analysis teams need to ask are the following: Improvement Efforts
• What have we done before? STEP 20: Take Additional Action
55
• What worked?
STEP 21:
55 Communicate the Results
• What didn’t work? Why not?
• Why are we continuing to have this type of event?
After a team is ready to move forward with a risk reduc- has been written about the engineering approach to failure
tion strategy, it might start by exploring relevant literature prevention and how it differs from the medical approach.
on risk reduction and error-prevention strategies. Much Some of the literature’s key points are described here.
91
The pervasive view of errors in the engineering field is that Taking Safety to the Next Level: High Reliability
humans err frequently and that the cause of an error is A high-reliability organization is an organization that has
often beyond the individual’s control. In designing systems succeeded in avoiding catastrophes in an environment
and processes, engineers begin with the premise that where normal accidents can be expected due to risk factors
anything can go wrong—and will. (Recall the Swiss cheese and complexity. A preoccupation with failure is one of
model discussed in Chapter 1.) Their role is a proactive one: the key characteristics of high reliability organizations
to design accordingly. Because engineering-based indus- (see Sidebar 6-1, on page 93). Organizations in the avia-
tries do not expect individuals to perform flawlessly, they tion and nuclear energy industries as well as the military
incorporate forcing functions into their designs—that is, have put programs and processes in place to become high
they try to design systems that make it extremely difficult reliability organizations.
for individuals to make mistakes. As shown in Figure 6-1,
below, forcing functions represent the strongest, most effec- In contrast, the still-pervasive view among many is that
tive intervention in systems design. By compensating for errors are the result of individual human failure and that
less-than-perfect human performance, engineering systems humans generally perform flawlessly. Hence, the design of
achieve a high degree of reliability through process stan- processes in health care organizations tends to be based on
dardization, backup systems, and designed redundancy. A the premise that nothing will ever go wrong. Education
failure rate as low as even 1% is not tolerated. The emphasis and training, more extensive in health care than in most
is on systems rather than individuals. other fields, focus on teaching professionals to do the
right thing. The assumption is that properly educated and
Figure 6-1. Strength of Intervention
This illustration shows the relationship between strength of intervention and effectiveness in systems design. The
stronger the intervention, the more effective it is.
92
CHAPTER 6 | Designing and Implementing an Action Plan for Improvement
Sidebar 6-1. to patients:

1. Design the process to minimize the risk of a failure.
Key Characteristics of High Reliability
2. Design the process to minimize the risk that a failure
Organizations will reach the patient.
3. Design the process to mitigate the effects of a failure that
►► Preoccupation with failure
reaches the patient.
►► Reluctance to simplify interpretation of problems
►► Sensitivity to operations
These levels are similar to the epidemiological concepts
►► Commitment to resilience of degrees of disease prevention, which can be defined
►► Deference to expertise as follows2:
• Primordial prevention: minimizes hazards to health
trained health care professionals will never make mistakes. • 1° prevention: reduces the risk of disease
Those who do are “retrained,” punished, or sanctioned—in • 2° prevention: reduces the prevalence of disease
essence, blamed. The immediate causes of errors are iden- • 3° prevention: reduces the impact of complications
tified and corrected but not planned or designed for. Root
causes are rarely identified. An example of a consideration for the first level of design
is ensuring a positive interlock between tubes in a venti-
Lucian L. Leape, MD, is a pioneer in comparing risk reduc- lator airflow circuit that might prevent a failure such as an
tion approaches in various other industries to those in the inadvertent disconnection of the tubes. On the second level
health care industry. He suggests four safety design charac- of design, an example of a consideration is ensuring that
teristics from the aviation industry that could, with some there is an airway pressure alarm that can detect a pressure
modification, prove useful in improving safety in the health drop due to a tubing disconnection and alert staff before a
care industry1: patient is harmed. An example of a consideration for the
• Built-in multiple buffers, automation, and third level of design is the ready availability of resuscitation
redundancy. Instrumentation in airplane cockpits equipment that can mitigate the effect of oxygen depriva-
includes multiple and purposely redundant monitoring tion due to a prolonged tubing disconnection.
instruments. The design systems assume that errors and
failures are inevitable and should be absorbed. It is also critical that error-prevention strategies used in
• Standardized procedures. Protocols that must be the health care industry include standardizing tasks and
followed exist for operating and maintaining airplanes. processes to minimize reliance on weak aspects of cogni-
• A highly developed and rigidly enforced training, tion, testing of professional performance, and institutional-
examination, and certification process. Pilots take izing safety through near miss and nonpunitive reporting.
proficiency exams every six months. For example, clinical practice guidelines and organization
• Institutionalized safety. The airline industry reports policies and protocols designed to reduce inappropriate
directly to two agencies that regulate all aspects of flying, variation in the care provided by practitioners can help
prescribe safety procedures, and investigate all accidents. reduce the likelihood of failures. In a general sense, the
A confidential safety reporting system established by tendency for a process to fail is diminished in relation to
the Aviation Safety Reporting System operated by the consistency with which it is carried out; that is, the
the National Aeronautics and Space Administration degree to which it is standardized. Yet efforts in recent
enables pilots, controllers, or others to report dangerous years to standardize health care processes through the
situations, including errors they have made, to a third introduction of practice parameters, protocols, clinical
party without penalty. This program greatly increases pathways, and so forth have been met with limited enthu-
error reporting in aviation, resulting in enhanced siasm among practitioners and are only slowly affecting
communication and prompt problem solving. the actual delivery of care. Achieving process consistency
while retaining the ability to recognize and accommodate
Organizations planning risk reduction strategies should variation in the input to the process (for example, the
consider three levels of design to reduce the risk of harm patient’s severity of illness, comorbidities, other treatments,
93
and preferences) is one of the major challenges to standard- Finally, a time-out is performed immediately prior to
ization in health care. Process variation to meet individual starting the procedure so that the medical team can confirm
patient needs is an essential principle of modern medicine; the correct patient, site, positioning, and procedure and, as
variation to meet individual health care organization or applicable, verify that all relevant documents, related infor-
practitioner preferences need not be. Standardization is mation, and necessary equipment are available.
advantageous—that is, it will get better overall results more
safely—even if each practitioner might individually get Risk points—specific points in a process that are suscep-
better results than the others by using a personally favored tible to failure or system breakdown—must be eliminated
but different process than that used by others. Practitioners through design or redesign efforts. They generally result
do not work alone within health care organizations; they from a flaw in the initial design of the process or system,
are members of teams, and those teams interact with other a high degree of dependence on communication, nonstan-
teams. Thus, assuming each personally favored practice is a dardized processes or systems, and/or failure or absence
good practice, it matters less which process is selected as the of backup.
basis for standardization than that the process is performed
consistently for safety. For example, risk points during the medication use process
include interpretation of an illegible order by a pharmacist
A Systems Approach to Risk Reduction and the time during which a registered nurse mixes the medi-
Risk reduction strategies must emphasize a systems rather cation dose to administer to a patient. In surgical procedures
than an individual human approach. A system can be requiring the use of lasers, a risk point occurs during the use
thought of as any collection of components and the rela- of anesthetic gases: The high concentration of oxygen, if not
tionships between them, whether the components are properly synchronized with use of the laser, can allow tubes,
human or not, when the components have been brought drapes, and other potentially flammable materials to ignite.
together for a well-defined goal or purpose.3 As Leape During preoperative procedures, verification of the body side
writes, “Creating a safe process, whether it be flying an and site constitutes a risk point.
airplane, running a hospital, or performing cardiac surgery,
requires attention to methods of error reduction at each Built-in buffers and redundancy, task and process simplifi-
stage of system development: design, construction, main- cation and standardization, and training are all appropriate
tenance, allocation of resources, training, and development design mechanisms to reduce the likelihood of failure at
of operational procedures.”1(p. 1854) If errors are made, if defi- risk points and elsewhere. For example, prior to the admin-
ciencies are discovered, individuals at each stage must revisit istration of medications, multiple and redundant checks
previous decisions and redesign or reorganize the process. such as asking the patient his or her name and checking the
patient’s armband can help confirm that a medication is
Designing for safety means making it difficult for humans given to the right patient. See Sidebar 6-2, on page 97, for
to err. However, those designing systems must recognize risk points for medication errors and risk reduction strate-
that failures do occur and that recovery or correction should gies to prevent such failures.
be built into the system. If that is not possible, failures must
be detected promptly so that individuals have time to take Systems engineering literature includes numerous other
corrective actions. For example, as required by The Joint design concepts that could be useful tools to prevent failures
Commission’s Universal Protocol for Preventing Wrong and sentinel events in health care organizations. Redundancy
Site, Wrong Procedure, Wrong Person Surgery™, a prepro- is one such concept familiar to the aerospace and nuclear
cedure verification process must be used. For example, a power industries, where systems have backups and even the
checklist could be used to confirm that appropriate docu- backups normally have backups. System reliability can be
ments (such as medical records and imaging studies) are increased by introducing redundancy into system design.
available. The Universal Protocol also states that the proce- However, the cost of designing in redundancy is an issue in
dure site must be marked. Ideally, the marking takes place most health care environments. The engineering literature
with the patient involved, awake and aware, if possible. also describes the benefits of simplification, standardization,
94
Sidebar 6-2. Medication Errors: Risk Points and Risk Reduction Strategies
Risk points or common causes: • Use auxiliary labels (such as “for IM

• Training/education [intramuscular] only”).
• Competence (lapses in performance, failure to • Support questioning of unclear orders.
comply with policies and procedures) • Eliminate guessing.
• Supervision
To reduce the likelihood of failures associated with access
• Staffing (excessive workload, incorrect mix)
to medications:
• Communication
• Remove high-risk medications from care units.
• Distraction due to environmental issues
• Label high-risk medications as such.
• Information availability
• Establish and implement policies and procedures
• Medication storage and access
for use of off-hours pharmacy.
• Labeling
• Nomenclature To reduce the likelihood of failures associated with
• Dosage calculation medication delivery:
• Equipment • Be sure that equipment defaults to the
• Abbreviations least-harmful mode.
• Handwriting • Use automated pharmacy units as a tool for
improving the process, not an inherent solution.
Risk reduction strategies:
• Recognize that polypharmacy equals higher risk.
To reduce the likelihood of failures associated with
prescribing errors: To reduce the likelihood of failures associated with
• Implement a system of computerized order human resources and competence factors:
entry by physicians (to decrease the likelihood • Address education and training issues
of dosage error, prompt for allergies, and (orientation, competence assessment, and
provide information on drug–drug and drug–food training with new medications and devices).
interactions). • Support professional ethics and judgment.
• Redefine the role of pharmacists to enable them • Implement systems involving double checks.
to perform daily rounds with physicians, work • Make safe staffing choices.
with registered nurses, and serve as on-site • Avoid reliance on illegible handwriting by
resources. providing alternatives such as requiring orders to
be printed or using electronic order entry.
To reduce the likelihood of failures associated with
• Discourage use of acronyms and abbreviations in
dispensing:
general and prohibit the use of specific, high-risk
• Do not rely on color coding.
acronyms and abbreviations.
• Remove or separate look-alike/sound-alike
• Control availability of high-risk drugs.
medications.
• Address environmental issues such as
• Use bar coding if possible.
distraction.
• Avoid lethal medications in bolus form.
• Standardize medication times.
• Use premixed solutions whenever possible.
• Use patients as safety partners.
• Minimize supplier and product changes.
and loose coupling to reduce the possibility of systems- making an electrical system safe. A destruct system on a
related problems. satellite or an air-to-air missile is an example of a fail-active
device. If the satellite or missile misses its target within a set
Fail-safe design is also a concept familiar to high reliability time, the destruct system blows the satellite or missile apart
industries, including aerospace and nuclear engineering. to halt its flight and limit any damage it might cause by
The design may be fail-passive, fail-operational, or fail-ac- falling to the ground. Checklist 6-1, on page 98, provides
tive. For example, a circuit breaker is a fail-passive device the key risk reduction strategies suggested in the medical
that opens when a dangerous situation occurs, thereby and engineering literature.
95
Risk points and risk reduction strategies for wrong-site

Checklist 6-1.
surgery, suicide, and infant abductions or release to wrong
families follow. Not all these strategies are specific Joint Identifying Risk Reduction Strategies
Commission requirements, but they are presented for
To reduce the likelihood of failures, the medical and
consideration by all health care organizations.
engineering literature offers the following tips:
Risk points for wrong-site surgery include the following: 33 Use an engineering (proactive, systems-based)
approach to failure prevention.
• Communication before reaching operating suite
• Communication in operating suite 33 Start with the premise that anything can and will go
wrong.
• Hierarchical issues of communication
• Communication with patient and patient’s family 33 Design systems that make the safest thing to do
the easiest thing to do.
• Information availability
33 Design systems that make it difficult for individuals
• Multiple surgery sites on patient’s body to err.
• Conflicting chart information
33 Build in as much redundancy as possible.
• Confused patient
• X-ray quality, labeling, and accuracy of interpretation
33 Use fail-safe design whenever possible.
33 Simplify and standardize procedures.
Risk reduction strategies to reduce the likelihood of failures 33 Automate procedures.
associated with preoperative procedures include the following: 33 Ensure rigidly enforced training and competence
assessment processes.
• Mark the operative site.
• Require the surgeon to obtain informed consent. 33 Ensure nonpunitive reporting of near misses.
• Require preoperative verification by the surgeon, 33 Eliminate risk points.
anesthesiologist/anesthetist, and patient or family.
• Personally review x-rays. Strategies to reduce the risk of suicide in a staffed around-
• Revise equipment setup procedures. the-clock care setting include the following:
• Revise assessment and reassessment procedures and
To reduce the likelihood of failures in the operating suite, do assure adherence.
the following: • Update the staffing model.
• Verify patient identity, intended procedure, and • Educate staff on suicide risk factors.
operative site before prep and drape. • Update policies on patient observation.
• Make sure site marking is visible after draping. • Monitor consistency of implementation.
• Obtain verbal verification with a time-out. • Revise information transfer procedures.
• Confirm the level of spinal surgery with intraoperative • Revisit contraband policies.
fluoroscopy. • Identify and remove nonbreakaway hardware.
• Weight test all breakaway hardware.
Risk points for suicide include the following: • Redesign or retrofit security measures.
• Suicide risk assessment • Educate family and friends on suicide risk factors.
• Communication of risk to direct care staff • Consider patients in all areas.
• Levels of monitoring required for those at risk • Ensure that staff members ask about suicidal thoughts
• Policies for risk assessment and monitoring outside every shift.
of behavioral health care units (for example, in the • Be cautious at times of shift change (admission,
emergency department) discharge, passes).
• Interventions for high-risk patients • Avoid reliance on pacts.
• Psychotropic medications • Be suspicious if symptoms lighten suddenly.
• Contraband • Involve all staff in solutions.
• Physical environment, including such potential hazards
as door frames, doorknobs, shower heads, and bedsheets
96
Risk points for infant abduction include the following: and effects analysis (FMEA) to identify risk reduction
• Mother and infant identification opportunities. Also known in the literature as failure mode,
• Staff identification effects, and criticality analysis (FMECA), FMEA offers
• Visitor identification a systematic way of examining a design prospectively for
• Physical layout of the obstetrics unit possible ways in which failure can occur. Potential failures
• Entry and exit security are identified in terms of failure modes or symptoms. For
• Mother’s education level or ability to comprehend each failure mode, the effect on the total system or process
security instructions is studied, and factors that might cause or enable those fail-
• Policy for issuing alarms when an infant is missing ures are identified.
• Public access to birth information
Actions (planned or already taken) can be reviewed for their
Strategies to reduce the risk of infant abductions include potential to minimize the probability of failure or to reduce
the following: the effects of failure. FMEA’s goal is to prevent poor results,
• Develop and implement a proactive infant abduction which in health care means harm to patients. Its greatest
prevention plan. strength lies in its capability to focus users on the process of
• Include information on visitor and provider redesigning potentially problematic processes to prevent the
identification as well as identification of potential occurrence of failures.
abductors or abduction situations during staff
orientation and in-service curriculum programs. Although the technique has been used effectively in the
• Enhance parent education concerning abduction risks engineering world since the 1960s, its use in the health care
and parent responsibility for reducing risk, and then world began as late as the 1990s. FMEA is now gaining
assess the parents’ level of understanding. broader acceptance in health care as a tool for prospective
• Attach secure identically numbered bands to the baby analysis due to the efforts of The Joint Commission, the
(wrist and ankle bands), mother, and father or significant Department of Veterans Affairs National Center for Patient
other immediately after birth. Safety, and the Institute for Safe Medication Practices,
• Footprint the baby, take a color photograph of the baby, among others. FMEA is discussed further in Chapter 7.
and record the baby’s physical examination within two
hours of birth. FMEA is described here because root cause analysis teams
• Require staff to wear up-to-date, conspicuous, color may also wish to use FMEA to proactively identify risk
photograph identification badges. reduction opportunities during their root cause analysis of
• Discontinue publication of birth notices in local a sentinel event or near miss or to carry out a proactive risk
newspapers. assessment on a process that is being redesigned in response
• Consider options for controlling access to the nursery or to the findings of a root cause analysis. The interrelation-
postpartum unit, such as swipe-card locks, keypad locks, ship between FMEA and root cause analysis is further
entry point alarms, or video surveillance. (Any locking discussed in Sidebar 6-3, on page 98.
systems must comply with fire codes.)
• Consider implementing an infant security tag or A performance improvement team at a community hospital
abduction alarm system. in Michigan used FMEA for one of the first times in health
care to proactively analyze and reduce medication errors
Failure mode and effects analysis associated with potassium chloride.5 The focus was on
(FMEA) developing strategies to reduce the risk of future fatal errors.
The team followed the steps outlined in Sidebar 6-4, on
Root Cause Analysis in Proactive Risk Assessment page 98.
Expanding the use of risk assessment has emerged as a
strong trend in health care. Facilities can implement a The process flow diagram developed by the team for the
strategy of assessing risk proactively by using failure mode medication use process from point of initiation through
97
Sidebar 6-3: Sidebar 6-4: Steps in FMEA

The Interrelationship Between
FMEA and RCA 1. Select a high-risk process and assemble a team.
2. Diagram the process.
Failure mode and effects analysis (FMEA) and root 3. Brainstorm potential failure modes and determine
cause analysis (RCA) differ in critical ways but also the effects of the failure modes.
share similarities. 4. Prioritize failure modes.
The fundamental difference is timing: RCA is a retro- 5. Identify root causes of failure modes.
spective approach, while FMEA is designed to keep 6. Redesign the process.
sentinel events from occurring in the first place. RCA 7. Analyze and test the new process.
asks “Why?” after an event occurred to identify the root 8. Implement and monitor the redesigned process.
causes of an event. FMEA asks “What if?” to explore
what could happen if a failure occurred at a particular Source: Joint Commission on Accreditation of Healthcare Organizations:
Failure Mode and Effects Analysis in Health Care: Proactive Risk
step in a process or link.
Reduction, 2nd ed. Oakbrook Terrace, IL: Joint Commission Resources,
FMEA and RCA have the following characteristics 2005, p. 5..
in common:
►► Nonstatistical method of analysis
►► Goal of reducing the possibility of future patient

analysis, or new members might be brought on board as
harm required by the recommended improvements.
►► Involves identifying conditions that lead to harm
Improvement actions fall into three basic categories:
►► Team activity 1. Actions to prevent errors
In addition, the two methodologies can be—and often 2. Actions to shield the patient from the effects of an error
are—interconnected. FMEA can be used during an 3. Actions to mitigate the effects of an error that reaches
RCA to help evaluate various improvement strategies the patient
that resulted from the RCA. FMEA can look at where
the various strategies might fail and identify any new “Mistake Proofing”
failure modes that have been introduced as a result of The ultimate goal of a root cause analysis is the develop-
new design processes. RCA can be used to identify the
ment of actions to reduce the potential for recurrence of
root causes of failure modes.
a similar event. The initial focus of improvement actions
Source: Adapted from Joint Commission International. Understanding should be on eliminating the circumstances that allowed
and Preventing Sentinel and Adverse Events in Your Health Care
Organization. Oak Brook, IL: Joint Commission Resources, 2008,
the outcome. If no action can be applied to eliminate the
pp. 89–90. cause, the team should seek appropriate measures to reduce
the possibility of recurrence.6
completion appears as Figure 6-2, page 99. The team’s As discussed in Step 12, the most effective actions are
outline of what could go wrong and its ranking appears as developed by using a systematic approach to mistake-
Sidebar 6-5, page 100. proofing processes to avoid inevitable human error in
care delivery. An early example of incorporating mistake
STEP 13 Formulate Improvement Actions proofing in a design was the 3.5-inch diskette.7 The diskette
With the list of root causes in hand, the team is now ready could be inserted only if it was oriented correctly. It could
to start devising potential solutions to systems-related not be inserted sideways because it was not square; the
problems. Known as corrective actions or improvement sides were too long to fit. It could not be inserted backward
actions, these solutions are required to prevent a problem or inverted.
from occurring or recurring due to the same root cause(s)
or interaction of root causes. The team may include the The drive was designed to stop the diskette unless the right
same members as during the early stages of the root cause front corner was chamfered (angled). When the disk was
98
Figure 6-2. The Medication Use Process at a Community Hospital
This figure illustrates the simplified process flowchart one community hospital created as step 1 in a failure mode and
effects analysis. In Step 2 the organization went through the flowchart, action by action, to brainstorm what might
go wrong. Sidebar 6.5, page 100, lists the possible risk points the hospital staff brainstormed. CMAR, computerized
medication administration record.
Source: Fletcher CE. Failure mode and effects analysis: An interdisciplinary way to analyze and reduce medication errors. J Nurs Adm. 1997
Dec;27(12):19–26. Reproduced by permission.
99
Sidebar 6-5. Possible Risk Points in the Medication Use Process
For each risk point in the process, a list of possible failures Order Transcribed onto CMAR
follows. • RN’s handwriting unclear.
• Order transcribed onto wrong computerized
Order Received
medical administration record (CMAR).
• Phone order not clarified.
• Order transcribed incorrectly.
• Verbal order not clarified.
• Order not transcribed in a timely manner.
• Written order illegible.
• Lack of nursing communication to RN or licensed
• Dose may be incorrect.
practical nurse (LPN) giving medication.
• Incorrect route.
• Forgetting or failure to transcribe.
• Order written on wrong chart.
• RN Mixes Medication
• Order written on right chart, but stamped with
• Miscalculated dose.
wrong name.
• Miscalculated measurement of volume.
• Wrong drug.
• Selects wrong drug.
• Drug not indicated.
• Prepares wrong drug brought by supervisor (after
• Language barrier present with verbal phone
hours).
order.
• Does not match the order to the chart.
Unit Secretary Processes Order • Labeling errors: drug is not actually added to IV
• Does not take order out of chart. or IV contains drug but no label.
• Sends order to department other than pharmacy. • Lacks knowledge of administration policies and
• Misplaces order and does not send it at all. procedures.
• Misinterprets and thus processes incorrectly. • Lacks knowledge regarding IV versus oral dosing
• Delays in processing occur. guidelines.
• Stamps wrong name on order sheet. • Interrupted during preparation.
• Lacks routine or organization when doing the
Registered Nurse Signs Off Order
task.
• Delays order.
• Does not read order carefully but processes it RN or LPN Administers Medication to Patient
anyway. • Wrong patient.
• Does not take time to read order but processes it • Incorrect labeling.
anyway. • Intravenous accurate control (IVAC) not used.
• Cosigns order without giving it to secretary; does • No IVAC available.
not go through correct process. • Wrong rate of flow set on IVAC.
• Cannot read or misreads order.
• Orders added by physician after initial order
processed; registered nurse (RN) does not see
additional orders.
Source: Fletcher CE. Failure mode and effects analysis: An inter
• Unfamiliar with drug: allergy, dose, and/or disciplinary way to analyze and reduce medication errors. J Nurs Adm.
cross sensitivity. 1997 Dec;27(12):19–26. Reproduced by permission.
inserted correctly, the mistake-proofing device was not for using design as an error-prevention strategy instead of
noticeable. When it was inserted incorrectly, however, the alternatives such as training, instructions, or warning labels.
device completely stopped the process. The only cost was
that of initial design implementation. No user training A more recent attempt to incorporate mistake proofing in
was required. design is the “smart pump” system for medication infusion.
This system prevents errors by “remembering” the organi-
The members of the design team that created the disk drive zation’s defined dosing limits and other clinical advisories
believed that getting the orientation right was important entered into the medication library and applying those
enough to design a process that allowed users only one way “rules” during pump programming, thereby helping ensure
to use the device. Their decision also indicated a preference that the right dosage of the medication is infused. The
100
system also warns clinicians about potential unsafe drug analysis teams could approach the identification of
therapy before the medication is administered.8 improvement actions.
When formulating action plans, teams also should In the suicide example, the team has completed a fishbone
analyze the strength of their proposed solutions. The US diagram indicating multiple system problems, including
Department of Veterans Affairs National Center for Patient assessment of suicide risk, environment of care, and emer-
Safety has devised a hierarchy of corrective actions that can gency procedures. The team might break into smaller
be used to gauge the effectiveness of improvement actions subgroups. One group (including the psychiatrist, medical
based on how likely they are to reduce vulnerability to an staff leader, and nurse) would address the failed patient
adverse event.9 assessment process. They might start by reviewing the
current standard for assessment of suicide risk and how this
Stronger actions typically involve modifications to the standard is communicated in the behavioral health care
physical environment. For example, after a patient unit. Another group (including the administrator and plant
attempted to commit suicide by hanging from an exposed safety representative) might start working on environment
pipe, the pipe was removed and the hazard thus eliminated. of care issues such as nonbreakaway shower heads. Another
Weaker actions include changes in policies, procedures, group (including emergency department physicians and the
and training. Although they are necessary components nursing staff) might work on emergency procedures.
of an improvement action, they are considered weaker
because they do not change the underlying conditions that In the elopement example, the root cause analysis team has
lead to errors, they tend to rely on human cooperation identified multiple system problems, including an unsafe
and vigilance to succeed, and they often show less imme- environment of care; inadequate assessment and reas-
diate results. In a study of root cause analysis action plans sessment; and inadequate staff orientation, training, and
reported by the NCPS, stronger actions were easier to ongoing competence assessment. One subgroup (including
implement and more effective than weaker actions.9 the safety director, a nurse from the unit, and a social
worker) might address possible actions to improve the
Improvement actions should be formulated by thinking in long term care organization’s security and safety measures.
terms of the everyday work of the organization. Work can Another small group (including the medical director, director
be defined in terms of functions or processes. A function of nursing, activity staff member, and unit staff nurse) might
is a group of processes with a common goal, and a process address opportunities to improve the process used to assess
is a series of linked, goal-directed activities. Improvement and reassess patients at risk for elopement. Another group
actions should be directed primarily at processes. As stated (including the medical director, the director of nursing, and
earlier in this book, process improvement holds the greatest the performance improvement coordinator) might address
opportunity for significant change, whereas changes related strategies to ensure that staff know elopement risk factors and
to an individual’s performance tend to have limited effect. who is at risk for elopement and are assessed regularly for
Competent people often find themselves carrying out competence in identifying and caring for at-risk patients.
flawed processes.
In the treatment delay example, the team has identi-
Practice guidelines or parameters and other standardized fied system problems that led to the missed diagnosis of
patient care procedures are useful reference points for metastasized breast cancer. They include communication
comparison. Whether developed by professional societies or problems between caregivers, insufficient staff orientation
in-house practitioners, these procedures represent an expert and training, and inadequate information management.
consensus about the expected practices for a given diagnosis Subgroups of the ambulatory health care organization’s
or treatment. Assessing variation from such established proce- team probe each of these areas for improvement opportu-
dures can help the team identify how to improve a process. nities, looking at such issues as care documentation and the
availability of clinical records, the timeliness and thorough-
Returning to the sentinel events described in Chapter 3, ness of initial and regular reassessments, and shift-to-shift
consider the following examples of how root cause communication of information related to patient needs.
101
In the medication error example, the root cause analysis 5. Objectively test and revise the solutions.
team has identified communication of medication orders 6. Develop a final list of potential solutions.
and the failure to ensure safe medication storage and access
as two key problems, among others. A subgroup (including These steps may assist the team through the process of both
the information technology staff member, pharmacist, developing and evaluating improvement actions.
medical director, and home health nurse) might investi-
gate possible strategies to improve the accuracy of orders STEP 14 E
valuate Proposed
communicated to local pharmacies. Another subgroup, Improvement Actions
including the nursing supervisor, pharmacy supplier, home When the list of possible improvement actions is as complete
health nurse, and medical director, might investigate strat- as possible, the team is ready to evaluate the alternatives and
egies to guard against medication theft and ensure proper select those actions to be recommended to leadership.
implementation of the home health agency’s medication
administration policies and procedures. To begin the evaluation process, the team should rank the
ideas based on criteria defined by the team. Gathering
For each root cause, the team should work interactively appropriate data is critical to this process. A simple 6-point
either as a whole or in smaller groups to develop a list of scale ranging from a low rank of 0 for the worst alterna-
possible improvement actions. Brainstorming can be used tive to a high rank of 5 for the best alternative can be used
to generate additional ideas. The emphasis at this point is at this point. Spath suggests asking the team to rank the
on generating as many improvement actions as possible, solutions in order of workability, reliability, risk, chance for
not on evaluating the ideas or their feasibility. The number success, management/staff/physician receptivity, cost, capa-
of suggested improvement actions may vary based on the bility to fix the problem, and other factors.11 To rank the
nature of the root cause and how it relates to other root proposed solutions, Ammerman suggests using criteria such
causes. To ensure as thorough a list as possible, the team as compatibility with other organization commitments and
may wish to review the analyses of information used to the possible creation of other adverse effects.12
identify root causes. Remember to encourage any and all
ideas without critiquing them. In the hands of a skilled Initially, to prevent groupthink, it is a good idea to ask
facilitator, even the seemingly wildest idea can lead to an each team member to rank the ideas on his or her own.
effective improvement action during later stages of anal- The rankings can then be consolidated into a team ranking.
ysis. Tools used such as flowcharts or fishbone diagrams can Worksheet 6-1 at the end of the chapter, page 119, can
prompt additional solutions. Ask questions of the group, be used to keep track of suggested improvement actions.
such as the following: Record the rankings assigned by individual team members
• What might fix this problem? and the team as a whole. FMEA may be a helpful tool at
• What other solutions can we generate? this point in the process. FMEA involves evaluating poten-
• What other ideas haven’t we thought of? tial problems (or improvement actions) and prioritizing
or ranking them on a proactive basis according to criteria
Brainstorming, flowchart, fishbone defined by a team.
diagram
Failure mode and effects analysis
Wilson and colleagues suggest using the scientific method to (FMEA)
develop a list of potential solutions. They restate the scientific
method in terms of steps used in developing solutions10: At the very least, every improvement action proposed by
1. Become familiar with all the aspects of the problem and the team should be objective and measurable. If it is objec-
its causes. tive, implementation is easier and those affected by the
2. Derive a number of tentative solutions. change are more likely to be receptive. If it is measurable,
3. Assemble as much detail as is needed to clearly define the team can ensure that improvement actually occurs. See
what is required to implement these solutions. Sidebar 6-6, right, for evaluative criteria for improvement
4. Evaluate the suggested solutions. actions. Before ranking the actions, ensure that the team
102
reaches a consensus on which criteria are most relevant to Sidebar 6-6.

the organization. Ranking the proposed ideas according to
Evaluating Improvement Actions
multiple criteria adds critical dimension to the evaluation.
Use the following criteria to evaluate improvement
In evaluating potential improvement actions, the team actions:
should consider the impact of the suggested improve-
►► Likelihood of success (preventing recurrence or
ment on organization processes, resources, and schedules.
occurrence) within the organization’s capabilities
Sentinel events or near misses frequently shake up the orga-
►► Compatibility with organization’s objectives
nization’s notions of the resources that should be expended
in particular areas. Organizations contemplating a design ►► Risk
or redesign effort will certainly weigh the availability ►► Reliability
of resources against the potential benefits for patients,
►► Likelihood to engender other adverse effects
customers, and the organization.
►► Receptivity by management/staff/physicians
Asking some key questions helps the team identify the ►► Barriers to implementation
potential barriers to implementation of each improvement ►► Implementation time
action. Relevant questions include the following.
►► Long-term (versus short-term) solution
►► Cost
Organization Processes
How does the proposed action relate to other projects ►► Measurability
currently under way in the organization? Are there
redundancies?
• How does the action affect other areas and processes? • What initial and ongoing training will be required?
• What process-related changes might be required? How will this impact the schedule, and how will its
• Can affected areas absorb the changes or additional impact be handled?
responsibilities?
Potential Negative Consequences
Resources • Could this action cause problems in other areas or have a
• What financial resources will be required to implement negative impact on other processes?
the action? (Include both direct and indirect costs—that • Is there a process in place to analyze and take action
is, costs associated with the necessary changes to other if there are unintended, negative consequences from
procedures and processes.) How will these resources be implementing this action?
obtained? • Will the amount of resources required for this action
• What other resources (staff, time, management) are detract from other quality improvement or patient safety
required for successful implementation? How will these initiatives?
resources be obtained?
• What resources (capital, staff, time, management) are With answers to these questions in hand, the team can
required for continued effectiveness? How will these better gauge whether the pluses outweigh the minuses.
resources be obtained? After completing this questioning process, the team may
• What other activities will have fewer resources as a result wish to revisit the ranking exercises described previously.
of shifting resource allocation for this change? Doing so can help clarify which corrective improvement
actions should be selected. To summarize the potential
Schedule of each proposed action, the team can ask, “What will
• In what time frame can implementation be completed? result from implementing this action?” and “What would
• Who will be responsible for making sure it happens? result from not implementing this action?” (as shown in
• How will implementation of this action affect other Worksheet 6-2 at the end of this chapter, page 120).
schedules? How can this be handled?
103
At this point, the team should be ready to select a finite resulting from the design or improvement. Without these
number of improvement actions. Each action must do expectations, the organization cannot determine the degree
the following: of success of the efforts. These expectations can be derived
• Address a root cause from staff expertise, consumer expectations, experiences
• Offer a long-term solution to the problem of other organizations, recognized standards, and other
• Have a greater positive than negative impact on other sources. What sequence of activities and resources will be
processes, resources, and schedules required to meet these expectations? How and what will the
• Be objective and measurable team measure to determine whether the process is actually
• Have a clearly defined implementation time line performing at the level expected?
• Be assignable to staff for implementation
• Be acceptable to staff most directly affected by The organization or group needs specific tools to measure
the change the performance of the newly designed or improved process
to determine whether expectations are met. These measures
The next section describes how the team designs improve- can be taken directly, adapted from other sources, or newly
ments and develops an action plan covering each of created, as appropriate. It is important for the measures to
these aspects. be as quantitative as possible, meaning that the measure-
ment can be represented by a scale or range of values. For
STEP 15 Design Improvements example, if improving staff competence in calculating medi-
The product of the root cause analysis is an action plan that cation doses is cited as a corrective solution, the measure
identifies the strategies that the organization intends to should evaluate competence before and after each training
implement to reduce the risk of similar events occurring in or educational session. If pretraining competence is tested
the future. The team is now ready to start drafting such a at 80% to 85% proficiency, posttraining competence might
plan. The plan should address the five issues of what, how, be set at 90% to 95% proficiency. Or, in the patient suicide
when, who, and where involved in implementing and evalu- example described in Sidebar 3-1, page 46, measures might
ating the effectiveness of proposed improvement actions. include the percentage of accurately and appropriately
completed suicide risk assessments as determined through
Issue One: What peer review and the percentage of rooms with breakaway
Designing what involves determining the scope of the shower fixtures.
actions and specific activities that will be recommended.
A clear definition of the goals is critical. To understand At times, it may be difficult to establish quantitative
the potential effects of the improvement activity, the measures—the improvement simply seems to lend itself
organization must determine which dimension of perfor- more to qualitative measures. Quantification of improve-
mance—safety, effectiveness, patient-centeredness, time- ment is critical, however, and even when solutions can
liness, efficiency, and equity13—will be affected. At times, be measured only in terms of risk reduction potential, it
the relationship between two or more dimensions must be is important to try to quantify such potential as much as
considered. Redesign in response to a sentinel event most possible through concrete measures. Use Worksheet 6-4 at
often focuses on safety, but it may affect any or all of the the end of the chapter, page 122, to articulate responses to
other dimensions. What specific activities will be needed questions concerning expectations, the sequence of activi-
to achieve the necessary improvement? Use Worksheet 6-3 ties, measures, and resources required.
at the end of the chapter, page 121, to articulate responses
to questions concerning goals, dimensions of performance, Issue Three: When
and required specific activities. Next, the team must define when the organization must
meet its improvement goals. What time frame will be estab-
Issue Two: How lished for implementing the improvement action? What
How does the organization expect, want, and need the time line will be established for each activity comprising
improved process to perform? The team carrying out the the steps along the way? What are the major milestones and
effort should set specific expectations for performance their respective dates? A Gantt chart of one organization’s
104
improvement plan appears as Figure 6-3, pages 106–107. staff members? Are the location, target population, and target
Use Worksheet 6-5 at the end of the chapter, page 123, to staff of the improvement action likely to expand with success?
articulate responses to questions concerning time frames Use Worksheet 6-7 at the end of the chapter, page 126, to
and milestones. indicate where the improvement action will be implemented.
Worksheet 6-8, page 127, can be used to provide a summary
Gantt chart look at the what, how, when, who, and where involved in
implementing proposed improvement actions.
Issue Four: Who Considering the Impact of Change

Who is closest to this process and therefore should “own” When designing improvements, the team should also
the improvement activity? Who should be accountable at consider the impact of change on the organization. No
various stages? To a great extent, the success of an improve- matter how minor, improvements require change, and it is
ment effort hinges on involving the right people from all normal for individuals and organizations to resist change.
disciplines, services, and offices involved in the process Resistance to change can come from inertia, the challenge
being addressed. The process for taking action consists of of managing the change process, the challenge of obtaining
several stages, each of which may have different players. necessary knowledge to ensure that the change can be
implemented effectively, and resource limitations.
The group that creates the process should include the
people responsible for the process, the people who carry Change comes in two basic forms: Physical and behavioral.
out the process, and the people affected by the process. As A physical change as a result of a root cause analysis might
appropriate, the group members could include staff from be the use of new equipment, such as medication adminis-
different units, branch offices, or teams, services, disciplines, tration carts. Behavioral change occurs when staff members
and job categories. When the group needs a perspective are asked to abandon established processes and procedures.
not offered by its representatives, it should conduct inter-
views or surveys outside the group or invite new members Understanding the underlying rationale and process
into the group. It is important to consider customers and whereby people modify their thinking and practice can help
suppliers such as purchasers, payers, physicians, referral team leaders gain better insight into why people behave the
sources, accreditors, regulators, and the community as a way they do, which in turn can help leaders identify moti-
whole. See Worksheet 6-6 at the end of the chapter, pages vators for helping personnel adopt and maintain desired
124–125, for more information on key players at each stage. behaviors. It can also help direct project resources to areas
where they are likely to have the greatest impact.
Leaders and managers must take an active role in over-
seeing and setting priorities for design and redesign. Stages of change theory is based on the idea that when indi-
Generally, managers are responsible for processes within viduals attempt to change a behavior, they progress along
their areas. Design or redesign of processes with a wider a continuum comprising five phases: Precontemplation,
scope may be overseen by upper management or by a team contemplation, preparation, action, and maintenance.14
of managers. Leaders must ensure that the people involved By gauging where stakeholders are on the continuum, root
have the necessary resources and expertise. Furthermore, cause analysis teams are can identify the motivators that
their authority to make changes should be commensu- encourage behavioral change.
rate with their responsibility for process improvements.
Although regular feedback and contact with management For example, an organization planning to introduce stan-
are important, rigid control can stifle creativity. dardized order sets learns through surveys, interviews, and
in-person discussions that cardiologists have not thought
Issue Five: Where about standardizing their order sets. Each cardiologist uses
Where will the improvement action be implemented? Will his or her own set and is content with it that way. These
its implementation be organizationwide, in a selected loca- physicians do not see a need for change; they are in the
tion, with a selected patient population, or with selected precontemplation stage.
105
Figure 6-3. Improvement Plan and Implementation Status
(continued)
106
Figure 6-3. Improvement Plan and Implementation Status (continued)
This Gantt chart shows a detailed plan created by an organization following a sentinel event involving a mechanical
failure. The chart shows the plan’s steps in implementing strategies, priorities, and expected time frames. The
status of each phase is recorded so that everyone involved has a clear idea of the progress being made.
107
The transition from precontemplation to contemplation is a

Checklist 6-2.
difficult one. People who do not recognize a need to change
are less likely to be willing to change behavior than people
Criteria for an Acceptable Action Plan
who are already considering it. In this situation, the team
Check to ensure that the action plan has the following
would have more success implementing standardized order attributes:
sets with a group of physicians who already recognize the
potential benefits. 33 Identifies changes to reduce risk or provides
rationale for not undertaking changes
These physicians would be in the contemplation stage, and 33 Identifies who is responsible for implementation
when provided with the necessary information, resources 33 Identifies when action(s) will be implemented
(for example, template order sets, best practices, the latest 33 Identifies how the effectiveness of action(s) will
be evaluated
evidence and quality measures), and tools (such as content
and process management systems), they will likely be ready
and willing to adopt the new behavior. Checklist 6-2, above, lists the criteria for an acceptable
action plan.
The team can identify areas where resistance to change
might arise and plan countermeasures using Worksheet 6-9 STEP 17 Implement the Improvement Plan
at the end of the chapter, page 128. When the goals for improvement have been established,
the organization can start planning and carrying out the
STEP 16 Ensure Acceptability of the improvements. A pilot test implementing improvement
Action Plan on a small scale, monitoring its results, and refining the
The team has defined the what, how, when, who, and where improvement actions is highly recommended. The pilot
in an improvement action plan. How does the team know test enables the team to ensure that the improvement is
whether it is acceptable to The Joint Commission or Joint successful before committing significant organization
Commission International (JCI) as part of a root cause resources. Pilot testing also aids in building support for the
analysis in response to a sentinel event? improvement plan, thereby facilitating buy-in by opinion
leaders. To pilot-test an improvement, the team should
As mentioned in Chapter 1, an action plan is considered follow a systematic method that includes performing
acceptable by The Joint Commission or JCI if it does the the subsequently described Steps 18 through 21 on a
following: limited scale.
• Identifies changes that can be implemented to reduce
risk or formulates a rationale for not undertaking such A systematic method for design or improvement of
changes processes can help organizations pursue identified oppor-
• Where improvement actions are planned, identifies who tunities. A standard yet flexible process for carrying out
is responsible for implementation, when the action will these changes should help leaders and others ensure that
be implemented (including any pilot testing), and how actions address root causes, involve appropriate personnel,
the effectiveness of the actions will be evaluated and result in desired and sustained changes. Depending
on an organization’s mission and improvement goals, any
The action plan also should include resource considerations, of the methods described here may be used to implement
leadership approval, and buy-in by affected staff. Teams a process improvement. Three improvement methods are
should consider conducting an FMEA of the proposed described:
action plan and then proceeding with some pilot testing. • The scientific method
• The Plan-Do-Study-Act (PDSA) cycle
• The Define, Measure, Analyze, Improve, Control
(DMAIC) process
108
The Scientific Method 5. Assess the results of the test. (Compare results of before
The fundamental components of any improvement process versus after the change.)
are the following: 6. Implement successful improvements, or rehypothesize
• Planning the change and conduct another experiment.
• Testing the change
• Studying its effects This orderly, logical, inclusive process for improvement
• Implementing changes determined to be worthwhile serves organizations well as they attempt to assess and
improve performance.
Many readers will readily associate the activities listed—
plan, test, study, implement—with the scientific method. The Plan-Do-Study-Act (PDSA) Cycle
Indeed, the scientific method is a fundamental, inclusive A well-established process for improvement that is based
paradigm for change and includes the following six steps: on the scientific method is the Plan-Do-Study-Act (PDSA)
1. Determine what is known now (about a process, cycle. This method is also called the Plan-Do Check-Act
problem, topic of interest). (PDCA) cycle. A brief explanation of this process follows
2. Decide what needs to be learned, changed, or improved. (see Figure 6-4, below). This process is attributed to
3. Develop a hypothesis about the effect of the change. Walter Shewhart, a quality improvement pioneer with Bell
4. Test the hypothesis. Laboratories in the 1920s and 1930s, and is also widely
Figure 6-4. The PDSA Approach to Performance Improvement
The Plan-Do-Study-Act (PDSA) approach to performance improvement includes identifying design or redesign
opportunities, setting priorities for improvement, and implementing the improvement project.
109
associated with W. Edwards Deming, a student and later

Checklist 6-3.
a colleague of Shewhart. Deming made the PDCA cycle
central to his influential teachings about quality. The cycle
Components of the PDSA Cycle
is compelling in its logic and simplicity.
Plan
During the Plan step, an operational plan for testing the 33 Develop or design a new process, or redesign or
improve an existing process.
chosen improvement action is created. Small-scale testing
can help determine whether the improvement actions 33 Determine how to test the new or redesigned
process.
are viable, whether they will have the desired result, and
whether any refinements are necessary before putting them 33 Identify measures that can be used to assess the
success of the strategy and whether the objective
into full operation. The list of proposed improvement actions was reached.
should be narrowed to a number that can be reasonably
33 Determine how to collect data to assess success.
tested—perhaps two, three, or four, but not often more.
33 Involve the right people in development and testing.
During the planning stage, a number of issues should Do
be resolved: 33 Run the test of the new or redesigned process,
preferably on a small scale.
• Who will be involved in the test?
• What must these people know to participate in the test? 33 Collect data to assess success.
• What are the testing timetables?
Study
• How will the test be implemented? 33 Assess the results of the test.
• Why is the idea being tested?
33 Determine whether the change was successful.
• What are the success factors?
33 Identify any lessons learned.
• How will the process and outcomes of the test be
measured and assessed? Act
33 If the test was successful, implement the change
The Do step involves implementing the pilot test and permanently.
collecting actual performance data. 33 If the test was not successful, modify it and run it
through another testing cycle.
During the Study (or Check) step, data collected during 33 If the subsequent test was not successful, abandon
the change and develop a new approach to test.
the pilot test are analyzed to determine whether the
improvement action was successful in achieving the desired
outcomes. To determine the degree of success, actual test of different testing phases or different change strategies and
performance is compared to desired performance targets can therefore require the use of consecutive PDSA cycles.
and baseline results achieved using the established process.
To help teams and individuals involved in design or
The last step is the Act step—to take action. But if the pilot improvement initiatives apply the method effectively, the
test is not successful, the Plan-Do-Study cycle repeats. organization, depending on the nature of the improvement
When actions have been shown to be successful, they are project, may want to consider the questions outlined in
made part of standard operating procedure. The process Sidebar 6-7, page 111, at each step of the method.
does not stop here. The effectiveness of the action should
continue to be measured and assessed to ensure that DMAIC
improvement is maintained. A newer process improvement model that has evolved
from PDSA goes by the acronym DMAIC (pronounced
The components of the four-step PDSA cycle as they relate duh-MAY-ick). The letters stand for Define, Measure,
to designing and improving processes appear as Checklist Analyze, Improve, and Control, the five phases of the
6-3, above right. A single initiative can involve a number process. Used in root cause analysis, DMAIC acts like
110
Sidebar 6-7. Key Questions to Consider During the PDSA Cycle
Plan: • Was data collection timely?

• How was a design or improvement strategy selected • Was data collection reliable?
for testing?
Study:
• Is there knowledge-based information (for example,
• How will the test data be assessed?
from the literature, other organizations, or other
• What process should be used?
external sources) supporting the new or improved
• Who should be involved in data analysis?
process?
• What methods or tools should be used to analyze
• What issues in the external environment (such as
data?
economy, politics, customer needs, competitors,
• Is training needed on data analysis methods and
regulations) will affect the performance of the new or
tools?
improved process?
• Are comparative data (internal or external) available?
• What issues in the internal environment will affect the
• Does data analysis lead to an understanding of
performance of the new or improved process?
problem areas?
• Who is (are) the customer(s) of the process?
• Are data analyses timely? Are the results available
• What is the current process?
soon enough to take needed actions?
• What is the desired process?
• Did the test data indicate that the design or
• Who are the suppliers of the process?
improvement was successful?
• What changes will have the most impact?
• What lessons were learned from the test?
• Is there a plan for testing the design or improvement?
• What measures will determine whether to implement
• Is there a time line for testing?
the tested design or improvement on a permanent
• What data will be collected to determine whether the
basis?
test was successful (that is, whether the objective
• How and to whom will the results of assessment
was met)?
activities be communicated?
• How is it determined that the measures actually
address the desired issue? Act:
• Can the measures used actually track performance? • Should changes be recommended to others (for
• How will data be collected? example, for purchasing equipment or implementing
• Who will collect data? specific processes)?
• Are systems in place to support planned • How will these changes be communicated to the
measurement? appropriate people?
• Is benchmarking feasible for this initiative? • Is any education or training needed?
• Are the right people involved? • How will gain be maintained and backsliding be
• What resources are needed to design or redesign the prevented?
process? What resources are available? • What measures should be used to assess the
performance of the new or improved product or
Do:
process?
• Was the testing plan followed?
• Should any of the measures identified previously be
• Were needed modifications discussed with the
included in ongoing measurement activities?
appropriate people?
a funnel, filtering through the proximate causes to the most Define: What is the problem? A team is formed and a project
likely root cause(s), as shown in Figure 6-5, page 112. charter developed.
Each phase carries with it a key question that must be Measure: What is the extent of the problem? Baseline data are
clearly answered before proceeding to the next phase. The collected and key measures determined.
questions and the corresponding actions are the following:
Analyze: Under what circumstances did the problem occur?
Statistical analysis tools are used to analyze the data and the
process to determine the root cause(s).
111
Figure 6-5. DMAIC Model for Root Cause Analysis
The DMAIC model for performance improvement includes five phases—Define, Measure, Analyze, Improve, and Control.
Cause investigation occurs during the analyze phase, as the process filters through the proximate causes to reach the
most likely root cause. Then improvements can be made and sustained.
Improve: How can the problem be solved? Solutions to the • Brainstorming can be used to create ideas for
identified problems are developed, evaluated, refined, and improvement actions.
implemented. • Multivoting can help a team decide between possible
improvement actions.
Control: How can the solution be sustained? Improvements • Flowcharts can help a team understand the current
and design controls are documented and monitored. process and how the new or redesigned process should
work.
Additional Improvement Tools • Fishbone diagrams can indicate which changes might
The following tools are useful for taking action to improve cause the desired effect, that is, the desired result or goal.
processes:
112
• Pareto charts can help determine which changes are likely • Works in multiple forms
to have the greatest effect in reaching the goal. • Involves doing instead of telling
• Control charts and scatter diagrams can measure the • Explains inconsistencies
effect of a process change or variation in processes • Involves give and take
and outcomes.
• Histograms can show how much effect each change 5. Empower broad-based action by doing the following:
has had. • Communicating sensible vision to employees
• Making organization structures compatible with
action
Brainstorming, multivoting, • Providing needed training
flowchart, fishbone diagram, • Aligning information and human resource systems
Pareto chart, control chart, scatter • Confronting supervisors who undercut change
diagram, histogram
These and the other tools shown in Chapter 7 may be used 6. Generate short-term wins in the following ways:
individually or in some combination as part of a perfor- • Fixing the date of certain change
mance improvement strategy such as Lean Six Sigma. • Doing the easy stuff first
• Using measurement to confirm change
Creating and Managing the Change
Some suggested actions the team might take to help manage 7. Consolidate gains and produce more change by doing
and lead the change or improvement process follow.15 These the following:
actions are based on eight sequential stages in the process of • Identifying true interdependencies and smooth
leading change in organizations.16 The steps in creating and interconnections
managing the change process are as follows: • Eliminating unnecessary dependencies
• Identifying linked subsequent cycles of change
1. Establish a sense of urgency by doing the following:
• Identifying the best anywhere and the gap between 8. Anchor new approaches in the culture with the
one process and another following:
• Identifying the consequence of being less than • Results
the best • Conversation
• Exploring sources of complacency • Turnover
• Succession
2. Create a guiding coalition to do the following:
• Find the right people STEP 18 Develop Measures of
• Create trust Effectiveness and Ensure
• Share a common goal Their Success
When a function or process is under way, the team
3. Develop a vision and strategy that is the following: should collect data about its performance. As described in
• Easily pictured Chapter 4, measurement is the process of collecting and
• Attractive aggregating these data, a process that helps assess the level
• Feasible and clear of performance and determine whether further improve-
• Flexible ment actions are necessary. Specifically, measurement can be
• Communicable used as an integral technique throughout the PDSA cycle to
do the following:
4. Communicate the changed vision in a way that has the • Assist in process design or redesign (the Plan step)
following characteristics: • Test whether process design or redesign is implemented
• Is simple properly (the Do step)
• Uses metaphor • Assess the results of the test (the Study step)
113
• Provide assistance in implementing the improvement

(the Act step)
TIP
• Maintain the improvement and determine whether Make every effort to coordinate any ongoing measure
the improvement should be part of the organization’s ment with data collection already taking place as part
of the organization’s everyday activities. For example,
ongoing monitoring process (repeat of the PDSA cycle)
if nurses are collecting blood pressure data as part
of their normal duties, these same data should be
The first step in measuring the success of improvement used when assessing the impact of a specific drug on
efforts is to develop high-quality measures of effectiveness. patient blood pressure.
The choice of what to measure is critical. Measurement
must relate to the improvement and validate the accom-
plishment of the goal (or failure to reach the goal). See of a frequently ordered medication, the team may want to
Checklist 6-4, right, for key questions the team should seek the help of information management staff with access
ask concerning what to measure. Answer the questions in to statistical software capable of analyzing a large volume
Worksheet 6-10 at the end of this chapter, pages 129–130, of data.
as the team identifies, measures, and designs the measure-
ment plan. Information management professionals and those respon-
sible for carrying out the process being measured are key
Some measures or performance indicators may require players in data collection and analysis. The people involved
specific targets, which should be set by the team prior to vary widely depending on the specific organization, the
data collection. For example, in the patient suicide case function being measured, and the measurement process.
described in Sidebar 3-1, the team should set 100% as
the target for bringing rooms in the behavioral health care STEP 19 Evaluate Implementation of
unit into compliance with breakaway shower fixtures. For Improvement Efforts
the treatment delay example in the same sidebar, the team After data are collected as part of measurement, they
should set a score of 95% as the target for all posttraining must be translated into information that the team can
test scores. use to make judgments and draw conclusions about
Data collection efforts should be planned and coordinated. Checklist 6-4.

Use a separate worksheet to plan and monitor the indicators Ensuring the Success of Measurement
selected to measure each improvement goal (see Worksheet
6-11 at the end of the chapter, page 131). An affirmative answer to the following questions gives
the team a good indication that it is on the right track
Who should be responsible for measurement? The team, with its efforts to measure the effectiveness of improve-
empowered to study the process and recommend changes, ment initiatives.
is usually responsible for designing and carrying out the Yes No
measurement activities necessary to determine how the
process performs. After making changes to improve the 33  Is there a plan for use of the data?
process, the team should continue to apply some or all of 33  Are the data collected reliable and valid?
its measures to determine whether the change has had the 33  Has ease of data collection been assured?
desired effect. Organizations may have various experts who 33  Have key elements required for improvement
been defined?
can help design measurement activities, including experts
in information management, quality improvement, and 33  Has a data rich–information poor syndrome
been avoided?
the function to be measured. The team can request such
contribution on an ad hoc basis. For example, if the team 33  Has a key point for information dissemination
been designated?
is investigating a medication error and has a large amount
of data to codify and process regarding the administration
114
the performance of improvement efforts. This assess- Assessment is not confined to information gathered within
ment forms the basis for further actions taken with a single organization. To better understand its level of
improvement initiatives. performance, an organization should compare its perfor-
mance against reference databases, professional standards,
Numerous techniques can be used to assess the data and trade association guidelines.
collected. Most types of assessment require comparing data
to a point of reference. These reference points may include Desired Performance Targets
the following: The team may also establish targets, specifications, or
• Internal comparisons, such as unit-to-unit or time- thresholds for evaluation against which to compare current
to-time performance. Such levels can be derived from professional
• Aggregate external reference databases literature or expert opinion within the organization.
• Desired performance targets, specifications, or thresholds
STEP 20 Take Additional Action
Internal Comparisons The team’s assessment of the data collected indicates
The team can compare its current performance with its past whether established targets or goals are being achieved. If
performance using statistical quality control tools. Two such the goals are being achieved, the team’s efforts now should
tools are particularly helpful in comparing performance focus on communicating, standardizing, and introducing
with historical patterns and assessing variation and stability: the successful improvement initiatives. The team can do
control charts and histograms. These tools respectively show the following:
variation in performance and the stability of performance. • Communicate the results, as described in subsequent
Step 21.
Control chart, histogram • Revise processes and procedures so that the improvement
is realized in everyday work.
• Complete necessary training so that all staff members are
Aggregate External Reference Databases aware of the new process or procedure.
In addition to assessing the organization’s own historical • Establish a plan to monitor the improvement’s ongoing
patterns of performance, the team can compare the orga- effectiveness.
nization’s performance with that of other organizations. • Identify other areas where the improvement could be
Expanding the scope of comparison helps an organization implemented.
draw conclusions about its own performance and learn
about various methods to design and carry out processes. Organizations frequently falter when continued measure-
Aggregate external databases come in a variety of forms. ment indicates that improvement goals are not being
Aggregate, risk-adjusted data about specific indicators sustained. More often than not, efforts tend to provide
help each organization set priorities for improvement by short-term rather than long-term improvement. If the team
showing whether its current performance falls within the is not achieving the improvement goals, then it needs to
expected range. revisit the improvement actions by circling back to confirm
root causes, identify a risk reduction strategy, design an
One method of comparing performance is benchmarking. improvement, implement an action plan, and measure the
Although a benchmark can be any point of comparison, effectiveness of that plan over time.
most often it is a standard of excellence. Benchmarking is
the process by which one organization studies the exem- There are a number of reasons a team’s improvements may
plary performance of a similar process in another organiza- falter and fail.17 If the team is having trouble effecting
tion and, to the greatest extent possible, adapts that infor- improvement, consider the reasons and remedies shown in
mation for its own use. Or the team may wish to simply Sidebar 6-8, page 116.
compare its results with those of other organizations or with
current research or literature.
115
Sidebar 6-8. When Improvements Falter: Reasons and Remedies
Problems ►► Ensure continued training of all appropriate staff.

►► Failure to hold the gains because the improvement
►► Build a lasting improvement by recognizing that individ-
required major changes, particularly behavioral
uals adopt innovations after passing through a series
changes
of stages, including knowledge (becoming aware that
►► Failure to hold the gains because the improvement a new idea exists), persuasion (forming a favorable
created extra work or hassle attitude toward the new idea), decision (choosing to
adopt the innovation), implementation (putting the idea
►► Failure to hold the gains because new staff or lead-
into use), and confirmation (seeking further verification
ership personnel were not trained in the improved
of the innovation, leading to either continued adoption
process
or discontinuance). Consider and plan for this process
►► Inability to replicate in other settings when bringing an improvement to every setting.
►► Not enough public or personal attention to improve- ►► Leaders must ensure that improvement successes are
ment success recognized and celebrated.
►► Inadequate institutional and administrative support ►► Leaders must provide time and talent.
►► Hidden barriers to needed changes ►► Leaders must empower improvement teams to identify
where changes are needed and help them make the
►► A cookie-cutter approach to replicating improvement
changes happen.
Remedies ►► Process improvements must be reinvented at each

►► Big changes are best arrived at one step at a time. new site, adapted to meet local circumstances, and
fingerprinted by the local owners of the newly improved
►► Design improvements so that the desired task is the
process.
easiest thing do. Design a robust process that makes
it easy to do things right and difficult to do things Source: Splaine ME, et al, editors: Practice-Based Learning and Improve
incorrectly. ment: A Clinical Improvement Action Guide, 3rd ed. Oak Brook, IL: Joint
Commission Resources, 2012.
STEP 21 Communicate the Results presentation should include those whose approval and help
Throughout the root cause analysis process, the team should is needed as well as those who could gain from the team’s
be communicating to the organization’s leadership the team recommendations. Consider the following questions in
conclusions and recommendations as outlined by the team communicating an improvement initiative:
early in the process. Hence, the communication process • How will implementation of this initiative be
occurs throughout the team’s effort and is critical to the communicated throughout the organization?
success of improvement initiatives. • Who needs to know?
• What communication vehicles will the team use
After determining what happened or could have happened for various audiences (individuals both directly and
and identifying root causes of the event or possible event, indirectly affected by the improvement)?
the team should provide leadership with the recommenda-
tions for improvement actions to prevent a recurrence of The team also needs to ensure that communication about
the event. Generally, a short written report provides leaders the results will be an ongoing activity that reinforces the
with the summary they need. An outline of the contents of reasons for the improvement initiative. Storytelling is
such a report appears as Sidebar 6-9, page 119. an effective strategy when reporting the results to a wide
audience at various staff meetings. Storytelling humanizes
The team should consider with care how and to whom the the event that caused the root cause analysis, which helps
report is to be presented. Participants during a formal oral catch people’s interest and get them emotionally invested in
116
Sidebar 6-9. the improvement actions being implemented. Storytelling

also helps safeguard an organization against communica-
Possible Content of Report to Leaders
tion leaks. Written reports have a greater tendency than
Event Description
oral accounts do to leak outside a facility to media outlets,
which in some cases leads to organizations being publicly
This section includes a brief description of the sentinel portrayed in the worst possible light when incident reports
event or possible event. It includes what, when, where,
and improvement plans are taken out of context.
who, and how information as articulated in the problem
definition (see Step 2 in Chapter 3, pages 49-53). The
emphasis is on facts related to the event and the areas Following implementation of such actions and measuring
involved. and ensuring their success, the team should report to
leadership on the results of the improvement actions. The
Scope of Analysis report should include information regarding applicability to
This section describes the team’s membership and other processes, areas, and locations and the lessons learned.
purpose as well as the analytical methods used to
investigate the event or possible event.
Proximate Causes and Immediate Responses

This section describes the circumstances leading to the
event, proximate causes identified by the team, and
any response strategies and corrective actions imple-
mented by individuals immediately following the event.
Contributing Factors
This section describes the circumstances, actions, or
influences thought to have played a part in the origin,
development, or increased risk of an incident.
Root Causes
This section describes the analyses conducted to deter-
mine root causes and lists the root causes identified by
the team.
Improvement Actions and Follow-Up Plan

This section describes the improvement actions
recommended by the team for each root cause. It also
describes the measures and time frame recommended
to evaluate the effectiveness of improvement actions.
117

References
1. Leape LL. Error in medicine. JAMA. 1994 Dec 21;272(23): 10. Wilson PF, Dell LD, Anderson GF. Root Cause Analysis: A Tool for
1851–1857. Total Quality Management. Milwaukee: ASQ Quality Press, 1993.
2. Last JM, editor. A Dictionary of Epidemiology, 4th ed. New York: 11. Veterans Health Administration. Systems Improvement Framework.
Oxford University Press, 2001. Jan 2010. Accessed Oct 20, 2015. http://www.paloalto.va.gov
3. Nash DB, Goldfarb NI. The Quality Solution: The Stakeholder’s /docs/improvementguide.pdf.
Guide to Improving Health Care. Sudbury, MA: Jones and 12. Ammerman M. The Root Cause Analysis Handbook: A Simplified
Bartlett, 2006. Approach to Identifying, Correcting, and Reporting Workplace
4. Joint Commission on Accreditation of Healthcare Organizations: Errors. New York: Productivity Press, 1998.
Failure Mode and Effects Analysis in Health Care: Proactive Risk 13. Institute of Medicine. Crossing the Quality Chasm: A New Health
Reduction, 2nd ed. Oakbrook Terrace, IL: Joint Commission System for the 21st Century. Washington, DC: National Academy
Resources, 2005. Press, 2001.
5. Mendes ME, et al. Practical aspects of the use of FMEA tool in 14. Schulte SK. Avoiding culture shock: Using behavior change
clinical laboratory risk management. J Bras Patol Med Lab. 2014 theory to implement quality improvement programs. J AHIMA.
Jun;49(3):174–181. 2007 Apr;78(4):52–56.
6. Hoffman C, et al. Canadian Root Cause Analysis Framework: 15. Nelson EC, Batalden PB, Ryer JC, editors. Clinical Improvement
A Tool for Identifying and Addressing the Root Causes of Critical Action Guide. Oakbrook Terrace, IL: Joint Commission on
Incidents in Healthcare. Edmonton, AB: Canadian Patient Safety Accreditation of Healthcare Organizations, 1998, pp. 116–117.
Institute, 2006. 16. Kotter JP. Leading Change. Boston: Harvard Business School
7. Grout J. Mistake-Proofing the Design of Health Care Processes. Press, 1996.
AHRQ Publication No. 07-0020. Rockville, MD: Agency for 17. James B, Ryer JC. Holding the gains. In Nelson EC, Batalden
Healthcare Research and Quality, 2007. PB, Ryer JC, editors: Clinical Improvement Action Guide.
8. Keohane C, et al. Intravenous medication safety and smart Oakbrook Terrace, IL: Joint Commission on Accreditation of
infusion systems: Lessons learned and future opportunities. Healthcare Organizations, 1998, 121–124.
J Infus Nurs. 2005 Sep–Oct;28(5):321–328.
9. Hughes D. Root cause analysis: Bridging the gap between ideas
and execution. Topics in Patient Safety. 2006 Nov–Dec;6(5):1, 4.
Accessed Oct 20, 2015. http://www.patientsafety.va.gov/docs
/TIPS/TIPS_NovDec06.pdf.
118
Worksheet 6-1. Prioritizing Improvement Actions
Use the following worksheet to catalog improvement actions suggested by the team. Separate sheets for each root
cause and its suggested improvement actions may be used. The team should also rate or rank improvement actions
based on agreed-upon criteria. Use this worksheet to record the rankings of individual team members and the team
as a whole.
Root Causes Suggested Improvement Actions Ranking
Root Cause 1:
Root Cause 2:
Root Cause 3:
119
Worksheet 6-2. Summarizing the Potential of Improvement Actions
Two questions help the team summarize the potential of each proposed improvement action. Use the space following
each question to provide a concise answer.
What will result from implementing this action?
What would result from not implementing this action?
120
Worksheet 6-3. Defining Improvement Goals, Scope, and Activities
This worksheet helps the team define what it is trying to improve. Use the space following each question to provide
as concise an answer as possible.
What goals does the organization have in implementing necessary improvements related to a sentinel event or
possible event?
What measures will be most affected by the change?

What specific activities must be carried out to reach the goals and affect the dimensions of performance? (Provide
a clear statement of the essential features of each proposed solution.) What are the sequential steps necessary to
accomplish the proposed improvement?
1.
2.
3.
4.
5.
6.
121
Worksheet 6-4. Defining Improvement Expectations, Sequence, Resources, and Measures
This worksheet helps the team define how the organization will meet its improvement goals. Use the space following
each question to provide as concise an answer as possible.
How must the improved process perform?

What sequence of activities will be required to meet these expectations?

What resources will be required to meet these expectations?

How and what will be measured to determine whether the process is actually performing at the level expected?
Improvement Action Quantitative Measure
122
Worksheet 6-5. Defining Time Frames and Milestones
This worksheet helps the team define when the organization will meet its improvement goals. Use the space
following each question to provide as concise an answer as possible.
What time frame will be established for implementing the overall improvement action?

What time line will be established for each activity comprising the steps along the way?
Activity Time Frame

What are the major milestones and their respective completion dates?
Milestone Completion Date
123
Worksheet 6-6. Involving the Right People
Involving the right people at each stage of the improvement process is critical to the success of the improvement
initiative. Consider which individuals should be involved at each stage and write their names in the appropriate spaces.
Designing the action. In general, the group that participated in the root cause analysis should have the necessary
expertise to recommend improvements and may be in the best position to design or redesign the improvements.
This group should include those who carry out or are affected by the process. They are

Approving recommended actions. When substantial resources are involved and the potential effects are signif-
icant, the organization’s leaders usually have to approve the action. This is most certainly the case with improve-
ments recommended following a sentinel event. If a group has obtained the necessary input and buy-in while
devising an improvement, the approval should come readily. The appropriate leaders are
(continued)
124
Worksheet 6-6. Involving the Right People (continued)
Testing the action. Testing should occur under real-world conditions, involving staff who will actually be carrying out
the process. Effects can be measured with the same methods used to establish a performance baseline. Appropriate
staff members include

Implementing the action. Although full-scale implementation of a process change should have positive results, any
change can create anxiety. Therefore, care should be taken to prepare people for change and to explain the reason
for the change in an educational, nonthreatening way. Cooperation is essential for changes to succeed, but it cannot
exist if people believe a change is being forced on them without good reason. An effective team should have already
acquired much of the necessary buy-in during earlier phases of the improvement process or during the early stages
in the root cause analysis. Appropriate staff members include
125
Worksheet 6-7. Determining the Location of Improvement Actions
This worksheet helps the team define where the organization will implement improvement goals.
Where will the improvement action be implemented?

Will its implementation be organizationwide or in a selected location with a selected patient population or selected
staff members?

Are the location, target population, and target staff of the improvement action likely to expand with success?
 No (If no, why not?)  Yes (If yes, how?)
126
Worksheet 6-8. Integrating the Improvement Plan
Define the time lines and responsibilities associated with each of the project steps using the following table.
(Customize column headers as desired.) Questions to consider include the following:
• What are the time lines for each step of the project and for the project as a whole?
• What will be the checkpoints, control points, or milestones for project assessment?
• Who is responsible for each step or milestone?
• Who is responsible for corrective course action?
• What staff members will be involved in the improvement project?
• What will be the nature and extent of their responsibilities?
Steps to Date of Areas for Individuals
Be Taken Implementation Implementation Responsible Other Considerations
127
Worksheet 6-9. Identifying Change Barriers and Solutions
Use this worksheet to identify possible barriers to change and solutions to overcome such barriers.
Areas where resistance to change might emerge include

Countermeasures to overcome such barriers include
128
Worksheet 6-10. Designing the Measurement Plan
What is the scope of measurement for the improvement project?

Have any portions of the process under study been measured in the past, or are they currently being measured?
If so, are assessments available?

What measurement tools will be used for this initiative?

Will the tools provide reliable data? Have they been tested?

What costs are associated with collecting the necessary data? Do benefits outweigh costs?

Can the data generated by the selected measurement tool be transformed into meaningful and useful information?

How can the team ensure that the data are complete, accurate, and unbiased?
(continued)
129
Worksheet 6-10. Designing the Measurement Plan (continued)
How will the staff collecting data be educated?

What format(s) will be used to report the data?

Where and how will any additional required data be obtained?

How will the success of the improvement be measured?
130
Worksheet 6-11. Evaluating Target Goals
Use a worksheet like this one to plan and monitor progress in measuring the effectiveness of each improvement goal.
Goal Measure Person Responsible Review Completed
131
132
CHAPTER
7
Tools and Techniques
Learning Objectives
• Become familiar with a variety of analytical tools and processes, when each is best used, and how best to
apply them
• Implement these tools as part of a comprehensive systematic analysis
Root cause analysis is a key element in an organization’s Six Sigma is an appropriate label because the Lean and
efforts toward performance improvement. The Joint Six Sigma methods are found in both companies and are
Commission has created its own organizationwide plan of considered inseparable.
Robust Process Improvement™ based on Lean Six Sigma
methodology and geared toward the specific issues of DMAIC
patient safety and quality in health care. Lean Six Sigma often is applied in a five-step problem-
solving approach commonly known by the acronym
Within this framework or other, similar methodologies, DMAIC:
numerous tools exist to facilitate the process of root cause • D, for Define, is a critical first step to have a clear
analysis. understanding of the issue or problem to be solved.
• M stands for Measure, wherein the team assesses the
What Is Lean Six Sigma? baseline performance of the process. This step also is
Lean Six Sigma is a term describing an improvement essential because without knowing the current state, one
methodology based on the scientific method, statistical cannot gauge whether the process improved.
methods, customer values, flow and pull principles, PDSA • A is for Analyze, which is what one must do for all the
(Plan-Do-Study-Act), and change management, among critical contributing factors—that is, the root causes.
other methods. In his book The Machine That Changed the Lean Six Sigma values the statistical validation of these
World, James Womack coined the term Lean to describe root causes to better ensure that improvements are
the methods of the Toyota Production System, in which all achieved.
employees work together to eliminate waste (of resources • I stands for Improve, and the team will statistically
and time) and to keep the process focused on the value of validate that the interventions achieve the results.
the product to the customer. The term Six Sigma comes • C is for Control to ensure sustained gains.
from Motorola’s attempt to describe its methods for
improving quality in a systematic way. It refers to a statis- (DMAIC is described in greater detail in Chapter 6, pages
tical concept in which a process or output remains within 112–114.) Earlier problem-solving methodologies often
six standard deviations of the mean value in a normal have shown limited success in creating interventions that
distribution. Both Toyota and Motorola have been widely last. The methodologies forwarded by Toyota and Motorola
recognized for high quality. For those who have worked thus include key tools to better control the process for long-
with both companies and have applied their methods, Lean term success.
133
Any employee can learn Lean Six Sigma. One’s level of

Sidebar 7-1. Ishikawa’s Seven Tools
expertise is often signified by a “belt” designation: Green
Belt, Black Belt, Master Black Belt, or Sensei, awarded According to performance improvement pioneer Kaoru
following training at that level. Ishikawa, if used skillfully, the following seven basic
quality control tools—all of which are visual, in the form
Green Belts are taught statistical methods such as perfor- of charts, graphs, or diagrams—will enable 95% of
mance improvement pioneer Kaoru Ishikawa’s seven tools workplace problems to be solved.
(see Sidebar 7-1, right) as well as the flow and pull concepts 1. Check sheet
from the Toyota Production System. Flow (continuous 2. Control chart
forward movement) in a process is desired as a sign that the 3. Fishbone diagram (also called cause-and-effect
process quality and productivity are higher than a process diagram or Ishikawa diagram)
that repeatedly stops and starts. Pull refers to how the 4. Flowchart (also called stratification)
inputs of the process are replenished based on the custom- 5. Histogram (also called frequency distribution)
er’s demand rather than the organization’s timetable. 6. Pareto chart
7. Scatter diagram
This chapter provides information on tools and techniques Each of these tools is described in a Tool Profile in this
that can be used during root cause analysis. The tools and chapter.
techniques are presented in a uniform format, called a Tool
Profile, to assist readers with their selection and use. Each Source: Ishikawa K. Introduction to Quality Control. Trans. J. Loftus.
Tokyo: 3A Corp., 1989.
Tool Profile identifies the stage of root cause analysis during
which the tool or technique may be used, its purpose, basic
usage steps, and tips for effective use. An example of the
tool or technique follows each profile.
When embarking on a root cause analysis, team members

may wish to start by consulting the tool matrix shown
in Table 7-1 on page 135. This matrix lists many of the
tools and techniques available during root cause analysis
and indicates the stages during which they may be
particularly helpful.
134
CHAPTER 7 | Tools and Techniques
Table 7-1. Tool Matrix

Identifying Implementing
Tool Name Identifying Identifying Root Improvement and Monitoring
(Alternate Name) Proximate Causes Causes Opportunities Improvements
Affinity diagram X X
Brainstorming X
Capability chart X X
Change analysis X X
Change management X X
Check sheet X
Control chart X
Failure mode and X
effects analysis
(FMEA)
Fishbone diagram X X
(Cause-and-effect
diagram, Ishikawa
diagram)
Flowchart X X X X
(Stratification)
Gantt chart X X X X
Histogram X X X X
Kaizen X
Multivoting X X X
Operational definition X
Pareto chart X
Relations diagram X X X
Run chart X X X
Scatter diagram X
(Scattergram)
SIPOC process map X
Stakeholder analysis X
Standard work X
Value stream mapping X X
This matrix lists many of the tools and techniques available during root cause analysis and indicates the stages
during which they may be particularly helpful. Each tool is described in greater detail in this chapter.
135
TOOL
Affinity Diagram TIPS FOR EFFECTIVE USE
PR
(See Figure 7-1) Overcome interviewee defensiveness by doing the
OFIL
E
following:
■■ Keep the team small (four to six people) and
Stages to Use: Identifying proximate causes, identifying
ensure varied perspectives.
improvement opportunities
■■ Generate as many ideas as possible using
brainstorming guidelines.
Purpose: To creatively generate a large volume of ideas or
■■ Record ideas from brainstorming on sticky notes.
issues and then organize them into meaningful groups.
■■ Sort the ideas in silence, being guided in sorting
only by gut instinct.
Simple Steps to Success:
■■ If an idea keeps getting moved back and forth
1. Choose a team.
from one group to another, agree to create a
2. Define the issue in the broadest and most neutral manner.
duplicate note.
3. Brainstorm (for example, for contributing factors or for
■■ Reach a consensus on how notes are sorted.
suggested improvements) and record the ideas.
■■ Allow some ideas to stand alone.
4. Randomly display sticky notes with the ideas so that
everyone can see them. ■■ Make sure that each idea has at least a noun and
5. Sort the ideas into groups of related topics. a verb when appropriate; avoid using single words.
6. Create header or title cards for each grouping. ■■ Break large groupings into subgroups with
7. Draw the diagram, connecting all header cards with subtitles, but be careful not to slow progress with
too much definition.
their groupings.
136
Figure 7-1. Example: Affinity Diagram
Laboratory reports not

on chart at 8 A.M.
Lab routine Lab results Lab ordering
Labs designated as daily not drawn Lab reports not printed until 10:30 AM Intensive care unit and stat orders
until 10 AM take precedence
Intensive care unit and telemetry Results available in lab system Too many stat orders when
drawn first before until system physicans discover no labs
Time to draw certain labs Lab results not printed by secretary Increased cost of stat labs
No way to tell if patient is

Labs not placed on front of chart Labs ordered incorrectly
preoperative
Main secretary ill Too many types or times to

choose from
This affinity diagram shows how a wide range of ideas can be arranged in manageable order. Using this type of diagram
presents ideas on why laboratory results are not available as needed in three categories: routine, results, and ordering.
137
TOOL
Brainstorming TIPS FOR EFFECTIVE USE
PR
(See Figure 7-2) ■■ Create a nonthreatening, safe environment for
OFIL
E
expressing ideas.
Stages to Use: Identifying improvement opportunities ■■ Tell the group up front that any idea is welcome,
no matter how narrow or broad in scope or how
serious or light in nature. All ideas are valuable, as
Purpose: To generate multiple ideas in a minimum amount
long as they address the subject at hand.
of time through a creative group process
■■ Remember that the best ideas are sometimes the
most unusual.
■■ Allow group members to occasionally say “pass”
1. Define the subject. Doing so ensures that the session has
if they can’t think of an idea when it’s their turn to
direction. speak.
2. Think briefly about the issue. Allow enough time for team
■■ Never criticize ideas. It is crucial that neither the
members to gather their thoughts, but not enough time leader nor the other group members comment on
for detailed analysis. any given idea.
3. Set a time limit. Allow enough time for every member ■■ In thinking briefly about the issue (Step 2), do
to make a contribution, but keep it short to prevent not give group members time to second-guess
premature analysis of ideas. their ideas. Be aware that self-censorship stifles
4. Generate ideas. Use a structured format in which the creative thought.
group members express ideas by taking turns in a ■■ Write down all ideas on a chalkboard or easel so
predetermined order, one idea per turn. The process that the group can view them.
continues in rotation until either time runs out or ideas ■■ Keep it short; enforce a time limit of 10 to 20
are exhausted. minutes.
5. Clarify ideas. The goal is to make sure that all ideas are ■■ In organizations where staff may not regularly be
recorded accurately and are understood by the group. in a centralized location, brainstorming can be
done by asking staff to submit as many ideas as
possible about the topic in writing, by voice mail, or
by e-mail.
■■ Make a note of deeper root causes that emerge
during brainstorming and place them in a “parking
lot” list for consideration later.
138
Figure 7-2. Example: Brainstorming List
Possible factors contributing to a surgical error include the following:

►► No timely case review
►► No mechanism to ensure patient identity
►► Informal case referral process
►► Untimely operative dictation
►► Inadequate presurgical evaluation
►► No review of patient care information prior to surgery
►► Inadequate informed consent
►► Patient care information unavailable for preoperative review
►► Failure to perform surgery in a safe manner
►► Laterality not clearly identified
►► Delay in reporting of incident
►► No multidisciplinary review
►► Ignored pathology reports
►► History of inadequate documentation in medical record
►► Procedures performed without adequate expertise
►► Failure to take responsibility for actions
►► No surgical plan/preoperative findings
This figure shows an excerpt from a list one organization created using brainstorming to identify possible contributing
factors in a surgical error that occurred. This list was used to create the fishbone diagram that appears as Figure 7-8,
page 148. As the example shows, the ideas are widely varied, and some seem more viable than others. Remember
that brainstorming is for generating ideas, not sorting or judging them.
139
TOOL
Capability Chart TIPS FOR EFFECTIVE USE
PR
(See Figure 7-3) ■■ Divide the process into steps the performance of which
OFIL
E
can be quantified.
Stages to Use: Identifying root causes; implementing and ■■ Determine upper and lower parameters for the
monitoring improvements. performance of each component of the process.

■■ For new processes, begin with a pilot program, which will
Purpose: To analyze processes in order to determine whether provide initial data on performance.
they are capable of satisfying the given requirements

1. Determine the upper and lower parameters for acceptable
performance within the process.
2. Collect or project data concerning actual or projected
performance.
3. Plot data on a graph showing the upper and lower
performance parameters.
4. Analyze graph to determine whether any part of the
performance curve falls outside of the parameters.
5. Decide whether performance values that fall outside the
parameters can be addressed by modifying the process, or the
whole process needs to be redesigned.
Figure 7-3. Capability Chart
Specification Limits (= Customer Requirements)
Source: George ML, et al. The Lean Six Sigma Pocket Toolbook: A Quick Reference Guide to Nearly 100 Tools for Improving Process Quality, Speed, and
Complexity. New York: McGraw-Hill, 2005. Used with permission.
140
TOOL
Change Analysis TIPS FOR EFFECTIVE USE
PR
(See Figure 7-4) ■■ Describe the problem as accurately and in as much
OFIL
E
detail as possible. Include in the description who

was involved, what might have been a factor in
Stages to Use: Identifying proximate causes; identifying
causing the event, where the event took place, and
root causes when it took place.
■■ After a change analysis is performed, additional
Purpose: To determine the proximate and root cause(s) of questions must be asked to determine how the
an event by examining the effects of change. This involves changes were allowed to happen.
identifying all changes, either perceived or observed, and all ■■ Continue the questioning process into the
possible factors related to the changes. organization’s systems.
■■ Remember that not all changes create problems;
Simple Steps to Success: rather, change can be viewed as a force that can
1. Identify the problem, situation, or sentinel event. either positively or negatively affect the way a
2. Describe an event-free or no-problem situation. Try to system, process, or individual functions.
describe the situation without problems in as much
detail as possible. Include the who, what, where, when,
and how information listed in Step 1. these affected the event. Did each difference or change
3. Compare the two. Take a close look at the event and explain the result?
nonevent descriptions, and try to detect how these 6. Integrate information and specify root cause(s). Identify the
situations differ. cause that, if eliminated, would have led to a nonevent
4. List all the differences. situation.
5. Analyze the differences. Carefully assess the differences
and identify possible underlying causes. Describe how
Figure 7-4. Change Analysis Worksheet
This generic worksheet shows a simple way of listing and comparing information for change analysis. The worksheet
is arranged in columns to show logically what happened, what did not happen, the difference between them, and an
analysis.
141
TOOL
Change Management TIPS FOR EFFECTIVE USE
PR
(See Figure 7-5) ■■ Calculate the costs of the change, both actual
OFIL
E
costs and opportunity costs, and be sure that the

need for the change outweighs these costs before
Stages to Use: Identifying improvements; implementing
moving forward.
and monitoring improvements
■■ Understand that resistance to change is a natural
human reaction, and assure key stakeholders in
Purpose: To encourage organizational acceptance of a the process that their voices will be heard.
change in process within a system
■■ Consider the change’s effect on overall
organizational culture, not just one process.
1. Determine the need for the change.
2. Secure leadership buy-in.
3. Anticipate resistance to change from staff who will be
affected by the change.
4. Communicate a vision of the improvement that will result
from the change.
5. Seek input from staff members in designing the change.
6. Implement the change.
7. Gauge the effectiveness of the change.
8. Modify the change, if necessary, to reach the desired
result.
9. Maintain support for the change.
Figure 7-5. Change Management
142
TOOL
Check Sheet TIPS FOR EFFECTIVE USE
PR
(See Figure 7-6) ■■ Know which data matter for your purpose.
OFIL
E
■■ Use the data recorded on the check sheet to

Stages to Use: Identifying proximate causes create a Pareto diagram or flowchart
Purpose: To sort and group data, using check marks

or similar symbols, onto a form that can be used with
other tools

1. Determine what data are to be collected.
2. Design a form on a blank sheet of paper with labeled
rows and columns representing the appropriate values.
3. Have team members record the data by placing a check
mark, hatch mark, X, or other symbol next to the value
to which each measurement corresponds.
143
Figure 7-6. Example: Check Sheet
144
TOOL
Control Chart TIPS FOR EFFECTIVE USE
PR
(See Figure 7-7) ■■ Obtain data before making any adjustments to the
OFIL
E
process.
Stages to Use: Implementing and monitoring ■■ In plotting data points, keep the data in the same
improvements sequence in which they were collected.

■■ Be aware that special causes of variation
Purpose: To identify the type of variation in a process and must be eliminated before the process can be
fundamentally improved and before the control
whether the process is statistically in control
chart can be used as a monitoring tool.
■■ The terms in control and out of control do not
signify whether a process meets the desired level
1. Choose a process to evaluate, and obtain a data set. of performance. A process may be in control
2. Calculate the average. but consistently poor in terms of quality, and the
3. Calculate the standard deviation. The standard deviation reverse may be true.
is a measure of the data set’s variability. ■■ Charting something accomplishes nothing; it
4. Set upper and lower control limits. Control limits should must be followed by investigation and appropriate
be three times higher or lower than the standard action.
deviation relative to the mean. ■■ Processes as a rule are not static. Any change
5. Create the control chart. In creating the control chart, can alter the process distribution and should trigger
plot the mean (that is, center line) and the upper and recalculation of control limits when the process
lower control limits. change is permanently maintained and sustained
6. Plot the data points for each point in time, and connect (that is, greater than 8 to 12 points on one side of
the process mean, or center line).
them with a line.
■■ Some special causes of variation are planned
7. Analyze the chart and investigate findings.
changes to improve the process. If the special
cause is moving in the right direction toward
improvement, retain the plan. It is working.
■■ The following four rules are used to identify out-of-
control processes:
1. One point on the chart is beyond three standard
deviations of the mean.
2. Two of three consecutive data points are on
the same side of the mean and are beyond two
standard deviations of the mean.
3. Four of five consecutive data points are on the
same side of the mean and are beyond one
standard deviation of the mean.
4. Eight data points are on one side of the mean.
145
Figure 7-7. Example: Patterns in Control Charts
These two control charts illustrate different patterns of performance an organization is likely to encounter. When
performance is said to be in control (top chart), it does not mean desirable; rather, it means a process is stable, not
affected by special causes of variation (such as equipment failure). A process should be in control before it can be
systematically improved. When one point jumps outside a control limit, it is said to be an outlier (bottom chart). Staff
should determine whether this single occurrence is likely to recur.
146
TOOL Failure Mode and Effects Analysis

(FMEA) TIPS FOR EFFECTIVE USE
PR (Synonym: failure mode, effects, ■■ Risk priority numbers may be calculated as the
OFIL
E
and criticality analysis, or FMECA) product of ratings on frequency of occurrence,

severity, and likelihood of detection.
(See Figure 7-8)
■■ Remember that this type of analysis is generally
proactive (used before an adverse event occurs),
Stage to Use: Identifying opportunities for improvement
although use during root cause analyses to
formulate and evaluate improvement actions is
Purpose: To examine a prospective design for possible ways also recommended and described in this book in
in which failure can occur so that actions can be taken to Chapter 6 on pages 99–100.
eliminate the possibility of failure, stop a failure before it
reaches people, or minimize the consequences of a failure

1. Select a high-risk process and assemble a team.
2. Diagram the process.
3. Brainstorm potential failure modes and determine their
effects.
4. Prioritize failure modes (often accomplished through
calculating a risk priority number).
5. Find root causes of failure modes.
6. Redesign the process.
7. Analyze and test the new process.
8. Implement and monitor the redesigned process.
147
Figure 7-8. FMEA/HFMEA® Steps

In conducting a health care FMEA (HFMEA®) the basic steps are:
• Identify the process to be examined.

• Assign FMEA team members, team leader and team facilitator.
• Explain the methodology to the team.
• Develop either a flowchart or detailed process flow (outline format) of the process under analysis. All steps in the
process should be included.
• Designate which of the steps in the process constitute “Functions.”
• Determine which Functions represent potential “Failure Modes” or points of potential failure. Determine the worst
potential adverse consequence or “Effect” of each of the Failure Modes.
• Determine the “Contributory Factors” for each Failure Mode. One or more Root Cause Analyses may be necessary
to complete this step. Note that we advocate the use of the term “Contributory Factor” rather than “Cause.”
• Identify any “Controls” in the process. Controls are components of the process which (a) reduce the likelihood of a
Contributory Factor or a Failure Mode, (b) reduce the severity of an Effect, or (c) detect the occurrence of a Failure
Mode or Contributory Factor before it leads to the adverse outcome (Effect).
• Rate the Severity of each Effect (usually on a scale of 1–10, with 10 being the most severe). The impact of Controls
that ameliorate the severity of an Effect are reflected in this rating as well.
• Rate the Occurrence (likelihood) of each Contributory Factor (usually on a scale of 1–10, with 10 being the most
frequent, or “certain to occur”). The impact of Controls that reduce the likelihood of occurrence of a Failure Mode
or Contributory Factor are reflected in this rating as well.
• Rate the effectiveness of each “Detection Control” (usually on a scale of 1–10, with 10 being the lack of a Detection
Control, or the presence of a wholly ineffective one, and 1 being a 100% flawless detection system).
• Multiply the three ratings by one another for each Contributory Factor and the corresponding Effect and Detection
Controls. The range of these products will be from 1 to 1,000. The resultant number is the Risk Priority number
(RPN) for that Contributory Factor.
• Rank-order the Contributory Factors according to the Risk Priority Numbers.
• Use a Pareto Chart with the traditional 80% rule to determine which contributory Factors should be addressed first.
• Add to the above listing ALL Contributory Factors which result in an Effect with Severity of 10, irrelevant of RPN.
• Develop a plan addressing how the selected Contributory Factors will be addressed, by whom, when, how the
improvement will be assessed, etc.
• Continue the improvement process.
Source: CCD Health Systems. The Basic Steps of Conducting a Healthcare FMEA. Accessed Nov 17, 2015. http://www.ccdsystems.com/Products
/FailureModeAnalystHFMEA/StepsinConductingaHealthcareFMEA.aspx. The HFMEA® process is reprinted with permission from CCD Health Systems
(www.ccdsystems.com). HFMEA is a registered trademark of CCD Health Systems.
148
TOOL Fishbone Diagram

(Synonyms: cause-and-effect diagram,
TIPS FOR EFFECTIVE USE
PR Ishikawa diagram) ■■ Make sure everyone agrees on the problem
OFIL
E
(See Figure 7-9) statement or outcome.

■■ Be succinct and stay within the team’s realm of
Stages to Use: Identifying proximate causes; identifying control.

root causes ■■ Place the outcome on the right side of the page,
halfway down, and then, from the left, draw an
arrow horizontally across the page, pointing to
Purpose: To present a clear picture of the many causal rela-
the outcome.
tionships between outcomes and the contributing factors in
■■ Represent common categories, including work
those outcomes
methods, personnel, materials, and equipment, on
the diagram by connecting them with diagonal lines
Simple Steps to Success: branching off from the main horizontal line.
1. Identify the outcome or problem statement. ■■ Brainstorm to come up with the important
2. Determine general categories for the causes. proximate causes. Place each proximate cause
3. List proximate causes under each general category. on a horizontal line connected to the appropriate
4. List underlying causes related to each proximate cause. diagonal line.
5. Evaluate the diagram. ■■ Gather data to determine the relative frequencies
of the causes.
■■ Look for causes that appear continually in the
evaluation process.
■■ Keep asking “Why?” to reach the root cause.
■■ Focus on system causes, not on causes
associated with individual performance.
149
Figure 7-9. Example: Fishbone Diagram
PATIENT PERSONNEL
Medical condition Lack of Education

Denial of suicide ideation
• History alcohol abuse • Depression assessment education
• No history of prior inadequate
psychiatric hospitalization • DTs (delirum tremens) Inadequate Staff
• Orientation procedures (content,
• “No Suicide” contract • Recent alcohol timeliness)
signed consumption (24 hr) • Cost
• Availability • 15-minute checks

• Psychosis trigger
Admit Status • Unassertive supervision
• Attrition
• EDO (emergency detention • Unavailability of staff
order)
• Danger to self and others
• No routine medications
given by law and state statute Suicide Death
in Mental
Health Unit
Safety Features
Initial Intervention
• Hazard present • “No Suicide” contract Admission Procedure
• Grab bar 33” from floor not • Psychiatric evaluation • Communication with case manager
previously determined to not performed not thorough
Hazard Rounds
be hazard • Written/verbal report incomplete
• By persons unfamiliar with • Inadequate orientation with community
• Direct observation of patient psychiatric needs mental health state
• No visualization of
• Assignment of personnel
bed areas Observation Status Inadequate
• Lack of education • Patient denied
• Physical layout
• Patient not seen by psychiatrist upon admit
• Lack of criteria Multiple Admissions
• Not an open dorm • Personnel not following procedure
• Access not • Time of day
restricted • Only provider
• Inadequate assessment by staff
ENVIRONMENT WORK METHODS
This figure illustrates how the generic diagram can be adapted to specific needs. This detailed diagram breaks down
the contributing factors that led to a sentinel event—the suicide of a patient in a mental health unit. By analyzing the
proximate and underlying causes listed, staff members can identify and prioritize areas for improvement.
150
TOOL
Flowchart TIPS FOR EFFECTIVE USE
PR
(See Figure 7-10) ■■ Ensure that the flowchart is constructed by the
OFIL
E
individuals actually performing the work being

charted.
■■ Be sure to examine a process within a system,
root causes; identifying opportunities for improvement;
rather than the system itself.
implementing and monitoring improvements
■■ If the process seems daunting and confusing,
create a simple high-level flowchart containing
Purpose: To help teams understand all steps in a process
only the most basic components. Do not include
through the use of common, easily recognizable symbols; too much detail; be wary of obscuring the basic
this illustrates the actual path a process takes or the ideal process with too many minor components.
path it should follow. ■■ Use adhesive notes placed on a wall to experiment
with sequence until the appropriate one is
Simple Steps to Success: determined.
1. Define the process to be charted, and establish starting and ■■ Make the chart the basis for designing an improved
ending points of the process. process, using spots where the process works well
2. Brainstorm activities and decision points in the process. as models for improvement.
Look for specific activities and decisions necessary to ■■ Create a separate flowchart that represents the
keep the process moving to its conclusion. ideal path of the process, and then compare the
3. Determine the sequence of activities and decision points. two charts for discrepancies.
4. Use the information to create the flowchart. Place each ■■ Keep in mind that difficulties probably reflect
activity in a box, and place each decision point in confusion in the process being charted, and work
a diamond. Connect these with lines and arrows to through them.
indicate the flow of the process.
5. Analyze the flowchart. Look for unnecessary steps,
redundancies, black holes, barriers, and any other
difficulties.
151
Figure 7-10. Example: Medication Administration Flowchart
Prescriber
writes order
Is order
Order is entered Pharmacy receives understandable No
Call perscriber to Prescriber clarifies
into computer computerized and compatible verify order or changes order
system order with patient’s
record?
Yes
Pharmacy fills
order
Order arrives on
unit
Is the
No medication
Call physician to Call pharmacist to correct?
verify correctness verify correctness Does it match
of order of medication the patient
chart?
Yes
Nurse administers
medication
This flowchart shows the basic steps in a traditional medication-use system. The process components are arranged
sequentially, and each stage can be expanded as necessary to show all possible steps.
152
TOOL
Gantt Chart TIPS FOR EFFECTIVE USE
PR
(See Figure 7-11) ■■ Leave enough space in the time line to write
OFIL
E
beneath each time period.

Stages to Use: Identifying proximate causes; identifying ■■ Write entries in a stair-step fashion, each step
root causes; identifying opportunities for improvement; below the one before it, so that overlapping steps
are clearly indicated.
■■ Color in the rectangles as each step is completed.
Purpose: To graphically depict the time line for long- term ■■ If the project is very complex and lengthy, consider
and complex projects, enabling a team to gauge its progress creating a Gantt chart for each phase or each
quarter of the year.

1. Agree on start and stop dates for the project, and outline
its major steps.
2. Draw a time line.
3. Write the first step of the project under the appropriate
time period. Enclose it in a rectangle long enough
to stretch across the length of time estimated for
completion.
4. Do the same for each of the succeeding steps.
Figure 7-11. Example: Gantt Chart
Task:
Person(s) Apr May Jun Jul Aug Sep Oct Nov Dec
Design Phase Responsible
Identify and appoint SH

credentialing committee
Identify performance SH and SR

measures
Define policies and SH and SR

procedures that outline
appointment, reappointment,
and privileging process
Develop credentialing SH and SR

application
This Gantt chart of a competency and privileging process helped one team determine what tasks to undertake in what
order. The chart details the target date and person(s) responsible for each task in the development process.
153
TOOL
Histogram TIPS FOR EFFECTIVE USE
PR
(See Figure 7-12) ■■ Data should be variable (that is, measured on
OFIL
E
a continuous scale such as temperature, time,

weight, speed, and so forth).
■■ Make sure data are representative of typical and
root causes; identifying opportunities for improvement;
current conditions.
■■ Use more than 50 data points to ensure the
emergence of meaningful patterns.
Purpose: To provide a snapshot of the way data are distrib-
■■ Be sure that the classes are mutually exclusive so
uted within a range of values and the amount of variation
that each data point fits into only one class.
within a given process, suggesting where to focus improve-
■■ Using K = 10 class intervals makes for easier
ment efforts
mental calculations.
■■ Be aware that the number of intervals can
influence the pattern of the sample.
1. Obtain the data sets, and count the number of data
■■ To construct the histogram, place the values for the
points.
classes on the horizontal axis and the frequency
2. Determine the range for the entire data set.
on the vertical axis.
3. Set the number of classes into which the data will be
■■ Be suspicious of the accuracy of the data if the
divided.
histogram suddenly stops at one point without
4. Determine the class width (by dividing the range by the some previous decline in the data.
number of classes).
■■ Remember that some processes are naturally
5. Establish class boundaries. skewed; do not expect a normal pattern every time.
6. Construct the histogram.
■■ Large variability or skewed distribution may signal
7. Count the data points in each class, and create the bars. that the process requires further attention.
8. Analyze the findings.
■■ Take time to think of alternative explanations for
the patterns seen in the histogram.
Figure 7-12. Example: Histogram
50
Number of Occurrences in Third Quarter
40
30
20
10
0
1–5 6–10 11–15 16–20 21–25
Days
This sample histogram was developed by an infusion therapy service to analyze turnaround time for authenticating
verbal orders from physicians. The irregular distribution suggests opportunities for improvement.
154
TOOL
Kaizen TIPS FOR EFFECTIVE USE
PR
(See Figure 7-13) ■■ Allow two to four weeks for the planning phase.
OFIL
E
■■ Make sure that the team includes members who

Stages to Use: Implementing and monitoring will contribute value to the process.
improvements ■■ If possible, give team members some training in
use of tools before the workshop.
Purpose: To improve a small, focused area in a week or less ■■ Keep communication lines with the project
champion open as the workshop progresses.
Simple Steps to Success: ■■ Include a celebration for the team at the end of the
1. Using a value stream map (see Value Stream Mapping project.
Tool Profile pages 172–173) as a road map, create a ■■ Allow at least three to four weeks for the follow-up
charter that clearly defines the issue or problem to be phase.
addressed through the kaizen effort.
2. Secure from the organization’s leadership a designated
project champion who will be willing to act upon the
result of the kaizen.
3. Assemble a multidisciplinary team of no more than
10 members, to be led by an experienced, impartial
facilitator, such as a Six Sigma Master Black Belt.
4. Conduct the kaizen workshop, which may last two to five
days, during which the team uses a variety of tools to
determine an action plan.
5. The team presents its plan at a report-out to the project
champion.
6. Implement and monitor.
155
Figure 7-13. Example: Kaizen Blitz Day-by-Day Plan
156
TOOL
Multivoting TIPS FOR EFFECTIVE USE
PR
(See Figure 7-14) ■■ Ensure that when combining ideas on the lists, the
OFIL
E
team members who suggested the ideas agree

with the new wording.
■■ Use letters rather than numbers to identify each
root causes; identifying opportunities for improvement
statement so that team members do not become
confused by the voting process.
Purpose: To narrow down a broad list of ideas (that is,
■■ Clearly define each idea so that it is easily
more than 10) to those that are most important and worthy
understood by everyone voting.
of immediate attention

1. Using a brainstorming list or other list, combine any
items on the list that are the same or similar.
2. Assign letters to items on the new list.
3. Determine the number of points each group member can
assign to the list. Each member uses a predetermined
number of points (typically between 5 and 10) to vote
on the items on the list.
4. Allow time for group members to assign points
independently.
5. Indicate each member’s point allocation on the list.
6. Tally the votes.
7. Note items with the greatest number of points.
8. Choose the final group, or multivote again.
Figure 7-14. Example: Multivoting

Improvement Opportunities Number of Votes
A. Facility safety management 3
B. Patient education 7
C. Staff orientation 5
D. Referral (authorization) 3
E. Care coordination and communication 1
F. Laundry 7
G. Medication profile 5
This figure shows the results of multivoting on priorities for improvement at an Indian health center. The team was able to
reach consensus on the need for prioritizing the laundering process.
157
TOOL
Operational Definition TIPS FOR EFFECTIVE USE
PR
(See Figure 7-15) ■■ Be as concise as possible.
OFIL
E
■■ If the definition has no numbers in it, it may not

Stages to Use: Identifying improvement opportunities be specific enough and could possibly be further
revised.
Purpose: To remove ambiguity so that everyone has the
same understanding of the issue being defined

1. Consider what is to be defined. It may be a project charter,
a customer requirement, or an aspect of data collection.
2. Determine the particular means of measurement that
would be appropriate in supporting the definition.
3. Write the definition, including the specific, concrete
measurement criteria to be used in determining whether
the terms of the definition have been met.
Figure 7-15. Example: Operational Definition
Medication errors: Number of patient deaths, paralysis,

coma, or other major permanent loss of function associated
with a medical error
158
TOOL
Pareto Chart TIPS FOR EFFECTIVE USE
PR
(See Figure 7-16) ■■ If the team is working from a fishbone diagram,
OFIL
E
the topic is the effect that has been targeted for

improvement. (See Fishbone Diagram Tool Profile,
Stages to Use: Identifying opportunities for improvement
pages 149–150.)
■■ When selecting factors for comparison, beware of
Purpose: To show which events or causes are most frequent
grouping several distinct problems together, which
and therefore have the greatest effect. Doing so enables a can skew the rank order. Refer to the fishbone
team to determine what problems to solve in what order. diagram, and use the most specific causes and
factors possible.
Simple Steps to Success: ■■ Be sure to mark the chart clearly to show the
1. Decide on a topic of study. The topic can be any outcome standard of measurement.
for which several potential causes have been identified. ■■ When analyzing the chart, keep in mind that
2. Select causes or conditions to be compared. Identify the numbers do not always tell the whole story.
factors that contribute to the outcome—the more Sometimes two severe complaints deserve more
specific, the better. attention than 100 minor complaints.
3. Set the standard for comparison. In many cases, the
standard is frequency, although factors may be compared
based on their cost or quantity. 8. Draw a bar for each factor. The bars represent how
4. Collect data. Determine how often each factor occurs often each factor occurs, the cost of each factor, or its
(or the cost or quantity of each, as appropriate). Use a quantity, as applicable.
checklist to help with this task. 9. Include additional features, if desired. By making a
5. Make the comparison. Based on the data collected in the few simple additions to the chart, a team can show
previous step, compare the factors and rank them from the cumulative frequency, cost, or quantity of the
most to least. categories in percentages.
6. Draw the chart’s vertical axis. On the left side of the 10. Add up the percentages. All the percentages for the
chart, draw a vertical line and mark the standard of causes need to add up to 100%.
measurement in increments.
7. List factors along the horizontal axis. Factors should be
arranged in descending order, with the highest ranking
factor at the far left.
159
Frequency of Responses
0
20
40
60
80
100
120
No review of patient
care information prior
to surgery
Inadequate presurgical
evaluation
No surgical plan/
preoperative findings
Failure to perform surgery

in a safe manner
Failure to take
responsibility for
actions
History of inadequate
documentation
160
Inadequate informed
Contributing Causes
members investigating a sentinel event involving a wrong-site surgery.

consent
Ignored pathology
reports
Figure 7-16. Example: Pareto Chart
Delay in reporting
incident
Performed procedures
without adequate
expertise
Untimely operative
dictation
One organization used a Pareto chart to rank the frequency of responses of selected root causes provided by team
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
TOOL
Relations Diagram TIPS FOR EFFECTIVE USE
PR
(See Figure 7-17) ■■ Have team members take turns placing ideas on
OFIL
E
the paper so that everyone participates.

Stages to Use: Identifying proximate causes; identifying ■■ Remembering the rules for brainstorming, do not
root causes; identifying opportunities for improvement criticize any ideas.

■■ Use bold markers for drawing the arrows.
Purpose: To generate understanding of how various aspects

of a problem are connected, including cause-and-effect
relationships

1. Write a definition of the issue (for example, a problem to 5. For each note, count the number of arrows to and from it,
be solved or a solution to be achieved) on a sticky note and write the totals beside each note. For example, three
and put it at the top of a large sheet of paper such as a arrows to and one arrow from could be written as “3/1.”
flipchart page on an easel. 6. Examine the totals. Notes with the highest totals will
2. Give each member of the group a pad of sticky notes and be the most important ideas to address in resolving the
then brainstorm for ideas about the issue, writing down issue. Those with the highest number of arrows out
one idea per note. represent proximate or possibly even root causes; those
3. One at a time, put each idea on the paper. As an idea is with the highest number of arrows in represent the main
added, discuss whether it seems related to any other effects. Circle these ideas for further analysis.
ideas already on the paper; if so, place it nearby (but not
touching; there should be space between the notes).
4. Determine what the relationships are between related ideas
and draw arrows to represent them. For example, if idea
A has an impact on idea B, draw a line with an arrow
pointing from A to B. Origination notes will indicate
possible causes, destination notes possible effects.
Ultimately each note will have one or more arrows
leading toward and/or away from it.
161
Figure 7-17. Example: Relations Diagram—Hospital
6 in 1 out
Scheduled
appointments
0 in 1 out 2 in 2 out
Doctors’ Emergency
pay level appointments
Equipment Administrative
quality and reliability workload
Support functions Changes in

availability scheduled appointments
Nurse
availability
1 in 5 out
Source: Andersen B, Fagerhaug T. Root Cause Analysis: Simplified Tools and Techniques, 2nd ed. Milwaukee. ASQ Quality Press, 2006.
Reproduced by permission.
162
TOOL
Run Chart TIPS FOR EFFECTIVE USE
PR
(See Figure 7-18) ■■ Be sure that the units and scale used in the chart
OFIL
E
present an accurate picture of the variation in order

to avoid creating a distorted picture of the problem.
■■ Create a hypothesis as to the cause of any
opportunities for improvement; implementing and moni-
excessive variation.
toring improvements
■■ Interview the responsible staff members to
determine the actual cause of excessive variation.
Purpose: To provide an overview of the variation in the
■■ Repeat the process and compare performance and
performance of a process
levels of variation on an ongoing basis.

1. Select appropriate units by which to measure variation.
2. Select appropriate units of time over which to measure
the variation.
3. Plot variation on the vertical axis along the horizontal
time line.
Figure 7-18. Run Chart
ED Wait Time: Average per Patient
140
135
130
Minutes
125
120
115
110
105
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Weeks
Source: Agency for Healthcare Research and Quality. Advanced Methods in Delivery System Research—Planning, Executing, Analyzing, and Reporting
Research on Delivery System Improvement. Webinar #2: Statistical Process Control. Jul 2013. Accessed Aug 21, 2015. http://www.ahrq.gov/professionals
/prevention-chronic-care/improve/coordination/webinar02/index.html.
163
TOOL Scatter Diagram

(Synonym: scattergram)
TIPS FOR EFFECTIVE USE
PR (See Figure 7-19) ■■ Select two variables with a suspected relationship
OFIL
E
(for example, delays in processing tests and total

volume of tests to be processed).
Stages to Use: Implementing and monitoring
■■ Use the horizontal (x) axis for the variable you
improvements
suspect is the cause and the vertical (y) axis for
the effect.
Purpose: To display the correlation—not necessarily the
■■ Construct the graph so that values increase while
cause-and-effect relationship—between two variables
moving up and to the right of each axis.
■■ The more the clusters form a straight line (which
could be diagonal), the stronger the relationship
1. Decide which two variables are to be tested. between the two variables.
2. Collect and record relevant data. Gather 50 to 100 paired
■■ If points cluster in an area running from lower left
samples of data involving each of the variables, and to upper right, the two variables have a positive
record them on a data sheet. correlation. This means that an increase in y may
3. Draw the horizontal and vertical axes. depend on an increase in x; if you can control x,
4. Plot the variables on the graph. If a value is repeated, circle you have a good chance of controlling y.
that point as many times as necessary. ■■ If points cluster from upper left to lower right, the
5. Interpret the completed diagram. variables have a negative correlation. This means
that as x increases, y may decrease.
■■ If points are scattered all over the diagram, these
variables may not have any correlation. (The
effect, y, may be dependent on a variable other
than x.)
■■ Remember, if the diagram indicates a relationship,
it is not necessarily a cause-and-effect relationship.
■■ Be aware that even if the data do not appear to
have a relationship, they may be related.
■■ Although scatter diagrams cannot prove a causal
relationship between two variables, they can offer
persuasive evidence.
164
Figure 7-19. Example: Scatter Diagram
165
This scatter diagram compares two variables associated with self-administration errors—the number of medications prescribed and the ages of the
patients involved. As might be expected, the clustering of points shows that the older the patient, the higher the number of medications involved in care.
TOOL
SIPOC Process Map TIPS FOR EFFECTIVE USE
PR
(See Figure 7-20) ■■ Use a cross-functional team to create the SIPOC
OFIL
E
process map.
Stages to Use: Identifying improvement opportunities ■■ Follow one of the key rules of brainstorming while
going through the steps to create the SIPOC
process map: Allow ideas to be expressed without
Purpose: To identify the basic elements or variables in
judgment or critique.
a process

1. SIPOC stands for suppliers, inputs, process, outputs, 4. Identify and list the inputs—that is, the information,
and customers. The SIPOC process map will have five materials, or personnel—needed for the process to
columns corresponding to each of these categories. produce the outputs.
Begin with a process map, perhaps generated from a 5. Identify and list the suppliers of the inputs—that is, the
brainstorming session (see Brainstorming Tool Profile, individuals or organizational groups that provide the
pages 138–139), that shows a process of about four or inputs.
five steps. List these steps in the center process column of 6. If needed, make further refinements; for example, by
the SIPOC process map. rewording, combining, or moving the items listed within
2. Identify and list the outputs—that is, the products or each column.
services being offered to the customers—that result from
this process.
3. Identify and list the customers—that is, the people who
are to receive and use the outputs. Customers may be
internal (another department in the organization) or
external (patients).
166
Figure 7-20. Example: SIPOC for a General Physician Visit
Suppliers Inputs Process Outputs Customers

Universities/ Doctors Treatment Patients
Patient visits
Institutions
medical center
Nurses Medical Report Patients’ Families
Pharmaceutical
Companies Experts Medical Record/ Management
Register the
Data
Support Staff patient Doctors
Medical Equipment
Companies Guidance
Medical Equipment Nurses
Drug Stores Patient receives Experience
Patient service(s) Support Staff
Pharmacists
Supplies
FDA Patient pays and
Laws leaves medical
Hospitals center
Facilities
Medical Centers
Insurance Plan
Laboratories
Drugs
Insurance
Companies
Source: Trusko BE, et al.: Improving Healthcare Quality and Cost with Six Sigma. Upper Saddle River, NJ: FT Press, 2007. Reproduced by permission.
167
TOOL
Stakeholder Analysis TIPS FOR EFFECTIVE USE
PR
(See Figure 7-21) ■■ Consider stakeholders at all levels, including both
OFIL
E
salaried and hourly employees.

Stages to Use: Implementing and monitoring ■■ If employees involved in the project are unionized,
improvements meet with the union representative to explain the

purpose of the project.
■■ Remember that while full support is ideal, not
Purpose: To gauge the level of support from key people
all stakeholders need to be at the same level of
involved in a process change
support for the project as a whole to succeed.

1. Obtain a project charter.
2. Based on the project charter, identify potential 4. Determine the level of support needed from each
stakeholders—that is, individuals who play a role in the stakeholder for the project to succeed. Assign numerical
project. Create a grid with column headings on the left values, using the same scale as in the previous step, and
side for the names and roles of these stakeholders. (Also place them in the second of the right-side columns.
allow space in the grid for at least three columns on the 5. Subtract the number in the columns from Step 3 from
right side, to be filled in during subsequent steps.) Row the number in the column from Step 4. The difference
by row, enter the name and role of each stakeholder. between the two numbers represents the “gap,” which is
3. Through focus groups or interviews, determine the level placed in the third of the right-side columns. Using the
of support (also known as buy-in) that each stakeholder 1-to-10 scale, if the gap number is 0, then the interests
shows for the project. Assign numerical values to reflect of the given stakeholder are perfectly aligned with the
each person’s level of support. For example, as shown in objectives of the process—congratulations! But more
Figure 7-21, on a scale of 1 to 10, a value of 1 represents likely, the number will be greater than 0. The higher the
low support, and 10 represents high support. Another number, the greater the discrepancy between support
possible type of scale uses zero (0) as a midpoint, needed and support shown.
representing neutral feelings about the project. A 6. Based on the stakeholder analysis, an action plan may
positive 1 or 2 (+1, +2) indicates mild or strong support, need to be developed to increase the level of support.
respectively, while a negative 1 or 2 (-1, -2) represents An optional column (shown at the far right in Figure
mild or strong resistance. In the first of the right-side 7-21) may be added to the grid to indicate the strategy
columns, fill in the level of support shown. called for.
168
Figure 7-21. Example: Stakeholder Analysis
169
TOOL
Standard Work TIPS FOR EFFECTIVE USE
PR
(See Figure 7-22) ■■ Strive for balance in the work required by each
OFIL
E
individual.
Stages to Use: Implementing and monitoring ■■ Encourage everyone involved to think of standard
improvements work as “the right way—every time.”
Purpose: To create a logical work flow with a minimum

of waste

1. Document the existing process steps for each individual
involved.
2. Identify which steps are waste—that is, they are of no
value to the patient.
3. Measure takt time (the rate of patient demand) and
cycle time (the time it takes to complete one cycle of an
operation).
4. Determine the optimum sequence of work steps for each
individual involved in the process.
5. Implement changes in the process to reflect the optimum
sequence.
6. Monitor the revised process and measure for success.
7. If the changes are deemed successful, standardize the
new process.
170
Figure 7-22. Example: Standard Work
171
TOOL
Value Stream Mapping TIPS FOR EFFECTIVE USE
PR
(See Figure 7-23) ■■ Keep the focus on the patient in any attempt to
OFIL
E
discern the value of a step in the process.

Stages to Use: Identifying opportunities for improvement; ■■ Choose and use the correct tools.
implementing and monitoring improvements ■■ Provide advanced education to leaders.

■■ Explain the purpose of the walk-through to unit,
Purpose: To help health care organizations operate success- section, or department staff.
fully by eliminating, or at least minimizing activities that ■■ Communicate progress and results throughout
do not add value—that is, any steps in a process that do not the organization.
contribute to a patient’s experience of value

1. Specify. Break down and analyze the entire value chain 4. Pull. After the true value stream is established,
to calculate the value delivered by each process area, services can be delivered in alignment with actual
as perceived by the patient. This task involves having a patient demand (that is, “pulled” along by the patient
cross-functional team literally walk through each step of according to the patient’s timetable of need) and not
the process, talking with workers in the value stream and by assumptions or arbitrary forecasts (that is, “pushed”
taking measurements of the time it takes to do the work onto the patient by the organization on a timetable set
and the time spent waiting for the next step. by the organization).
2. Map. After the value chain is broken down, the various 5. Perfect. The final, crucial step is bringing the principle of
steps are assigned value from the point of view of the continuous improvement to the initiative. It is essential
patient and mapped in their respective places according that an organization not allow the initial taste of success
to value delivered. Those areas that do not contribute to get in the way of its continual pursuit of perfection.
value are considered waste and either minimized or Ongoing process improvement is absolutely key to
eliminated. The remaining elements form the foundation sustaining a competitive edge.
for developing a true patient-centric value chain.
3. Flow. The purpose of this step is to create a smooth and
efficient process flow between the value-added steps
identified in order to transform the chain into a value
stream for the organization. The increase in flow will
ultimately improve lead times and eliminate “bullwhip”
effects, creating maximum efficiency and productivity.
172
Figure 7-23. Example: Value Stream Mapping
173
174
CHAPTER
8
Root Cause Analysis Case Studies
from the Field
CASE STUDY ONE
Root Cause Analysis of Article-at-a-Glance
Serious Adverse Events Among
Older Patients in the Veterans Background: Preventable adverse events are more likely to occur
Health Administration among older patients because of the clinical complexity of their care.
The Veterans Health Administration (VHA) National Center for
Patient Safety (NCPS) stores data about serious adverse events when a
Alexandra Lee, MS; Peter D. Mills, root cause analysis (RCA) has been performed. A primary objective of
MS, PhD; Julia Neily, MS, RN, MPH; this study was to describe the types of adverse events occurring among
Robin R. Hemphill, MD, MPH older patients (age ≥ 65 years) in US Department of Veterans Affairs
(VA) hospitals. Secondary objectives were to determine the underlying
I n the United States, adults 65 years of

age and older account for 13.0% of the
population, and it is estimated that this
reasons for the occurrence of these events and report on effective action
plans that have been implemented in VA hospitals.
demographic will grow to comprise 19.3% Methods: In a retrospective, cross-sectional review, RCA reports were
of the population by 2030.1 Hospitalized, reviewed and outcomes reported using descriptive statistics for all VA
older patients often have more complex hospitals that conducted an RCA for a serious geriatric adverse event
medical needs than younger patients2–4 from January 2010 to January 2011 that resulted in sustained injury or
and are at higher risk for adverse events.5,6 death.
Studies have shown that older adults,
compared to other age groups, are more Results: The search produced 325 RCA reports on VA patients
likely to experience significant morbidity (age ≥ 65 years). Falls (34.8%), delays in diagnosis and/or treatment
and mortality as a result of falls in the (11.7%), unexpected death (9.9%), and medication errors (9.0%) were
hospital, polypharmacy, and surgical the most commonly reported adverse events among older VA patients.
errors.7–9 Communication was the most common underlying reason for these
events, representing 43.9% of reported root causes. Approximately 40%
In 1996, The Joint Commission required of implemented action plans were judged by local staff to be effective.
hospitals in the United States to conduct
a root cause analysis (RCA) for adverse Conclusion: The RCA process identified falls and communication as
events causing sustained injury or death.10 important themes in serious adverse events. Concrete actions, such as
The Veterans Health Administration (VHA) process standardization and changes to communication, were reported
National Center for Patient Safety (NCPS) by teams to yield some improvement. However, fewer than half of the
was established in 1999 to lead the Depart action plans were reported to be effective. Further research is needed to
ment of Veterans Affairs’ (VA’s) develop- guide development and implementation of effective action plans.
ment of safety culture and to investigate
175
Figure 1. Root Cause Analysis (RCA) Process Within the Veterans Health Administration
The PSM assigns a

Patient safety The PSM charters an
Adverse event safety assessment
manager (PSM) interdisciplinary
occurs. score (SAC) for the
notified. team.
event.
Determine root cause

Gather and analyze
Determine actions and contributing
Determine how to information leading
to address the root factors, with a primary
measure outcomes. up to the adverse
cause(s). focus on systems
event.
issues.
Implement actions
Present the analysis
and examine
and proposed actions
outcomes to
to hospital leadership
determine
Administration for concurrence.
effectiveness.
The root cause analysis (RCA) process is intended to answer three major questions:
(1) What happened? (2) Why did it happen? (3) What can be done to prevent it from happening again?
Source: Adapted from Department of Veterans Affairs, Veterans Health Administration. VHA National Patient Safety Improvement Handbook (Page B-1). Mar
2011. Accessed Apr 25, 2014. http://www.va.gov/vhapublications/ViewPublication.asp?pub_ID=2389.
adverse events. By the end of 2000, the RCA process had summarized in Figure 1 (above), is intended to answer
been implemented across the VHA’s 152 medical centers, three major questions: (1) What happened? (2) Why
which serve more than six million veterans.11,12 did it happen? (3) What can be done to prevent it from
happening again?13
There is limited research on systems-level action plans to
improve care for hospitalized older adults. To date there When a serious adverse event occurs in the hospital, the VA
are no studies describing reports of iatrogenic events in the facility’s patient safety manager determines whether an RCA
older adult population within the VA health care system. should be initiated. He or she codes the event according to
The primary objective of this study was to describe the a standardized Safety Assessment Code (SAC) Matrix.11,14
types and frequency of reported adverse events occurring This matrix has two dimensions, one dimension based on
among older patients (age ≥ 65 years) in VA hospitals the severity of the event (catastrophic, major, moderate, and
nationally for one year. Secondary objectives were to deter- minor injuries) and the other, on the probability of recur-
mine the underlying reasons (root causes) for the occurrence (Figure 2, page 177). The adverse event may result in a
rence of these events, report on action plans that have been score of 3 (high risk), 2 (intermediate risk), or 1 (lowest risk).
implemented in VA hospitals, and to determine the extent The patient safety manager is required to conduct an RCA
of their effectiveness. review if the adverse event has scored a 3, which includes
all events resulting in a catastrophic patient outcome and
Methods events resulting in major injury that are likely to reoccur
Overview of the Veterans Health several times in one year. Patient safety managers may also
Administration National Center for Patient decide to conduct an RCA on events that have scored a 2,
Safety Root Cause Analysis Program which includes all other events resulting in major injury or
RCA is a systematic nonpunitive process to retrospectively events resulting in moderate injury and likely to reoccur
analyze adverse events and develop action plans to prevent several times in one year. It is unlikely that the patient safety
them from occurring in the future. The RCA process, as manager will conduct an RCA on a case that has scored a 1.
176
CHAPTER 8 | Root Cause Analysis Case Studies from the Field
Figure 2. Safety Assessment Code (SAC) Matrix

Severity
Catastrophic Major Moderate Minor

Patient death, or Permanent lessening Event leading to No injury, no increased
permanent loss of of bodily functions, not increased length length of stay, or
function, not related to related to the patient’s of hospital stay, or change in level of
the patient’s illness or natural course of illness increased level of patient care has
underlying condition patient care occurred
Frequent
Event is likely to
3 3 2 1
reoccur several times
in 1 year
Occasional
Probability
Event may occur

3 2 1 1
several times in
1 to 2 years
Uncommon
May happen once in 3 2 1 1
2 to 5 years
Remote
May happen once in 3 2 1 1
5 to 30 years
The Safety Assessment Code (SAC) Matrix is used to score cases on the basis of the severity of the event and the
probability of recurrence.
Source: Adapted from Bagian JP, et al. Developing and deploying a patient safety program in a large health care delivery system: You can’t fix what you don’t
know about. Jt Comm J Qual Improv. 2001;27(10):522–532.
Within the VHA, tools to standardize the RCA process as “worse,” “same,” “better,” “much better,” “not measured,”
have been developed, such as a computer-aided tool and a “not reported,” or “not implemented”).
flipbook containing triage questions to help teams deter-
mine a systems-based root cause for the event.11 Team Action items are informed by human factors concepts. This
members, who are chosen by the hospital’s patient safety means that proposed actions are developed to improve the
manager, include any staff member who works in the process or environment to minimize the likelihood that a
hospital, including physicians, nurses, social workers, and mistake would happen in the future, rather than placing
facilities management personnel. The team members may blame on an individual.16 For example, if a patient fell
be subject matter experts or nonexperts who were not because he or she did not have proper monitoring, the RCA
directly involved in the event being investigated. An expert team may consider whether fall prevention policies were
of the subject matter will be able to offer insight on how comprehensively written and easily understood by staff
the process is carried out. A nonexpert may be helpful in members, if overall staffing was sufficient, or if providers
identifying vulnerabilities in the current system.15 had clear means of communication to know that there was
a patient at high risk for falls under their care. These cases
Having individuals involved in the adverse event on the are stored in a secure computerized reporting system, where
RCA team may introduce bias on the basis of his or her they are available for review by NCPS staff.15
own perspective of the event. Instead, such individuals are
typically interviewed by team members to obtain further
details on the event. Finalized RCA reports include a brief
summary of the adverse event, root causes, action items,
how actions will be measured, and effectiveness (reported
177
Selection of Included Geriatric Defining an Effective Action Plan

Adverse Events RCA Reports Each case was reviewed to determine if implemented action
To identify all RCA reports of geriatric adverse events plans had any reported effect on patient outcomes or
occurring within a one-year period, we searched the NCPS changes in provider practices. An effective action plan was
database from January 2010 to January 2011. All RCA determined based on two criteria, as follows:
reports that listed the patient as 65 years old or older were • The action must have resulted in a “better” or “much
included in the analysis. Reports in which the patients’ better” outcome (on the basis of the report from the VA
age was unknown were manually reviewed and assessed facility where staff assessed the impact of the change).
on a case-by-case basis to determine eligibility (n = 6). • The type of the outcome measure must have been a
We read each case summary and included RCA reports process or outcome indicator as defined above.
that mentioned characteristics that were likely to indicate
that the patient was 65 years old or older, such as the fact We excluded structural indicators, as they do not clearly
that the patient was receiving care from a nursing home, translate into improvements in patient care.17
community living center, or hospice or had a diagnosis
typical for older patients. Results
Number of Included Root Cause
Cases were coded on the basis of the type of adverse event, Analysis Reports
root causes, action items, outcome measures, and effec- From January 2010 to January 2011, we identified 504
tiveness of each outcome. The root causes were classified potentially eligible RCA reports. We focused on events that
by using a predefined rubric developed by the NCPS that resulted in mortality or major to severe injury (permanent
organizes root causes into five categories—communication, lessening of bodily functioning, events requiring surgical
environment/equipment, training, rules/policies/proce- intervention or increased length of stay),11 resulting in 325
dures, and staff fatigue/scheduling problems, to which we RCA reports that were included in our final analysis. These
added “patient characteristics.” reports revealed 333 adverse events, 699 root causes, and
1,183 proposed action plans.
The type of outcome measure was classified into three cate-
gories using Donabedian’s framework for assessing quality Geriatric Adverse Events
of care: outcome, process, and structural indicators.17 An Falls were the most commonly reported event leading
outcome indicator describes the effect of care on the health to significant injury or death among patients (34.8%,
status of a patient or patients.18 In the context of measuring n = 116). Other events included delays in diagnosis and/
the effectiveness of fall prevention strategies, an example or treatment (11.7%, n = 39), unexpected death (that is,
of an outcome indicator includes a reduced rate of patient the clinical prognosis did not reflect the imminent death of
falls resulting in injury. A process indicator assesses changes the patient) (9.9%, n = 33), adverse drug events (9.0%,
in the provision of care18 (for example, the proportion of n = 30), and surgical errors (8.7%, n = 29). Figure 3
fall risk assessments conducted that include assessment (page 179) provides an overview of these adverse events.
of orthostatic hypertension), and a structural indicator
measures changes made to the care setting18 (for example, Root Causes
hiring of additional nursing staff.) A major limitation is Figure 4 (page 180) summarizes root causes as identified
that the progression from structure to process to outcome is using the NCPS framework.14 Communication, including
poorly established.17 both verbal exchanges and written documentation, was the
most common root cause of adverse events among older
Two researchers [A.L., P.D.M.] coded the RCA reports. VA patients (43.9%, n = 307). Within the communication
The first 10 cases were independently coded, with discrep- category, communication among providers (13.4% of all
ancies discussed until consensus was reached. The subse- root causes, n = 94), missed assessments (7.9%, n = 55),
quent 10 cases were coded achieving acceptable interrater lack of documentation (5.0%, n = 35), and communica-
reliability (kappa = 0.825). The remaining 305 cases were tion with patients and/or their caregivers (3.6%, n = 25)
independently coded. accounted for most of the communication issues.
178
DSSM
<L1> Figure 8-3. Root Cause Analysis Reports of Geriatric Adverse Events
Resulting in Serious Injury or Death in 2010, Categorized by Type of Event, N =
333
Figure 3. Root Cause Analysis Reports of Geriatric Adverse Events Resulting in
Serious Injury or Death in 2010, Categorized by Type of Event, N = 333
Further description of adverse event

categories:
• Unexpected death: Patient death in
which the clinical picture did not reflect
the imminent death of the patient
• Surgery: Retained object, incorrect
surgery (wrong side, site, or patient),
surgical wound infection, surgical
mortality
• Other: Adverse events that did not fall in
the listed categories
• Oxygen: Oxygen-related fires, oxygen
tank running out, others that led to an
adverse event
• Intubations: Difficult intubations
• Magnetic resonance imaging (MRI):
Adverse event related to the use of MRI
equipment
Falls were the most commonly reported event leading to significant injury or death among patients (34.8%, n = 116).
<L2> Further description of adverse event categories:
Problems•related
Unexpected
to equipmentdeath: Patient
and the death
hospital in which the
environ- Staffclinical picture
fatigue and did not
scheduling reflectaccounted
problems the for 7.3%
ment accounted for 21.6% (n = 151) of the root causes, (n = 51) of root causes, with work overload (3.6%, n = 25)
imminent
with issues mostly arising death
from lack of medical equipment and lack of available services/access delays (0.9%, n = 6)
(4.3%, n = 30), practical issues that arise from work with as the main contributing factors.
various technology systems (2.1%, n =15), and equipment
• of the patient
malfunction (1.9%, n = 13). The NCPS root cause analysis process advises RCA teams to
focus on systems issues when developing root causes; there-
• Surgery: Retained object, incorrect
Rules, policies, and procedures accounted for 16.9% surgery (wrong
fore, few side,
reportssite,
citeor patient),
patient surgicalas a root cause
characteristics
(n = 118) of the identified root causes, with the predom- for an adverse event.
wound infec-
inant reasons including procedures having a lack of a
standardized protocol (10.4%, n = 73) and lack of staff However, in some cases, when the team cannot determine
adherence • tion, surgical mortality
to a protocol (3.0%, n = 21). a systems-level root cause, the patient’s medical condition
or compliance with the treatment plan may be reported as
Insufficient or lack of staff training accounted for 7.7%
• Other: Adverse events that did not fall in thethelisted underlying cause for an adverse event. In our analysis,
categories
(n = 54) of root causes. Adverse events related to insuf- patient characteristics accounted for 2.6% (n = 18) of the
ficient/lack of training for one staff member (personal) root causes, specifically medical condition (2.1%, n = 15)
accounted for 0.7% (n = 5) of the root causes. Insufficient/ and compliance (0.4%, n = 3).
lack of training for two or more staff members (institu-
tional) accounted for 7.0% (n = 49) of the root causes.
179
Serious Injury or Death in 2010, N = 699
Figure 4. Root Causes for Reported Geriatric Adverse Events Resulting in

Serious Injury or Death in 2010, N = 699
General categories were derived from the Veterans Health Administration (VHA) National Center for Patient Safety
(NCPS) framework.
Figure 8-4. General categories were derived from the Veterans Health Administration (VHA) National
Action Plans
Of the 1,183 proposed actions,
Center 96.7% Safety
for Patient (n = 1,144)
(NCPS)wereframework.
Examples of process standardization included a neurolog-
implemented, of which 60.8% (n = 696) had outcomes ical examination in fall risk assessments, development of
assessed and were documented in
[END FIGURE] the RCA report. Of the order sets for high-risk procedures, and implementation
action plans for which outcomes were assessed, 39.5% of a standard protocol for the transfer of patients. Both
(n = 275) were considered effective on the basis of the two changes made to verbal communication (52%, 13 of 25
criteria defined in the Methods. action plans) and written documentation (29.9%, 40 of
Examples of action plans to enhance 134 action plans) resulted
communication includedin effective outcomes.
conducting
Figure 5 (page 181) summarizes all actions that were pro-
posed by the RCA teams, as well as effective
multidisciplinary action plans.
meetings Examples
to ensure that of action
patients plans
at high risktoofenhance
fallingcommunication
were receiving
The most common action plans included process standard- included conducting multidisciplinary meetings to ensure
ization (19.1%, n = 226), staff training (15.5%, n = 183), that patients at high risk of falling were receiving appro-
changes made to written documentation (11.3%, n = 134), priate interventions, standardizing documentation to
changes made to information display (8.9%, n = 105), and keep advance care directives up to date, and streamlining
revisions/updates made to a current policy (6.4%, n = 76). communication between providers and caregivers on the
The most effective actions plans were related to process patient’s plan of care.
standardization, communication, equipment, and environ-
mental changes. Equipment-related (50%, 9 of 18 action plans) and envi-
ronmental changes (25.8%, 8 of 31 action plans) also led
Eighty-nine of the 226 (39.4%) action plans involving to improved outcomes. Examples of these changes included
process standardization yielded an effective outcome. purchasing and implementing the use of nonskid socks for
180
Root Cause Analysis for Health Care, 5th ed.

Senior Editor: Jim Parker, x5459, [email protected]
Reviewers:Figure
Ronald5. RootMD;
Wyatt, Cause Analysis
Gerald (RCA)
Castro, Anita Reports
Guintoli, ofVanOsterberg,
Paul Action PlansDoreen
for Geriatric
Finn, Adverse Events
DSSM Resulting in Serious Injury or Death in 2010, N = 1,183
This figure summarizes all actions that were proposed by the RCA teams, as well as effective action plans.
high fall-risk patients, removing clutter from patient rooms, fall-risk status and needed precautions. In addition, physi-
and using a checklist to verify and restock crash carts. cians may not have been contacted when a patient was
Figure 8-5. This figure summarizes all actions that were proposed by the RCA
found teams,orthostatic
to have as well ashypotension,
effective and pharmacists
As discussed previously, falls, delays in care, unexpected may not have been consulted when a medication review was
actiondeath,
plans. medication errors, and surgical errors were the needed. Other significant contributors included problems
most commonly reported adverse events among older VA with equipment and the hospital environment.
[END patients. Implemented action plans resulting in a “much
FIGURE]
better” patient outcome for these events are summarized in The NCPS root cause analysis process focuses on iden-
Tables 1–5 (page 182). tifying root causes from a systems-level perspective and
implementing subsequent action plans that provide a
Discussion structure to minimize human error rather than placing
[TABLE]
Our findings indicate that communication problems were blame on individual providers. Approximately 40% of the
the most common underlying reason for reported adverse completed action plans resulted in an effective outcome.
events among older VA patients in 2010. Such problems Concrete actions, such as process standardization, changes
included inconsistent communication between providers to communication, and changes to equipment or the
Table 1. Root
and lack Causes
of clear and Sampling
documentation regarding of
the Action
patient’s Plans
fall- with a “Much
environment, Better”
were Reported
most likely to yield Outcome
improvement. The
risk status. For example, handoff communication between report from bedside clinicians regarding the effectiveness of
Related
nurses atto Falls*
change of shift did not always include patients’ action plans is a unique feature of this report.
181
Table 1. Root Causes and Sampling of Action Table 5. Root Causes and Sampling of Effective
Plans with a “Much Better” Reported Outcome Action Plans with a “Much Better” Reported
Related to Falls* Outcome Related to Adverse Events in Surgery*
Root Cause, % (n) Root Cause, % (n)
N = 225 N = 64
Communication 39.1% (n = 88) Rules/policies/procedures 32.8% (n = 21)
Environment/equipment 32.9% (n = 74) Communication 31.3% (n = 20)
Rules/policies/procedures 11.6% (n = 26) Environment/equipment 21.9% (n = 14)
Training 7.6% (n = 17) * Not included: Training, 9.4% (n = 6); Fatigue/scheduling, 4.7% (n = 3).
Staff fatigue/problems with scheduling 4.0% (n = 9)

* Not included: Patient characteristics, 4.9% (n = 11). There have been many studies conducted assessing falls,
medication errors, and surgical errors among older patients.
Table 2. Root Causes and Sampling of Effective Falls, in particular, are a common and serious event in the
Action Plans with a “Much Better” Reported hospital. Generally, fall rates among all ages have been
Outcome Related to Delays in Care* reported from 2.9 to 13 falls per 1,000 bed-days of care.7
Root Cause, % (n) Falls may result in injury (both psychological and phys-
N = 81 ical), leading to longer hospital stays and increased risk of
Communication 50.6% (n = 41) discharge to a long term care facility.19,20 Reducing falls in
Rules/policies/procedures 17.3% (n = 14) the hospital is a complex problem that involves creating a
* Not included: Staff fatigue/problems with scheduling, 14.8% (n = 12);
safer environment and implementing individualized inter-
Environment/equipment, 12.3% (n = 10); Patient characteristics, 1.2% ventions to modify patient risk factors for falls and fall-re-
(n = 1); Other, 3.7% (n = 3). lated injuries.21
Table 3. Root Causes and Sampling of Effective Although most medication errors do not result in injury,22,23
Action Plans with a “Much Better” Reported
polypharmacy is common among older patients, which
Outcome Related to Unexpected Death*
suggests that they may be at greater risk of experiencing
Root Cause, % (n) an adverse drug event.8 For example, among 1,523 adverse
N = 92 drug events identified by Gurwitz and colleagues in older
Communication 47.8% (n = 44) patients in an ambulatory setting, 27.6% were judged
Environment/equipment 14.1% (n = 13) preventable. These events occurred in the prescribing
Rules/policies/procedures 13.0% (n = 12)
(58.4%) and monitoring (60.8%) stages of pharmaceutical
care. Patient adherence was implicated in 21.1% of the
Training 10.9% (n = 10)
cases.8 In our analysis of the RCA database, issues that arose
* Not included: Fatigue/scheduling, 12.0% (n = 11); Patient characteristics,
from the inpatient bar-coding medication administration
2.2% (n = 2).
system, patient monitoring, lack of medication reconcil-
Table 4. Root Causes and Sampling of Effective
iation, and patient compliance were contributory factors
Action Plans with a “Much Better” Reported toward adverse drug events in older VA patients.
Outcome Related to Medication Errors*
Root Cause, % (n)
Older patients have the highest mortality rate in the adult
N = 64 surgical population.9 Preoperative assessment can help
guide the treatment course and determine if high-risk
Communication 53.1% (n = 34)
surgery is necessary for the older patient, or if a less invasive
Rules/policies/procedures 14.1% (n = 9) procedure is more appropriate.24 Furthermore, postopera-
Training 7.8% (n = 5) tive monitoring may prevent the occurrence of an adverse
* Not included: Environment/equipment, 20.3% (n = 13); Fatigue/scheduling, event. Again, we found problems with monitoring patients
4.7% (n = 3).
182
and missed or incorrect assessment to contribute toward the reviewed cases and that they seemed feasible to imple-
adverse events. ment at other hospital facilities.
• Ensure that the patient’s advance care directives are
Among reported action plans, staff training (15.5%, documented and clearly communicated to hospital
n = 183) and improving a policy (6.4%, n = 76) were providers.
frequently implemented; however, they did not necessarily • Standardize handoff communication among providers,
translate into improved processes and patient outcomes. using tools such as pocket guides or electronic templates.
Such actions are considered weak compared with actions • Enhance education provided to hospital staff through
that might change the environment in a way that will make interactive drills, such as a mock code or simulation
a fall less likely to happen (for example, removing a rug that training.
a patient might slip on). Moreover, our findings suggest • Provide patients at high risk of fall with proper
that reporting a patient’s medical condition or compli- equipment (for example, nonslip footwear, bed alarms,
ance with the treatment plan as an underlying cause does or chair alarms) and mitigate environmental hazards
not promote the development of effective action plans for (for example, unnecessary furniture and clutter). If these
hospital patient safety and quality improvement efforts. interventions cannot be put in place, one-on-one staff
supervision may be necessary.20
After the RCA team has developed action plans, the next • Standardize the stocking of crash carts through the use of
step is implementation. The NCPS root cause analysis a checklist. In addition, human factors principles can be
framework does not specifically comment on how to imple- implemented in organizing the crash cart (for example,
ment actions; such strategies will differ for each institution, arranging supplies according to order or frequency of use
depending on numerous factors, such as the type of action in a way that is common to all units).
and how they will meet the needs of patients and providers • Implement a time-out procedure before all surgical
within a particular setting. However, the RCA team assigns procedures, include discussion of potential complications
an individual or team the responsibility of implementation of the case involving an older patient, and standardize
within a predefined time period. Following implementation the count of supplies prior to closing the patient.
and outcome assessment, effectiveness of the action plan is Support from hospital leadership, appropriate human
reported back to the patient safety manager. We found that factors education, and regular multidisciplinary team
59.3% of the proposed action plans were implemented and communication may help promote staff compliance with
measured for effectiveness—a relatively low rate that high- this process.25,26
lights difficulty of implementing changes. • Review of patient’s medications and medication
reconciliation may avert many iatrogenic events, such as
The NCPS root cause analysis process provides a standard- falls, unexpected death, and medication errors.
ized framework for identifying root causes of an adverse
event and developing an action plan. The NCPS uses a Limitations and Conclusions
variety of strategies to share aggregated RCA findings, such Results obtained are based on the written reports of patient
as sharing lessons learned, publishing manuscripts, issuing safety managers at local VA facilities, but the staff who
patient safety alerts, and sharing information among patient originally submitted the reports of adverse events and
safety managers. conducted the RCAs did not have this study in mind. At
the same time, the main purpose of submitting RCAs to a
Recommendations for Hospital Quality central database is to enable the aggregation and subsequent
Improvement Teams dissemination of the lessons with the field. This study does
We recommend implementation of the following action not include adverse events caused by individual practi-
plans to provide effective care to patients and minimize tioner-specific practice issues or errors in judgment, which
geriatric adverse events. We were able to determine that are complex and difficult to determine and manage in a
these action plans or similar action plans were implemented system other than RCA. We did not independently verify
at multiple VA facilities and yielded effective outcomes in the process used to identify root causes or develop action
183
plans. Adverse events in the VA system are based on volun- References

tary reports submitted to the facility patient safety manager. 1. Department of Health & Human Services, Administration
There is likely variation among facilities regarding the on Aging. Projected Future Growth of the Older Population.
Accessed May 3, 2014. http:// www.aoa.gov/aoaroot/aging
percentage of actual adverse events that are reported. We
_statistics/future_growth/future_growth.aspx.
do not know how the actual rate of adverse events compares
2. Kleinpell RM, Fletcher K, Jennings BM. Reducing functional
to the reported rate. We do know that events that are
decline in hospitalized elderly. In Patient Safety and Quality: An
reported provide us with rich opportunities to improve the Evidence-Based Handbook for Nurses. Rockville, MD: Agency for
safety of care for older patients. Despite the variation that Healthcare Research and Quality, 2008. Chapter 11. Accessed
may occur in reporting among national VA hospital facil- May 3, 2014. www.ahrq.gov/qual/nurseshdbk/ docs/KleinpellR
ities, the NCPS does take steps to standardize the process. _RFDHE.pdf.
Parts of the program include a three-day training session 3. Keehan SP, et al. Age estimates in the national health accounts.
for all new patient safety managers, resources via the NCPS Health Care Financ Rev. Epub 2004 Dec 2.
intranet and Internet sites, and monthly conference calls.11 4. Mattie AS, Webster BL. Centers for Medicare and Medicaid
Services’ “never events”: An analysis and recommendations to
In addition, patient safety managers at VA facilities hospitals. Health Care Manag (Frederick). 2008;27(4):338–349.
report on the effectiveness of their action plans. There is a 5. Leape LL, et al. The nature of adverse events in hospitalized
potential for bias, as facilities may want to positively report patients. Results of the Harvard Medical Practice Study II.
N Engl J Med. 1991 Feb 7;324(6):377–384.
on their efforts. However, all patient safety managers are
6. Thomas EJ, Brennan TA, Incidence and types of preventable
trained in a standardized manner, and there are no penal-
adverse events in elderly patients: Population based review of
ties for reporting on an action plan that was found not to
medical records. BMJ. 2000 Mar 18;320(7237):741–744.
be effective.
7. Hornbrook MC, et al. Preventing falls among community-
dwelling older persons: Results from a randomized trial.
In some cases, multiple actions were implemented simul- Gerontologist. 1994;34(1)16–23.
taneously, and therefore it was difficult to determine the 8. Gurwitz JH, et al. Incidence and preventability of adverse drug
specific actions responsible for improved patient care. events among older persons in the ambulatory setting. JAMA.
However, in the RCA reports, when multiple components 2003 Mar 5;289(9):1107– 1116.
of an action plan are implemented, patient safety managers 9. Pedersen T, Eliasen K, Henriksen E. A prospective study
are prompted on a computerized system to report outcomes of mortality associated with anaesthesia and surgery: Risk
on each individual action. Therefore, we were able to draw indicators of mortality in hospital. Acta Anaesthesiol Scand.
limited conclusions on the effectiveness of these compo- 1990;34(3):176–182.
nents. Finally, patients included in the analysis are from VA 10. O’Leary DS. President’s Column: The Sentinel Event Policy:
A work in progress. Jt Comm Perspect. 1996;16(6):2–3.
hospitals and may not accurately reflect the general popu-
lation. Despite these differences in patient characteristics, 11. Bagian JP, et al. Developing and deploying a patient safety
program in a large health care delivery system: You can’t
we believe that our findings fill a gap in knowledge toward
fix what you don’t know about. Jt Comm J Qual Improv.
describing adverse events among older patients in a national 2001;27(10):522–532.
health care system and that the implemented action
12. Weeks WB, Bagian JP. Developing a culture of safety
plans may provide insight for other institutions’ patient in the Veterans Health Administration. Eff Clin Pract.
safety efforts. 2000;3(6):270–276.
13. Bagian JP. Patient safety; It’s not rocket science. Paper presented
Disclaimer: This material is the result of work at the 2011 Northwest Patient Safety Conference, Seattle, May
supported with resources and the use of facilities at the 19, 2011. Accessed May 3, 2014. http://www.wapatientsafety
US Department of Veteran Affairs (VA) National Center .org/downloads/1107_Bagian_ Handout_May_2011.pdf.
for Patient Safety (NCPS) Field Office in White River 14. Department of Veterans Affairs, Veterans Health Administration.
Junction, Vermont, as well as the NCPS. The views VHA National Patient Safety Improvement Handbook,
expressed in this article do not necessarily represent Appendix B, “The Safety Assessment Code (SAC) Matrix.”
the views of the Department of Veterans Affairs or of Mar 4, 2011. Accessed May 3, 2014. http:// www.va.gov
the United States government. /vhapublications/ViewPublication.asp?pub_ID=2389.
184
15. Bagian JP, et al. The Veterans Affairs root cause analysis system in 22. Bates DW, et al. Relationship between medication errors and
action. Jt Comm J Qual Improv. 2002;28(10):531–545. adverse drug events. J Gen Intern Med. 1995;10(4):199–205.
16. Nolan TW. System changes to improve patient safety. BMJ. 2000 23. Institute of Medicine. To Err Is Human: Building a Safer Health
Mar 18;320(7237):771–773. System. Washington, DC: National Academies Press, 2000.
17. Donabedian A. Evaluating the quality of medical care. Milbank 24. Jin F, Chung F. Minimizing perioperative adverse events in the
Q. 2005;83(4):691–729. elderly. Br J Anaesth. 2001;87(4):608–624.
18. Mainz J. Defining and classifying clinical indicators for quality 25. Neily J, et al. Association between implementation of a medical
improvement. Int J Qual Health Care. 2003;15(6):523–530. team training program and surgical mortality. JAMA. 2010 Oct
19. Nevitt MC, Cummings SR, Hudes ES. Risk factors for injurious 20;304(15):1693–1700.
falls: A prospective study. J Gerontol. 1991;46(5):M164–170. 26. Haynes AB, et al. A surgical safety checklist to reduce morbidity
20. Peel NM, Kassulke DJ, McClure RJ. Population based study and mortality in a global population. N Engl J Med. 2009 Jan
of hospitalised fall related injuries in older people. Inj Prev. 29;360(5):491–499.
2002;8(4):280–283.
21. Oliver D, Healey F, Haines TP. Preventing falls and fall related
injuries in hospitals. Clin Geriatr Med. 2010;26(4):645–692.
Source: Lee A, et al. Root cause analysis of serious adverse events among older patients in the Veterans Health Administration. Jt Comm J Qual Patient Saf. 2014
Jun;40(6):253-262.
185
CASE STUDY TWO

Using Root Cause Analysis and Form Article-at-a-Glance
Redesign to Reduce Incorrect Ordering
Background: Advances in molecular biology and
of HIV Tests
changes in microbial nomenclature may subject diag-
nostic microbiology to errors. A patient diagnosed
Reed A.C. Siemieniuk; Kevin Fonseca, PhD;
with Pneumocystis jiroveci pneumonia and then with
M. John Gill, MB, FRCPC
AIDS had received a negative “AIDS test”—“negative
T he unrealistic expectation for physicians to prac- for antibodies to HTLV 1 and 2.” The test requisition
tice flawless medicine persists to a great extent, both showed that the physician had requested HTLV-I/II
within the medical community and within the population testing but not an HIV-1/2 test. A root cause anal-
at large, more than a decade of focused efforts to unearth ysis was performed to determine if the erroneous
and “humanize” the entrenched nature of medical errors testing represented a systemic problem. A study was
notwithstanding.1,2 Although there have been sporadic conducted to identify and address such testing errors.
advances in the identification and prevention of errors,
particularly with regard to medication administration3,4 Methods: For the 1,952 HTLV-I/II test requests in
and surgical procedures,5,6 errors in arriving at the correct a 17-month period in the Southern Alberta region,
diagnosis remain elusive. a random representative sample of 555 requests for
HTLV-I/II testing were evaluated for appropriate-
Errors in arriving at the correct diagnosis may make up a ness. Physicians ordering “inappropriate” tests were
large proportion of all medical errors and cause considerable surveyed to determine root causes, and the HTLV-I/
morbidity and mortality.7,8 The human factors components, II check box was subsequently removed from the
including formulating a diagnosis (which involves appro- requisition.
priate clinical suspicion and acumen, requesting the orig-
inally intended tests, and accurate interpretation of these Results: Some 318 (94%) of the 340 clinically
tests), are at high risk of error in the increasingly complex directed HTLV tests were likely or definitely inappro-
nature of medical practice. Reports of such errors to date, priate—that is, only an HIV-1/2 test was required.
however, have been scarce8 and have relied on convenience At least 81% (127/156) of the 8% (156/1,948) of
samples such as retrospective self-reports by physicians9 or the HTLV-I/II tests ordered without an HIV-1/2 test
reviews of malpractice claims.10 concurrently were ordered inappropriately. In the
telephone survey, all 69 physicians suspected to have
Given the rapid changes in microbial nomenclature, it may incorrectly ordered HTLV-I/II tests reported errone-
be particularly difficult for practicing physicians to remain ously requesting HTLV for HIV. A root cause analysis
current and ensure appropriate ordering of tests for infec- identified confusing viral nomenclature, diagnostic
tious diseases. With the increasing use of novel molec- testing menu, and form design as contributing factors.
ular techniques, more precision has become available for A requisition recall and redesign has reduced erro-
the classification and consequently for naming and even neous laboratory testing.
renaming of microorganisms and viruses. Taxonomy experts
have identified the classification and naming of viruses as Conclusions: A high proportion of HTLV-I/
a “logistical challenge,” with thousands of new viruses and II tests were ordered erroneously and confused with
sequences needing classification.11 In the similar situa- HIV-1/2. Careful attention to routine test menus
tion of naming of both new and generic drugs, confusing and form design, including the exclusion of rare and
look-alike, sound-alike names have been shown to cause confusing pathogens, reduces risk of error for prac-
serious errors in drug administration, resulting in adverse ticing physicians.
patient outcomes. To protect practicing physicians from
this systemic risk for medical errors, the World Health
Organization and the US Food and Drug Administration
186
have actively initiated major programs to preclude the use A root cause analysis (RCA) was performed to determine if
of confusing drug names.3 the erroneous testing for HTLV-I/II rather than HIV-1/2
(when AIDS was a concern) represented a systemic
Human immunodeficiency virus (HIV) and human T-cell problem. We report on the identification and intervention
lymphotropic virus (HTLV) share some history. The of testing errors to reduce such risks for the future.
virus responsible for causing AIDS was first described in
1983.12 In North America the name most commonly used Methods
was HTLV-III. This name was initially chosen because Design Overview and Setting
the virus was closely related to other members of the All laboratory requisition forms requesting HTLV-I/
family Retroviridae_HTLV-I and -II.12,13 In 1986 a special II testing submitted to the Southern Alberta Laboratory
subcommittee of the International Committee on Viral of Public Health (SALPH; Calgary, Alberta, Canada)
Nomenclature standardized the “AIDS virus” nomencla- between March 1, 2008, and July 31, 2009, were reviewed
ture, officially naming it HIV.14 HTLV-IV was discovered in for appropriateness on the basis of clinical information
198615,16 and also had several names before its current name provided by the physician (Figure 1, page 188).
of HIV-2 was established in 1987.17
The SALPH is the sole provider of HIV-1/2 and HTLV-I/II
HTLV-I is a human pathogen associated with two rare testing services (apart from blood transfusion service) for
conditions: Tropical spastic paraparesis and adult T-cell a diverse region composed of approximately 1.6 million
leukemia.18 These conditions are seldom seen in North people, most of whom live within the city of Calgary. There
America but are reported more widely in southwestern are 39 hospitals, with more than 3,500 practicing physi-
Japan, parts of Africa, parts of South America, and the cians, within the laboratory’s catchment area. A single stan-
Caribbean.19 Recent evidence suggests that HTLV-I may dardized requisition form that requests clinical information
also be associated with bronchiectasis in some indigenous is employed in the region.
populations20 and a rare form of infective dermatitis.21
To date, the only evidence of HTLV-II–associated disease All known HIV-1/2–positive patients, including the index
comes from occasional case reports; as such, its pathoge- patient, were excluded from the study, as were all tests
nicity is uncertain.22 Because both viruses may be trans- performed within recognized and well-established fertility
mitted parenterally, testing for HTLV-I and HTLV-II has and tissue donation programs. This study was approved
been undertaken routinely for more than a decade for by the Conjoint Medical Bioethics Committee at the
tissue/blood screening purposes and on a case-by-case basis University of Calgary.
for clinical diagnosis of HTLV-I disease. Reports have
described two potentially clinically significant retroviruses Requisition Review
that have been named HTLV-3 and HTLV-4.23,24 The “study,” which we conducted from August 2009
through March 2011, consisted of two parts, known as
In 2009 a 21-year-old man presented to the hospital with Study 1 and Study 2. In Study 1, drawing on a comput-
cough, fever, shortness of breath, and weight loss. He was er-generated random number list, we selected a represen-
diagnosed with Pneumocystis jiroveci pneumonia and subse- tative sample of 555 HTLV- I/II requisitions and reviewed
quently with AIDS. This diagnosis surprised him because each requisition for appropriateness. In Study 2, we evalu-
he had recently received a negative “AIDS test” after ated all HTLV-I/II tests requisitions that did not concur-
disclosing a high-risk lifestyle to his physician. He produced rently order HIV-1/2 testing. Reviews of all requisitions was
a copy of the report of this earlier test, which showed performed by one of the authors [R.A.C.S.] and reviewed
that he was “negative for antibodies to HTLV 1 and 2.” by a second [M.J.G.].
A review of the test requisition showed that the physi-
cian noted significant HIV risks and requested chlamydia, The requisition’s mandatory fields, which were used to
hepatitis, and HTLV-I/II testing but not an HIV-1/2 test. judge appropriateness of testing, included patient iden-
Subsequent testing of the original sample confirmed that tifiers, ordering physician information, and check boxes
this index patient was positive for HIV-1/2 at that time. for testing requested. Additional clinical information
187
Figure 1. Flow Diagram of the Study Design
HTLV-I/II testing requests were reviewed from a random sample of all tests ordered without HIV-1/2. Physicians ordering
HTLV-I/II tests without HIV-1/2 likely or definitely inappropriately were surveyed by telephone. These samples were also
blinded and tested for HIV-1/2.
requested included check boxes for HIV risk factors, such specimen testing for chlamydia and other sexually trans-
as homosexual/bisexual orientation, HIV–positive partner, mitted pathogens.
multiple sexual partners, risks associated with an endemic
country, sexual assault, HIV exposure, intravenous drug In both Study 1 and Study 2, each HTLV-I/II test was as-
use, hemophilia, HIV–positive mother, tuberculosis, and signed into one of the following four categories for appro-
anxiety because of nondefined high-risk behavior. Check priateness on the basis of the information available:
boxes for symptoms included fever, rash, respiratory 1. Fertility/tissue donor screening: Fertility screening
symptoms, adenopathy, neurological symptoms, immune prior to tissue or sperm donation ordered by individual
status, and systemic symptoms, as well as free-text space for physicians without formal laboratory preapproval
details. The detailed risk information requested for public 2. Appropriate: HTLV-I/II testing supported by disease-
health purposes during the study period has now been compatible history (for example, paresis or leukemia)
eliminated as we move to the recommended widespread or ordered by a neurologist, oncologist, or infectious
population-based testing with limited need for risk-based disease specialist
assessments. Additional check box testing included sero- 3. Likely inappropriate: Unfocused HTLV-I/II testing
logic testing for hepatitis (B and C) and syphilis, as well as ordered by a physician other than a neurologist,
188
oncologist, or infectious disease specialist with a lack box on the test menu for a rare pathogen, the adjacency
of information for HTLV-I/II testing and testing for of HIV-1/2 and HTLV-I/II check boxes, and a small
multiple unrelated viral and bacterial conditions font (Arial, size 7)
4. Inappropriate: HTLV-I/II testing with symptoms, risk 2. Office practices. In some cases, nonprofessional office
factors, or concurrent sexually transmitted pathogen staff filled out requisition forms, and/or the exact source
screening strongly suggestive of HIV-1/2 without any of the error remained unidentified but often suggestive
clinical information suggestive of HTLV-I/II of busy clinics with hasty form completion.
3. Confusing viral nomenclature. Some physicians
Blinded Seroprevalence Study were under the impression that HTLV and HIV were
Provided there was an adequate residual sample, samples synonymous. This was particularly noted for HIV-1
originating from inappropriate or likely inappropriate versus HTLV-I/II.
testing (categories 3 and 4) for HTLV-I/II alone were
blinded and tested for antibodies to HIV-1/2 with enzyme- Intervention
linked immunosorbent assay (Abbott, North Chicago, Following the RCA, HTLV-I/II testing and other
Illinois). The primary outcome was prevalence of HIV-1/2 uncommon tests were removed as check box options, on
infection among tests where the original intention was the requisition, and the form was officially recalled. The
likely HIV-1/2 testing. results of the interventions were reviewed by the testing
laboratory, and process adjustments implemented as below.
Physician Contact
All physicians ordering HTLV-I/II without concurrent Requisition Review, Recall, and Redesign
HIV-1/2 testing and evidence of an intention for HIV-1/2 Because the RCA implicated, in part, test menu and poor
testing were sent registered mail in October 2009 to recom- form design in the inappropriate testing, a recall of the
mend contacting the patient for HIV-1/2 testing. A letter forms was initiated, with a replacement that excluded
from the University of Calgary’s Chair of Ethics and a new HTLV-I/II testing from the open menu but allowed
laboratory requisition with completed patient identifiers a write-in option. The impact of this intervention on
were included. These physicians were also contacted by HTLV-I/II testing was assessed by evaluating the rate of
telephone by one of the investigators [K.F.] to determine HTLV-I/II tests likely mistaken for HIV-1/2 tests. On the
the original testing intention and to solicit ways to improve basis of the results from the original review, the identifica-
the requisition process. tion of any HIV-1/2 risk factor was again used as a marker
of inappropriate HTLV-I/II testing. The primary outcome
Root Cause Analysis was a decrease in inappropriate tests per month before the
The incident case described in the introduction was brought form redesign (January 2008–April 2010) versus after the
to the attention of the medical director of the regional HIV redesign (May 2010–November 2010). P values were calcu-
program [M.J.G.], who, with the HIV testing laboratory, lated with a two-tailed t-test.
initiated an investigation into the root cause of the error.
A voluntary group of three persons, representing the labo- Results
ratory [K.F.] end-users [R.A.C.S., M.J.G.], and infectious Requisition Reviews
diseases [M.J.G.], was established to determine the extent Between March 2008 and July 2009, 1,952 HTLV-I/II tests
of the problem, all contributing factors, and any correct- were ordered, for which the requisitions were available for
able systemic issues. Additional and extensive voluntary 1,948. In the random representative sample of 555 requi-
input from interested parties with expertise, including the sitions, 241 (43%) were requested from within Calgary, 96
laboratory academic staff, the Public Health Office, and the (17%) from rural communities, and 218 (39%) did not
University Bioethics Committee, was formally sought out list a location. Some 214 (39%) of the requisitions orig-
and provided. inated from fertility or tissue donation programs, and 1
requisition was not found; these screening programs were
The following three sources of concern were identified: excluded from further analysis. Some 340 (61%) tests were
1. Test menu and form design. The provision of a check ordered for clinical purposes, of which 318 (94%) were
189
inappropriate or likely inappropriate HTLV-I/II testing Table 1. Study 1: Analysis of a Random Sample
using criteria listed (Table 1, right). The most frequent of 555 HTLV-I/II Test Requisitions
risk factors listed were patient anxiety (49%) and multiple Indication n (%)
sexual partners (19%), while the most frequent ancil- 1. F
ertility/tissue donor screening 214 (39)
lary tests requested were hepatitis B and/or C (61%) and Clinically directed HTLV-I/II tests 340 (61)
syphilis (44%) (Table 2, right). 2. Appropriate or likely appropriate 22 (6.5)
3. Likely inappropriate 80 (24)
One hundred fifty-six (8%) of the 1,952 HTLV-I/II tests 4. S
trong evidence of inappropriate 238 (70)
ordered in the 17-month period were not ordered with an testing
HIV-1/2 test. In 29 (19%) of the 156 cases, the testing
was likely appropriate, whereas HIV-1/2 testing was likely
Table 2. Clinical Indications for HTLV-I/II Testing*
the intent in the other 127 (81%) cases (Table 3, on page
193). The ordering physician was identifiable in 123 of the Study 1 Study 2
Clinical Information, n (%) n (%) n (%)
127 tests—69 physicians were identified. Thirteen of the
physicians ordered more than one inappropriate HTLV-I/II HIV Risks
test, and 3 physicians inappropriately tested more than 10 Gay/bisexual 10 (2.9) 3 (1.9)
HIV–positive partner 14 (4.1) 6 (3.8)
patients exclusively for HTLV-I/II.
Multiple sexual partners 65 (19.1) 15 (9.6)
HIV–endemic area 10 (2.9) 7 (4.5)
Blinded Seroprevalence Study HIV–positive mother 2 (0.6) 0
Some 97 (76%) of the 127 tests that were ordered inappro Bodily fluid exposure 34 (10.0) 9 (5.8)
priately or likely inappropriately without concurrent Intravenous drug use 8 (2.4) 1 (0.6)
Hemophilia 1 (0.3) 0
HIV-1/2 testing had an adequate residual serum for further Anxiety following nondefined risk 165 (48.5) 63 (40.4)
testing. These samples were blinded and tested for antibodies Symptoms
to HIV-1/2 by enzyme-linked immunosorbent assay—two
Fever 5 (1.5) 3 (1.9)
(2%) of the samples were reactive. Confirmatory testing was Rash 5 (1.5) 3 (1.9)
positive in one case, while there was insufficient residual Adenopathy 3 (0.9) 1 (0.6)
sample in the other for confirmation. Respiratory 4 (1.2) 3 (1.9)
Neurological 7 (2.1) 8 (5.1)
Immune status 3 (0.9) 1 (0.6)
Physician Contact General/other 9 (2.6) 7 (4.5)
Before proceeding with the planned mail-out in March of
Representative Ancillary Testing
2010, a review of more recent HIV-1/2 testing revealed
Hepatitis B or C 207 (60.8) 58 (37.2)
that 25 (20%) of the 127 patients had subsequently been Syphilis 148 (43.5) 40 (25.6)
tested for HIV-1/2 (all negative), supporting the notion Rubella 15 (4.4) 4 (2.6)
that HIV-1/2 testing was the original intention. One Varicella zoster virus 15 (4.4) 2 (1.3)
patient included in the blinded seroprevalence study was HIV-1/2 300 (88.2) —
* Study 1: Random sample of all HTLV-I/II testing requests (fertility/tissue
later identified as having previously received an HIV-1/2
donation screening excluded). Study 2: All HTLV-I/II test requisitions ordered
diagnosis outside the region. The reason for the HTLV-I/ during a 17-month period without concurrent HIV-1/2 testing.
II test was unclear but may have represented the patient’s
physician’s attempt to confirm the HIV diagnosis. The 69 design, improper office practices, and inadequate knowl-
physicians who inappropriately or likely inappropriately edge contributed to the inappropriate tests.
ordered HTLV-I/II testing without concurrent HIV-1/2
testing were contacted by telephone in March 2010 to Testing Requests Postintervention
determine the original intention of the HTLV-I/II test. No complaints or inquiries regarding the redesigned form
They all confirmed that they did not intend to order HTLV have been received in the two years since it was imple-
testing but rather intended HIV testing. The physicians mented. No requests for inappropriate HTLV-I/II tests have
suggested that confusing viral nomenclature, poor form been received on the form.
190
Table 3. Study 2: HTLV-I/II Testing During the Institute of Medicine’s landmark 2000 report To Err
a 17-Month Period Is Human.1,2 In contrast, form design has received little
Tests n (%) attention; it is essentially unregulated, with no international
HTLV-I/II requisitions 1,948 or national directives on how to limit potential for errors.
HIV-1/2 ordered concurrently 1,791 (92) Erroneous laboratory test ordering and interpretation,
HIV-1/2 not ordered concurrently 156 (8.0) which are not easily systematically monitored or assessed,
1. Fertility/tissue donor screening 0 have likely been underrecognized and underreported.2,8,26 In
2. Likely appropriate testing 29 (19) our study, modification of the requisition form and moni-
3. Likely inappropriate testing 27 (17) toring for testing requested on original forms has immedi-
4. Strong evidence of inappropriate 100 (64) ately reduced risks. A broader, forward-looking oversight
testing of the issue is required, particularly given the transition to
* For the 1,952 HTLV-I/II tests ordered, 4 requisitions could not be found for electronic ordering.
Study 2 and therefore could not be evaluated for appropriateness.
Although this study was conducted with paper-based

Requests for HTLV-I/II testing on clinical grounds with ordering, there are significant implications for computer-
noted HIV risk factors continue to be received on the orig- ized provider order entry (CPOE) systems. As the number
inal forms but decreased from 24.8 to 17.3 tests per month of available laboratory tests in clinical medicine continues
seven months after the intervention (p = .008). The 29 to proliferate and the tests become increasingly special-
ordering physicians were contacted promptly, all confirming ized, CPOE systems offer the opportunity to both educate
that they did not intend to order HTLV-I/II testing. physicians for optimal test ordering and to redirect any
erroneous orders. An integrated clinical decision support
Discussion system may be as simple as an automatic alert that HTLV is
Confusing viral nomenclature, exacerbated by an exten- often confused for HIV. We recommend this approach, as
sive test menu and suboptimal form design, led to a high opposed to one that prevents nonspecialists from ordering
number of erroneous HTLV-I/II diagnostic tests, each rare virology tests, which could lead to hazardous delays,
incurring unnecessary costs and having the potential for as learned from experience with medication ordering
patient harm. The potential for serious medical harm was systems.27 CPOE systems have a capacity to bring effi-
evidenced not only by the index case but also by a higher- ciency to our own current system, in which physicians are
than-expected seroprevalence of HIV-1/2 (> 3%) among contacted directly by a laboratory specialist when HTLV
inappropriately ordered HTLV-I/II tests, compared with a is likely to have been ordered in error. Although similar
population prevalence of about 0.1%. In our series, exces- clinical decision support systems have been successful
sive and inappropriate testing for HTLV-I/II on clinical in increasing meaningful use of radiology testing28 and
grounds was widespread—well over 90% of the HTLV-I/II in reducing medication errors,29,30 these clinical decision
tests were ordered erroneously. support systems have not been widely adopted for reducing
erroneous laboratory testing.
Although we could not elucidate the precise contribution
and relationship between the factors leading to the high As the number of recognized pathogens, new names,
error rates, the causation was likely multifactorial. We and available diagnostic tests expands, this increasing
believe that this study reveals problems, as reflected in, for complexity will inevitably lead to pressure on ordering
example, test menu, form design, and practitioner educa- physicians. Our study suggests that a broader approach
tion—which can be addressed on a local-systems level—as to minimize the potential risks of look-alike, sound-
well as the wider problem associated with viral nomencla- alike names in viral nomenclature is needed. Sequential
ture, which can be addressed only at the international level. numbering of new viruses based on scientific grounds opens
up the opportunity for clinical errors. This is a particularly
Health system issues such as optimizing teamwork in pressing issue within the family Retroviridae, given that
surgery5,6 and reducing medical errors from long work HIV- 1 and -2 were originally named HTLV-III12–14 and
hours25 have received considerable attention following -IV15,16 and that recently described human retroviruses are
191
currently named HTLV-323 and HTLV-4.24 HTLV-III and harm may be further amplified or diminished or mitigated
HTLV-IV are outdated terms for HIV-1 and -2, which by testing of menu designs.35
contributed to confusion among some of the practitioners
we interviewed. It is of further concern that the two Conclusions
recently discovered viruses have been named HTLV-3 and Physicians should be aware that test-ordering errors may
HTLV-4. If they are found to be clinically important, then not be infrequent and can lead to adverse patient outcomes.
the potential for confusion and error is significant. The Two concerning areas leading to medical error were
renaming of microbial pathogens such as HTLV-III and identified by our study at a regional reference laboratory.
HTLV-IV to HIV-1 and HIV-2 and the recent recycling of First, the issue of renaming pathogens for precise genetic
the older names (with adjustments) for newly discovered phylogeny and the even more concerning practice of recy-
viruses HTLV-3 and -4 adds precision to nomenclature but cling older names (for example, HTLV-III and the recent
introduces an unnecessary risk into clinical practice. The HTLV-3) can be addressed only at the international level.
risk of look-alike, sound-alike names has been prohibited in Second, test-menu and requisition-form design can either
the naming of pharmaceutical agents, but no such consider- amplify or mitigate the risk of testing errors. Laboratories
ation had been taken yet in laboratory diagnostics. should take explicit action to make test requisitions as clear
as possible to mitigate this risk. A broad assessment of
Currently, the International Committee on Taxonomy physician-ordering errors in the entire field of laboratory
of Viruses, a subdivision of the International Union of diagnostics is a logical but difficult next step in the quest of
Microbiological Societies, is tasked with recognizing and enhancing patient safety.
naming viral taxa. Although they attempt to “avoid or reject
the use of names which might cause error or confusion,” References
their scope does not extend below the rank of species.31 The 1. Institute of Medicine: To Err Is Human: Building a Safer Health
absence of formal oversight below the level of species can System. Washington, DC: National Academy Press, 2000.
lead to confusion (for example, HTLV viruses are viruses 2. Pronovost PJ, Goeschel CA. Time to take health delivery research
of the species Primate T-lymphotropic virus, and therefore seriously. JAMA. 2011 Jul 20;306(3):310–311.
the committee has no influence on the common naming of 3. The Joint Commission. Look-alike, sound-alike drug names.
Jt Comm Perspect. 2001;21(7):10–11.
HTLV viruses).32,33 The continuing exponential growth in
the number of viruses recognized34 increases the possibility 4. Ostini R, et al. Quality use of medicines—Medication safety
issues in naming; look-alike, sound-alike medicine names. Int J
for further confusion between viruses.33
Pharm Pract. Epub 2012 May 12.
5. Pronovost PJ, Freischlag JA. Improving teamwork to reduce
Limitations surgical mortality. JAMA. 2010 Oct 20;304(15):1721–1722.
Our study was restricted to one geographic area and
6. de Vries EN, et al.; SURPASS Collaborative Group. Effect of
therefore may not be fully representative of other areas. a comprehensive surgical safety system on patient outcomes.
However, it is a large and diverse area with urban and rural N Engl J Med. 2010 Nov 11;363(20):1928–1937.
representation and included both outpatient and inpatient 7. Graber M. Diagnostic errors in medicine: A case of neglect.
samples from all types of health care settings, thus miti- Jt Comm J Qual Patient Saf. 2005;31(2):106–113.
gating the risk of sample bias. We determined appropriate- 8. Plebani M. The detection and prevention of errors in laboratory
ness of HTLV-I/II testing on the basis of a limited number medicine. Ann Clin Biochem. 2010;47(Pt 2):101–110.
of risk factors, symptoms, and ancillary tests, which may 9. Schiff GD, et al. Diagnostic error in medicine: Analysis of
have led to under- or overestimates of inappropriate 583 physician-reported errors. Arch Intern Med. 2009 Nov
testing. However, corroborating evidence from our tele- 9;169(20):1881–1887.
phone survey showing 100% specificity for inappropriate 10. Gandhi TK, et al. Missed and delayed diagnoses in the
testing suggests that, if anything, we present a conservative ambulatory setting: A study of closed malpractice claims. Ann
Intern Med. 2006 Oct 3;145(7): 488–496.
estimate of the number of inappropriate HTLV-I/II tests.
Moreover, although the requisitions were paper based, we 11. Fauquet CM, Fargette D. International Committee on Taxonomy
of Viruses and the 3,142 unassigned species. Virol J. 2005 Aug
believe that the risk factors identified are directly transfer-
16;2:64.
able to electronic health records, where the risk of patient
192
12. Barré-Sinoussi F, et al. Isolation of a T-lymphotropic retrovirus 25. Volpp KG, Landrigan CP. Building physician work hour
from a patient at risk for acquired immune deficiency syndrome regulations from first principles and best evidence. JAMA. 2008
(AIDS). Science. 1983 May 20;220(4599):868–871. Sep 10;300(10):1197–1199.
13. Popovic M, et al. Detection, isolation, and continuous 26. Newman-Toker DE, Pronovost PJ. Diagnostic errors—
production of cytopathic retroviruses (HTLV-III) from The next frontier for patient safety. JAMA. 2009 Mar
patients with AIDS and pre-AIDS. Science. 1984 May 11;301(10):1060–1062.
4;224(4648):497–500. 27. Strom BL, et al. Unintended effects of a computerized physician
14. Coffin J, et al. What to call the AIDS virus? Nature. 1986 May order entry nearly hard-stop alert to prevent a drug interaction:
1–7; 321(6065):310. A randomized controlled trial. Arch Intern Med. 2010 Sep
15. Kanki PJ, et al. New human T-lymphotropic retrovirus related to 27;170(17):1578–1583.
simian T-lymphotropic virus type III (STLV-IIIAGM). Science. 28. Ip IK, et al. Adoption and meaningful use of computerized
1986 Apr 11;232(4747):238–243. physician order entry with an integrated clinical decision support
16. Clavel F, et al. Isolation of a new human retrovirus from system for radiology: Ten-year analysis in an urban teaching
West African patients with AIDS. Science. 1986 Jul hospital. J Am Coll Radiol. 2012;9(2):129–136.
18;233(4761):343–346. 29. Kaushal R, Shojania KG, Bates DW. Effects of computerized
17. Clavel F, et al. Human immunodeficiency virus type 2 infection physician order entry and clinical decision support systems on
associated with AIDS in West Africa. N Engl J Med. 1987 May medication safety: A systematic review. Arch Intern Med. 2003
7;316(19):1180–1185. Jun 23;163(12):1409–1416.
18. Gonçalves DU, et al. Epidemiology, treatment, and prevention 30. Milani RV, Oleck SA, Lavie CJ. Medication errors in patients
of human T-cell leukemia virus type 1-associated diseases. Clin with severe chronic kidney disease and acute coronary syndrome:
Microbiol Rev. 2010;23(3):577–589. The impact of computer-assisted decision support. Mayo Clin
19. Proietti FA, et al. Global epidemiology of HTLV-I infection and Proc. 2011;86(12):1161–1164.
associated diseases. Oncogene. 2005 Sep 5;24(39):6058–6068. 31. International Committee on Taxonomy of Viruses. The
20. Einsiedel L, et al. Bronchiectasis is associated with human International Code of Virus Classification and Nomenclature.
T-lymphotropic virus 1 infection in an Indigenous Australian Aug 2002. Accessed Sep 27, 2012. http://ictvonline.org
population. Clin Infect Dis. 2012 Jan 1;54(1):43–50. /codeOfVirusClassification_2002.asp.
21. de Oliveira Mde F, et al. Infective dermatitis associated 32. International Committee on Taxonomy of Viruses. ICTV Master
with human T-cell lymphotropic virus type 1: Evaluation Species List 2009, ver. 10. Aug 24, 2011. Accessed Sep 27, 2012.
of 42 cases observed in Bahia, Brazil. Clin Infect Dis. http://talk.ictvonline.org/files/ictv_documents/m/msl/1231.aspx.
2012;54(12):1714-1719. 33. Kuhn JH, Jahrling PB. Clarification and guidance on the
22. Araujo A, Hall WW. Human T-lymphotropic virus type II and proper usage of virus and virus species names. Arch Virol.
neurological disease. Ann Neurol. 2004;56(1):10–19. 2010;155(4):445–453.
23. Calattini S, et al. Discovery of a new human T-cell lymphotropic 34. Brister JR, et al. Towards viral genome annotation standards:
virus (HTLV-3) in Central Africa. Retrovirology. 2005 May Report from the 2010 NCBI Annotation Workshop. Viruses.
9;2:30. 2010;2(10):2258–2258.
24. Wolfe ND, et al. Emergence of unique primate T-lymphotropic 35. Radecki RP, Sittig DF. Application of electronic health records
viruses among central African bushmeat hunters. Proc Natl Acad to the Joint Commission’s 2011 National Patient Safety Goals.
Sci U S A. 2005 May 31;102(22):7994–7999. JAMA. 2011 Jul 6;306(1):92–93.
Source: Siemieniuk RAC, Fonseca K, Gill MJ. Using root cause analysis and form redesign to reduce incorrect ordering of HIV tests. Jt Comm J Qual Patient Saf.
2012 Nov; 38(11):506.
193
CASE STUDY THREE

Counting Matters: Lessons from the Root Article-at-a-Glance
Cause Analysis of a Retained Surgical Item
Background: Retained surgical items (RSIs), such
as a sponge, instrument, or needle, after a surgery or
Abha Agrawal, MD, FACP
invasive procedure is an uncommon but potentially
A retained surgical item (RSI)—a sponge, instrument, serious event associated with significant morbidity
needle, or other item inadvertently left after a surgery and mortality. A 27-year-old woman was discov-
or invasive procedure—is an uncommon but potentially ered to have a retained vaginal sponge a week after
serious event associated with significant morbidity and she underwent the repair of a vaginal tear following
mortality for the patient1 and malpractice risk for the normal vaginal delivery. The retained sponge was
provider and the institution.2 In addition, a “foreign object removed with no further complications.
retained after surgery” is one of 10 health care–acquired
conditions for which the Centers for Medicare & Medicaid Root Cause Analysis: The fundamental error
Services (CMS) does not provide hospitals with addi- involved the obstetric team’s failure to perform the
tional payment.3 RSI is subject to voluntary reporting as standard protocol of counting sponges before, as well
a sentinel event to The Joint Commission4 and manda- as after, the procedure. This was attributed to a lack
tory reporting to the New York State’s incident reporting of reminders to perform the count, relatively recent
system (New York Patient Occurrence and Tracking System implementation of the sponge-count policy, and a
[NYPORTS]).5 breakdown in teamwork and communication between
physicians and nurses.
RSIs are uncommon but represent a persistent surgical
complication. According to a 2003 report by Gawande et Corrective Actions: The corrective actions
al., RSIs are estimated to occur in 1 of every 1,000 to 1,500 focused on systems improvement, as opposed to
abdominal operations.1 the human error of the memory lapse. The sponge-
counting process was reinforced by incorporating a
In the report, of the 61 RSIs (detected in 54 cases), sign-out at the end of obstetric procedures to ensure
69% involved sponges, and 31% involved instruments; that the counts have been done and any discrepancies
54% of the RSIs were left in the abdomen or pelvis; addressed. A specialized delivery note with mandatory
22% in the vagina; 7.4% in the thorax; and 17% elsewhere, field to document sponge count was implemented in
including the spinal canal, face, brain, and extremities. the electronic health record as an additional reminder.
Retained sponges are more common than instruments All staff participated in a teamwork and communica-
because of their small size, frequent use, and the ability to tion training program.
mimic intraabdominal contents when saturated in blood.
The Minnesota Adverse Health Events Reporting System Tracking Compliance: Since the incident’s
revealed that a quarter of all 161 RSIs between 2003 and occurrence in 2010, the staff has demonstrated 100%
2008 occurred during obstetrical procedures, with nearly all compliance with the corrective actions, and a retained
of those cases involving vaginal deliveries.6 surgical item complication has not recurred.
RSIs can cause serious clinical complications, such as small- Conclusion: Individual accountability must be
bowel fistulae, obstruction, visceral perforation, sepsis, or balanced with systems improvement, given that most
even death. They often lead to reoperation for removal of medical errors are a result of fallible humans working
the object and management of complications.1,7 In addition in chaotic unpredictable, and complex clinical
to clinical complications, RSIs frequently lead to malprac- environment.
tice lawsuits. In one report, 87% cases of RSI resulted in
malpractice litigation, with an average claim of $52,581.1
194
Another review of 40 cases of RSIs found that the average examination was normal except for mild lower abdominal
defense costs were $572,079 and the total indemnity tenderness. She had a white blood cell count of 11,830
payments were more than $2 million.2 RSI malpractice per mm3 with 76% neutrophils, hemoglobin of 13.5 gm
claims often fall under the doctrine of Res ipsa loquitur per dL, and hematocrit of 40.6%. On sterile speculum
(“the thing speaks for itself ”), implying that the presence of examination, she was found to have a retained vaginal
RSI itself is proof of negligence on the part of the defendant sponge, which was removed with ring forceps. There
(the surgeon). In addition, RSIs drive up the cost of care was no active bleeding and no clots. She was discharged
and often attract critical media coverage implying proven home on oral antibiotics for one week, and recovered fully
negligence or recklessness on the part of the surgeon.8 without complications. The attending obstetrician made
Regardless of the clinical outcome, this is a psychologi- a full disclosure of the error to the patient, who appeared
cally devastating complication both for patients, because it comfortable with and accepting of the disclosure. To date,
erodes their trust in our commitment to care for them, and no further action has been taken by the patient, and she
for caregivers, in that it evokes a sense of personal failure, continues to follow up in the outpatient clinic.
guilt, and embarrassment.9
Methods
This article describes the root cause analysis (RCA) and Setting
lessons learned from the case of an unintentionally retained The RCA for this case was conducted at a large academic,
vaginal sponge in a patient following normal vaginal urban, tertiary care public hospital that performs about
delivery. RCA is a structured method to analyze serious 2,600 deliveries annually, in addition to approximately
adverse events. Initially developed to analyze industrial acci- 27,000 admissions, 130,000 emergency room visits, and
dents, RCA is now widely deployed as an error analysis tool 750,000 outpatient visits annually. The labor and delivery
in health care.10 A central tenet of RCA is to identify under- (L&D) area in the hospital has nine rooms, and the staff
lying systems problems that increase the likelihood of errors includes 19 obstetricians, 8 certified nurse midwives,
(called “latent errors”) while avoiding the trap of focusing 24 house officers, and 40 nurses.
on mistakes by individuals (called “sharp-end errors”).11
Root Cause Analysis
This article provides a clinical summary of the case; the Process. The RCA process at this institution is managed by
purpose and methodology of the RCA process; root cause the risk management department (RM). Following report
findings; and corrective actions, including follow-up. of an adverse event, which is usually made by the clinical
department where the incidence occurred, the RM director
Case Summary determines whether the event is reportable to NYPORTS
A 27-year-old woman presented to the hospital at 40 weeks and whether an RCA needs to be performed. NYPORTS
of gestation with the complaints of lower abdominal pain. has a mandatory requirement to conduct RCAs on some
She had normal vital signs and category 1 fetal heart events, such as unexpected death or RSI.5
tracing. Labor was induced using vaginal misoprostol, and
she delivered a healthy boy by normal vaginal delivery. For each RCA, RM convenes a work group, called an RCA
She sustained a readily identifiable right vaginal sulcus Committee, which includes members from the key clin-
tear during the delivery, which was repaired. The patient ical areas involved in the case. In this case, the committee
was monitored in the hospital for two days and discharged included the chief of obstetrics and gynecology, another
home in good condition. senior obstetric attending, the director of nursing for labor
and delivery, and a staff nurse, as well as leadership from
She returned to the hospital a week later with lower abdom- perioperative services. The committee also included the
inal pain and foul discharge from the vagina. On triage, the RM director; representatives from patient safety and quality
blood pressure was 145/88 mm Hg, the heart rate 94 beats management; and senior members of hospital administra-
per minute, and the temperature 98.7 degrees F (37.06°C). tion, such as the chief medical officer (CMO [A.A.]). The
She was noted to have dark brown, malodorous vaginal hospital policy does not permit the inclusion of clinicians
discharge, as well as pelvic tenderness. The abdominal involved in the adverse event in the RCA meetings. A senior
195
physician is asked to chair the committee; for this case, the be done to prevent it from happening again? Figure 1
chief of orthopedic surgery served as chair. The director of (below) provides a schematic diagram of the three ques-
RM facilitates the meetings, and a staff member from RM tions and how they were applied to this case. The RCA
documents the discussion and minutes. starts with clinical narrative followed by a discussion of the
root cause(s). Participants are provided a copy of the locally
RM prepares a detailed clinical summary of the case, which developed “rules of conduct” that specifically prohibit the
is distributed to RCA participants in advance for review. In “name, blame, and shame” approach. The “Just Culture”
addition, the electronic health record (EHR), as well as the algorithm is frequently used during the discussion.12 The
supplemental paper chart, is available for access to clinical discussion of “Why did it happen?” involves digging deeper
information to facilitate the discussion. by repeatedly asking “Why?” until no further causes can
be identified. This allows the group to focus on root causes
The RCA process is designed to answer three basic ques- instead of proximal causes and permits discovery of poten-
tions: What happened, why did it happen, and what can tial systems vulnerability for improvement.
Figure 1. Schematic Diagram of the Root Cause Analysis (RCA) Process
What happened? A sponge was left behind inadvertently after

repair of a vaginal tear postdelivery.
Why did it Physician and nurse “forgot” to do the count:

happen? Human fallibility.
Why?
They failed to follow the mandatory

sponge-count policy.
Why?
- No reminders to do sponge count

- Breakdown in teamwork and communication
between nurse and physician
- Relatively recent implementation of the
policy about one year ago
What can be - Reminders

done to prevent - Time -out checklist and procedure at
it from the end of the procedure
happening - Mandatory reminder in postdelivery
again? note
- TeamSTEPPS™ training for L&D staff
- Continuing education sessions for staff to
reinforce policy
The RCA process is designed to answer three basic questions: What happened, why did it happen, and what can be
done to prevent it from happening again? L&D, labor and delivery.
196
At the end of the RCA, the findings are summarized, and a On further analysis, this error of omission was attributed
standard of care determination is made with the following to four root causes, as illustrated in Figure 2 (below) and as
options in accordance with the NYPORTS requirements: listed as follows:
(a) standard of care was met, (b) standard of care was met 1. Although the standard practice of performing sponge
with room for improvement, (c) standard of care was not counts before and after a procedure has been widely
met attributable to systems, (d) standard of care was not adopted in the operating room (OR) for many years, it
met attributable to an individual practitioner. Options was only implemented in L&D about one year before
“c” and “d” can be combined. The standard of care deter- the incident. Staff was provided in-services on the new
mination is made by carefully following the just culture policy, but no formal monitoring was in place to ensure
approach of balancing no blame with accountability. compliance with the sponge-counting protocol after
normal vaginal deliveries. In a busy clinical area with
Findings. The RCA elucidated the fundamental error in multiple rotating house officers, it was likely that the
this case to be the failure of the obstetric team, including recently implemented sponge counting had not yet
the physician and the nurse, to perform the standard become an automated part of L&D staff ’s work flow,
protocol of counting sponges before, as well as after, the leading to this RSI case.
repair of the vaginal tear. This type of error, called slips, 2. The L&D is usually a busy area with multiple
is described by Wachter as “. . . inadvertent, unconscious distractions; the presence of a newborn postdelivery
lapses in the performance of some automatic task,” which, unit was a particular source of distraction for the
he states,” probably represent the greater overall threat to nurse involved.
patient safety.”13(p. 20) In this case, there were no obvious 3. There was no step downstream in the work flow to
extraneous factors to facilitate the memory lapse such as remind the team to do the sponge count if team
unexpected emergencies, other sick patients, understaffing, members forgot to do it at the end of the procedure.
or excessive work hours. The team included a skilled nurse In the absence of a reminder or a forcing function,
with years of experience in deliveries; a chief resident this step remained dependent on imperfect human
with excellent track record who was about to graduate as memory and therefore susceptible to failure. In the
an independent physician; and an attending physician, OR, unlike in L&D, sponge count is a part of the
also with an impeccable record of safety, who was closely comprehensive World Health Organization (WHO)
supervising the team. Surgical Safety Checklist,14 which includes a sign-out
Figure 2. Root Causes in the Case of a Postdelivery Retained Vaginal Sponge
Policy or Human Environment/

Procedure Resources/ Equipment/
People Supplies
Lack of visual Addition of a Failed human

reminders newborn baby memory None
Busy clinical
Recently implemented
environment
policy for 1 yr
Failure to follow Retained Vaginal
sponge count policy Sponge
Hierarchical boundaries Breakdown of
teamwork among None None
Culture of poor physicians and nurses
communication
Information Leadership: Miscellaneous

Management and Corporate Factors
Communication Culture
In the root cause analysis, the error of omission was attributed to a number of root causes.
197
at the end of each procedure. The sign-out is held jointly tasks are completed.15 Our institution has implemented
by the surgeon and the nurse, and works as a final an enterprisewide EHR, which is widely adopted
check to ensure that everything has been counted and across all areas of the hospital. Our own experience
discrepancies addressed. has demonstrated the positive role of computer-based
4. Finally, because of a breakdown in the teamwork and reminders in improving adherence with specific clinical
communication between the physician and the nurse, tasks, such as prescribing appropriate medications for
they did not check on each other to prevent this error. congestive heart failure and performance of medication
A culture of hierarchy and resistance to escalation among reconciliation on admitted patients.16,17
clinical team members further exacerbated the lack of 3. The literature on RSIs suggests that a failure of
communication. communication among team members is the most
frequent contributor to the event.18,19 Therefore, to
Following an extensive discussion of the root causes, improve teamwork and communication, all L&D staff
the RCA Committee concluded that the standard of care members, including physicians, nurses, nurse midwives,
was not met and was attributable to various systems and clerks, were required to participate in a structured
vulnerabilities, as described above. team-building program. Because the institution
had already been implementing one such program,
Corrective Actions TeamSTEPPS® (Team Strategies and Tools to Enhance
Because the fundamental error leading to RSI was an Performance and Patient Safety)20 across all departments,
inadvertent omission of the already existent sponge-count we decided to use TeamSTEPPS as the training tool to
protocol by the L&D staff, the corrective actions, including improve teamwork and communication in L&D as well.
visual and electronic reminders, focused around strategies
to reduce the probability of such omission in the future. TeamSTEPPS, a health care version of the aviation indus-
The corrective actions were as follows: try’s crew resource management program, was developed
1. The sponge count sheet used in L&D was modified by the US Department of Defense and the Agency for
to include the documentation of a sign-out process to Healthcare Research and Quality.20 The TeamSTEPPS
verify that a sponge count has been performed correctly framework is based on teaching four core skills—leadership,
and any discrepancies have been reconciled (Figure 3, communication, situation monitoring, and mutual support,
page 199). This sign-out is jointly performed at the end which are all intended to lead to desired outcomes in knowl
of the procedure by the surgeon and the nurse. The edge, attitudes, and performance (Figure 5, on page 200).
documentation of sign-out on the sponge-count sheet
serves as a reminder and should help minimize omission Tracking Compliance with Corrective Actions. The RM
of the sponge-counting protocol. Of note, the sign-out department compiles a list of corrective actions for all
is already a part of the WHO Surgical Safety Checklist RCAs done at the hospital and distributes it to various
protocol in the ORs14; incorporating it in the L&D departments for action. The clinical leadership of the
work flow should also improve standardization of the department—the chief of service and the associate director
work across clinical areas. It is noteworthy that we have of nursing—is responsible for implementing the corrective
not had a case of RSI in ORs since the case in 2010. actions and reporting the status quarterly at the hospital’s
2. The postdelivery note in the EHR was modified to quality council meeting, chaired by the CMO. RM presents
include a mandatory field to document the process a compiled report to the executive quality council, chaired
and accuracy of the sponge/instrument count. A by the CEO, and is responsible for ensuring compliance
screenshot of the delivery note in the EHR is provided with corrective actions.
in Figure 4 (page 200). Although the EHR reminder
will not prevent the retained sponge complication, it Results
will alert the team of a potential omission sooner— For this case, the outcome measure to evaluate the effec-
that is, before the patient is discharged. There is strong tiveness of corrective actions was the occurrence of an RSI
evidence that computerized reminders and forcing in L&D—and we have not had any occurrence of an RSI
functions are effective in ensuring that specific care in the department of obstetrics since the implementation
198
Figure 3. Sponge Count Sheet
DATE: WARD/LOCATION:
BIRTHING ROOM-LABOR & DELIVERY
VAGINAL DELIVERY COUNT RECORD M.R. # D.O.B.
NAME:
ACCOUNT NUMBER:
ITEM BEFORE FINAL COUNT

PROCEDURE (AFTER PROCEDURE)
DETECTOS (4 X 4)
ADDED
TAMPONS
ADDED
LAP RINGS
ADDED
NEEDLES
ADDED
Pre-procedure Count ________________________ AND ________________________________

Date/Time: _________ Signature/Title Signature/Title
Relief Count ________________________ AND ________________________________

Time: ________ Signature/Title Signature/Title
Final Count ________________________ AND ________________________________

Time: ________ Signature/Title Signature/Title
Correct:  YES  NO → If NO complete additional information

Notified: ____________________________ Time: ____________________
X-Ray Performed:  YES  NO Time: ____________________
Outcome: ______________________________________________________________________________
______________________________________________________________________________
Vaginal Packing:  YES  NO Type of Packing: ______________________________
Date to be removed: ___________________ ______________________________

Signature/Title
Date packing removed: __________________ By: _____________________________________
Physician Signature: __________________________ Date/Time: _____________________________________
RN Signature: ______________________________ Date/Time: _____________________________________
The sponge-count sheet used in labor and delivery (L&D) was modified to include the documentation of a sign-out
process to verify that a sponge count has been performed correctly and any discrepancies have been reconciled. M.R.,
medical record; D.O.B., date of birth.
199
Figure 4. Screenshot of the Postdelivery Note
The postdelivery note in the electronic health record, as shown in this screenshot, includes a mandatory field to
document the process and accuracy of the sponge/instrument count. Left panel: Structured fields in the note;
right panel: Details of the sponge-count field.
of the corrective actions. Because an RSI is a rare event, the

Figure 5. Framework of the Team
following two process indicators are tracked as proxies for
Strategies and Tools to Enhance
the effectiveness of the interventions:
Performance and Patient Safety
Random sampling of 45 charts per month from the (TeamSTEPPS® Program)
L&D area to determine compliance with the sign-out
and sponge-counting process. A sample monthly tracking
form is provided in Figure 6 (page 201). The percentage
of compliance with the process is tracked and reported
on a monthly basis and has been sustained at 100% since
January 2011.
Participation of 100% of L&D staff in the TeamSTEPPS

training program. This goal was achieved in May
2011 within six months of the RCA—and has been
maintained since.
The TeamSTEPPS framework is based on teaching
Discussion four core skills— leadership, communication, situation
Lessons learned from adverse events are disseminated monitoring, and mutual support, which are all intended
widely throughout the hospital at departmental meetings. to lead to desired outcomes in knowledge, attitudes,
and performance.
In addition, selected cases are discussed at the monthly
Source: Agency for Healthcare Research and Quality (AHRQ).
case-based patient safety grand rounds. The format of the
TeamSTEPPS Instructor Guide. Rockville, MD: AHRQ, 2006. AHRQ
grand rounds includes a clinical presentation of the case Publication No. 06-0020. Used with permission.
200
Figure 6. Sample Monthly Tracking Form to Audit Compliance with Corrective Actions
followed by a discussion of the key lessons learned from not all reports of RSIs have been associated with these
the adverse outcome. The event is open to all staff and risk factors.24
is widely attended by nurses, physicians, social workers,
respiratory therapists, and administrators. As suggested in Limitations of Manual Sponge-Counting
the literature, case-based learning is a valuable educational Process and Emerging Technology-Based
tool; sharing an actual clinical situation stimulates much Solutions
greater clinician interest than discussing a generic patient Although ORs throughout the United States have widely
safety topic.21,22 adopted the protocol, sponge counting is frequently
omitted after obstetrical procedures.1,2,25 In addition, as a
Strategies that were implemented in January 2011 on the manual practice, counting of sponges and other surgical
basis of the findings of the RCA described in this article items is not only labor intensive but fraught with vulnera-
have helped us achieve sustained and perfect compliance bilities.26,27 Cima et al. concluded that the manual counting
with the corrective actions and no recurrence of an RSI. of surgical items is an inherently error-prone process and
Three factors have been found to be significantly associ- that a new technology was needed to break beyond this
ated with an increased risk of RSI: emergency procedure, “performance boundary.”28(p. 131) Two technological adjuncts
unplanned change in the procedure, and a high body-mass to the counting process have emerged that may hold the
index (BMI).1 In a meta-analysis, Cima et al. concluded promise to make it more reliable in preventing RSI—bar
that the following surgical procedure characteristics and coding (or data-matrix coding) of sponges and sponges
clinical variables were associated with a significantly embedded with radiofrequency identification (RFID).
increased risk for RSI and hence should require greater vigi-
lance: complex surgical procedures, damage control (open In a randomized controlled trial in 300 general surgery
abdominal procedures and staged and abbreviated proce- operations, Greenberg et al. demonstrated that a bar-coded
dures), emergency surgery, increased BMI, involvement of sponge system detected significantly more counting
more than one surgical team, procedures that involve more discrepancies than the traditional protocol. However, the
than one body cavity (such as thoracoabdominal proce- system introduced new technical difficulties and increased
dures and trauma), prolonged surgery, unexpected change the time spent counting sponges.29 An 18-month use
in the course of a surgical procedure, and use of unusually of a data-matrix–coded sponge system eliminated RSIs
large number of instruments.23 It is noteworthy that the from a high-volume surgical practice, as opposed to the
RSI occurrence in this case was not associated with any finding of a retained sponge occurring an average of every
of these risk factors; indeed, the literature suggests that 64 days when the former manual sponge-counting system
201
was used.30 Early experiments with RFID yielded encour- against the almost inevitable memory lapses is to improve
aging results.31 It is important to note that cost can be our system by creating memory aids such as checklists and
an important consideration in the adoption of these new read-backs, standardizing and simplifying routine tasks,
technological systems. improving teamwork among various care providers so that
they can correct each other when they observe memory
Teamwork and Communication lapses, and most importantly, by learning from our errors
Better communication and teamwork among the OR team, when they do occur.
including surgeons, assistants, and nurses, is an important
component of reducing surgical errors.32 The literature Conclusion
suggests that is critical to not only provide training but also The case of an unintentionally retained vaginal sponge
ensure that the training is continuously reinforced until the provides a vehicle for discussion of a balanced approach
teamwork is fully integrated in the work flow and culture of to analysis of surgical complication and formulation of
the organization.33 strategies to prevent a recurrence of the event. This case
also underscores the limitations of system improvement.
Human Error and Accountability No matter how rigorous the corrective actions, such as
The path of least resistance in our RCA would have been additional reminders after the already “mandatory” sponge-
to attribute the complication to an error by the nurse or count process, there is no guarantee that someday somehow
the physician and pursue punitive actions. However, with another team won’t forget to perform these new steps. Even
the progress in patient safety during the last decade and a with the application of technology, many vital tasks in clin-
greater understanding of the concept of a just culture—or, ical medicine will remain dependent on humans and their
a “safety culture”34—we recognize that expecting humans inherent fallibility.
to be perfect and therefore have no memory lapses will not
solve the problem of medical errors. Instead, we must fight References
the instinct to blame individuals and replace it with a focus 1. Gawande A, et al. Risk factors for retained instruments and
on identifying and improving systems flaws that allow the sponges after surgery. N Engl J Med. 2003 Jan 16;348(3):229–
235. Accessed Oct 22, 2012. http://www.nejm.org/doi/full
inevitable human error to occur and cause harm.
/10.1056/NEJMsa021721.
2. Kaiser CW, et al. The retained surgical sponge. Ann Surg.
As Reason observed, “Human fallibility, like gravity,
1996;224(1):79–84.
weather and terrain, is just another foreseeable
3. Centers for Medicare & Medicaid Services. Hospital-Acquired
hazard.”10(p. 25) Now, a safety culture does not mean a lack
Conditions (Present on Admission Indicator). (Updated Sep 20,
of accountability; on the contrary, “balancing systems and 2012.). Accessed Oct 22, 2012. http://www.cms.gov/Medicare
individual accountability is . . . central to the operation of /Medicare-Fee-for-Service-Payment/HospitalAcqCond/index
a safety culture.”34(p. 1) It provides a framework to distin- .html?redirect=/HospitalAcqCond/.
guish between “human error” (inevitable, and managed 4. The Joint Commission. Sentinel Event Policy and Procedures. Jan
through systems change), “at-risk behavior” (such as 1, 2012. Accessed Oct 22, 2012. http://www.jointcommission
work-arounds—managed by understanding and fixing the .org/Sentinel_Event_Policy_and_Procedures/.
systems factors that promote such behavior), and “reck- 5. New York State Department of Health. The New York Patient
less behavior (“the conscious disregard of a substantial Occurrence and Tracking System (NYPORTS): Annual Reports.
and unjustifiable risk”).35(p. 43) A surgeon refusing to sign Accessed Oct 22, 2012. http://www.health.ny.gov/nysdoh
/hospital/nyports/.
the operative site or to participate in time-out process will
6. Minnesota Department of Health. Spotlight on Patient Safety.
be considered reckless behavior and is worthy of punitive
Retained Foreign Objects. Apr 2009. Accessed Oct 22, 2012.
action. Fortunately, such instances are rare, and most errors
http://www.health.state.mn.us/patientsafety/publications
fall into the category of human error or at-risk behavior. /newsletter0409_rfo.pdf.
7. Gonzalez-Ojeda A, et al. Retained foreign bodies following
In the case described in this article, it was concluded that intraabdominal surgery. Hepatogastroenterology. 1999;46(26):
in the absence of a behavioral pattern, the problem of the 808–812.
memory lapse was a typical human error. Our best defense
202
8. New York Times. The Biggest Mistake of Their Lives. Burton S. 22. Piotrowski MM, Saint S, Hinshaw DB. The Safety Case
Mar 16, 2003. Accessed Oct 22, 2012. http://www.nytimes Management Committee: Expanding the avenues for addressing
.com/2003/03/16/magazine/the-biggest-mistake-of-their-lives patient safety. Jt Comm J Qual Improv. 2002;28(6):296–305.
.html?pagewanted=all&src=pm. 23. Cima RR, et al. Incidence and characteristics of potential and
9. Wu AW. Medical error: The second victim: The doctor who actual retained foreign object events in surgical patients. J Am
makes the mistake needs help too. BMJ. 2000 Mar 18; Coll Surg. 2008;207(1):80–87.
320(7237):726–727. 24. Wang CF, et al. Risk factors for retained surgical foreign bodies:
10. Reason J. Managing the Risks of Organizational Accidents. A metaanalysis. OPUS 12 Scientist. 2009;3(2):21–27.
Burlington, VT: Ashgate, 1997. 25. Agency for Healthcare Research and Quality WebM&M. Did we
11. Wu AW, Lipshutz AK, Pronovost PJ. Effectiveness and forget something? Gibbs C. Sep 2003. Accessed Oct 22, 2012.
efficiency of root cause analysis in medicine. JAMA. 2008 Feb. http://www.webmm.ahrq.gov/case.aspx?caseID=27.
13;299(6):685–687. 26. Beyea SC. Counting instruments and sponges. AORN J.
12. Marx D. Patient Safety and the “Just Culture”: A Primer for 2003;78(2):290, 293–294.
Health Care Executives. New York City: Columbia University, 27. Christian CK, et al. A prospective study of patient safety in the
2001. Accessed Oct 22, 2012. http://psnet.ahrq.gov/resource operating room. Surgery. 2006;139(2):159–173.
.aspx?resourceID=1582. 28. Cima RR, et al. A multidisciplinary team approach to retained
13. Wachter RM. Understanding Patient Safety. New York City: foreign objects. Jt Comm J Qual Patient Saf. 2009;35(3):
McGraw-Hill Medical, 2008. 123–132.
14. World Health Organization. WHO Surgical Safety Checklist and 29. Greenberg CC, et al. Bar-coding surgical sponges to
Implementation Manual. World Alliance for Patient Safety. 2008. improve safety: A randomized controlled trial. Ann Surg.
Accessed Oct 22, 2012. http://www.who.int/patientsafety 2008;247(4):612–616.
/safesurgery/ss_checklist/en/index.html. 30. Cima RR, et al. Using a data-matrix-coded sponge counting
15. Kawamoto K, et al. Improving clinical practice using clinical system across a surgical practice: Impact after 18 months.
decision support systems: A systematic review of trials to identify Jt Comm J Qual Patient Saf. 2011;37(2):51–58.
features critical to success. BMJ. 2005 Apr 2;330(7494):765. 31. Macario A, Morris D, Morris S. Initial clinical evaluation of a
16. Agrawal A. Medication errors: Prevention using information handheld device for detecting retained surgical gauze sponges
technology systems. Br J Clin Pharmacol. 2009;67(6):681–686. using radiofrequency identification technology. Arch Surg.
17. Agrawal A, Wu WY. Reducing medication errors and improving 2006;141(7):659–662.
systems reliability using an electronic medication reconciliation 32. Catchpole K, et al. Teamwork and error in the operating room:
system. Jt Comm J Qual Patient Saf. 2009;35(2):106–114. Analysis of skills and roles. Ann Surg. 2008;247(4):699–706.
18. Leonard M, Graham S, Bonacum D. The human factor: The 33. Neily J, et al. Association between implementation of a medical
critical importance of effective teamwork and communication team training program and surgical mortality. JAMA. 2010 Oct
in providing safe care. Qual Saf Health Care. 2004;13 Suppl 20;304(15):1693–1700.
1:i85–90. 34. Hickson GB, et al. Balancing systems and individual
19. Lingard L, et al. Communication failures in the operating room: accountability in a safety culture. In From Front Office to Front
An observational classification of recurrent types and effects. Qual Line: Essential Issues for Health Care Leaders, 2nd ed. Oak Brook,
Saf Health Care. 2004;13(5):330–334. IL: Joint Commission Resources, 2011, 1–35.
20. Agency for Healthcare Rearch and Policy. TeamSTEPPS®: 35. Marx D. Whack-a-Mole: The Price We Pay for Expecting Perfection.
National Implementation. Accessed Oct 22, 2012. http:// Plano, TX: By Your Side Studios, 2009.
teamstepps.ahrq.gov/.
21. Wachter RM, et al. Learning from our mistakes. Ann Intern Med.
2002 Jun 4;136(11):850–852.
Source: Agrawal A. Counting matters: Lessons from the root cause analysis of a retained surgical item. Jt Comm J Qual Pat Saf. 2012 Dec;38(12):566.
203
204
Glossary
action plan The product of the root cause analysis that similar activities in the organization to find and implement
identifies the strategies that an organization intends to ways to improve it. This is one of the foundations of both
implement to reduce the risk of similar events occurring total quality management and continuous quality improve-
in the future. The plan should address responsibility for ment. Internal benchmarking occurs when similar processes
implementation, oversight, pilot testing as appropriate, within the same organization are compared. Competitive
time lines, and strategies for measuring the effectiveness of benchmarking occurs when an organization’s processes
the actions. are compared with best practices within the industry.
Functional benchmarking refers to benchmarking a similar
adverse drug event An injury resulting from medical inter- function or process, such as scheduling, in another industry.
vention related to a medication, including harm from an
adverse drug reaction or a medication error. brainstorming The process of capturing people’s ideas,
without censoring or editing them, and organizing those
adverse drug reaction A response to a medicinal product thoughts around common themes.
that is noxious and unintended, and that occurs at doses
normally used in humans for the prophylaxis, diagnosis, or capability chart An analytical tool that uses upper and
treatment of disease or for the restoration, correction, or lower parameters for acceptable performance of tasks or
modification of physiological function. processes in order to determine whether a given change in
the process is capable of reducing variation in performance.
adverse event A patient safety event that resulted in harm
to a patient. care The provision of accommodations, comfort, and treat-
ment to an individual, implying responsibility for safety,
affinity diagram An illustrative tool used to organize a including care, treatment, service, habilitation, rehabilita-
large volume of ideas or data into meaningful groups with tion, or other programs instituted by the organization for
logical connections. the individual served.
aggregate To combine standardized data and information. change analysis A study of the differences between the
expected and actual performance of a process. Change
aggregate data (measurement data) Measurement data analysis involves determining the root causes of an event by
collected and reported by organizations as a sum or total examining the effects of change and identifying causes.
over a given time period (for example, monthly, quarterly)
or for certain groupings (for example, health care organiza- change management A tool used to encourage organiza-
tion level). tional acceptance of a change in process within a system.
benchmarking Continuous measurement of a process, check sheet A form used to sort and group data, using
product, or service compared to those of the toughest check marks or similar symbols.
competitor, to those considered industry leaders, or to
205
clinical error A commission or an omission with poten- elopement The unauthorized departure of a patient from
tially negative consequences for the patient that would a controlled care setting.
have been judged incorrect by skilled and knowledgeable
peers at the time it occurred. Synonym: medical mistake. emergency An unexpected or sudden event that signifi-
cantly disrupts the organization’s ability to provide care,
common-cause variation See variation. treatment, or services, or the environment of care itself, or
that results in a sudden, significantly changed or increased
complexity A high number of steps and handoffs in work demand for the organization’s services. Emergencies can be
processes. either human-made or natural (such as an electrical system
failure or a tornado), or a combination of both, and they
complication A detrimental patient condition that arises exist on a continuum of severity.
during the process of providing health care, regardless of
the setting in which the care is provided. For instance, error of commission A mistake that occurs as a result of
perforation, hemorrhage, bacteremia, and adverse reactions an action taken. Examples include a drug being adminis-
to medication (particularly in the elderly) are four compli- tered at the wrong time, in the wrong dose, or using the
cations of colonoscopy and its associated anesthesia and wrong route; surgeries performed on the wrong side of the
sedation. A complication may prolong an inpatient’s length body; and transfusion errors involving blood crossmatched
of stay or lead to other undesirable outcomes. for another patient.
comprehensive systematic analysis A process for identi- error of omission A mistake that occurs as a result of an
fying basic or causal factors underlying variation in perfor- action not taken. Examples include a delay in performing
mance, including the occurrence or possible occurrence of a an indicated caesarean section, resulting in a fetal death;
sentinel event. A root cause analysis is one type of system- a nurse omitting a dose of a medication that should be
atic comprehensive analysis. administered; and a patient suicide that is associated with a
lapse in carrying out frequent patient checks in a psychiatric
contributing factor A circumstance, action, or influence unit. Errors of omission may or may not lead to adverse
that is thought to have played a part in the origin, develop- outcomes.
ment, or increased risk of an incident.
failure Lack of success, nonperformance, nonoccurrence,
control chart A visual representation of tracking progress breaking down, or ceasing to function. In most instances,
over time. and certainly within the context of this book, failure is what
is to be avoided. It takes place when a system or part of a
corrective action See improvement action. system performs in a way that is not intended or desirable.
coupled system A system that links two or more activities failure mode and effects analysis (FMEA) A systematic
so that one process is dependent on another for completion. way of examining a design prospectively for possible ways
A system can be loosely or tightly coupled. in which failure can occur. It assumes that no matter how
knowledgeable or careful people are, errors will occur in
direct cause See proximate cause. some situations and may even be likely to occur. Synonym:
failure mode, effects, and criticality analysis (FMECA).
discovery In the legal sense, the required disclosure of
pertinent facts or documents by parties in a legal action or fishbone diagram (also called cause-and-effect diagram
proceeding. or Ishikawa diagram) The visual representation to clearly
display the various factors affecting a process. This can be
DMAIC The acronym for the five phases of Six Sigma a structured approach to root cause analysis. The diagram
implementation: Defining, Measuring, Analyzing, identifies the inputs or potential causes of a single output
Improving, Controlling. or effect. In a hospital, for example, work can be divided
206
Glossary
into categories: responding to instructions (orders), using and effective. The organization retains the option to
supplies and medications (materials), using equipment complete self-assessment with all applicable accreditation
(machinery), providing the needed care or service in accor- standards in the FSA tool, available on the organization’s
dance with established procedures (methods), and the Joint Commission Connect™ extranet site. See also Intracycle
environment itself (environment). Those categories can be Monitoring (ICM).
listed on a fishbone diagram as the different branches from
which mistakes can arise. To complete the branches, brain- function A group of processes with a common goal.
storm primary causes, ask together why they are occurring,
analyze the causes, and prioritize and identify the likely Gantt chart A graphical depiction of the time frame for a
root causes. long-term, complex project, including the numerous phases
of the project, person(s) responsible, and targeted comple-
five whys A tool used to probe beyond the contributing tion dates, used to help the project team gauge its progress.
factors or proximate causes to find the root cause(s) of an
adverse event by repeatedly asking the question Why? hazardous condition A circumstance (other than a
patient’s own disease process or condition) that increases
flowchart A pictorial summary that shows with symbols the probability of an adverse event.
and words the steps, sequence, and relationship of the
various operations involved in the performance of a func- histogram A visual representation displaying the
tion or a process. distribution of the data within a range of values.
Focused Standards Assessment (FSA) A requirement of human factors The study of the interaction of human
the accreditation process whereby an organization reviews performance (capabilities and limitations) in relation
its compliance with a selected subset of applicable Joint to elements of the system (the design of machines,
Commission accreditation requirements (including the jobs, and other aspects of the physical environment).
applicable National Patient Safety Goals, a subset of direct Synonym: ergonomics.
and indirect impact standards, a selection of standards that
address accreditation program–specific high-risk areas, and immediate cause See proximate cause.
the organization’s Requirements for Improvement [RFIs]
from its last triennial survey); completes and submits to The improvement action A solution required to prevent a
Joint Commission a Plan of Action (POA) for any accredi- problem from occurring or recurring due to the same
tation requirement with which it is not in full compliance, root cause or interaction(s) of root causes. Synonym:
including identifying any required Measure of Success corrective action.
(MOS); and chooses whether to engage in a telephone
discussion with a member of the Standards Interpretation incident report (occurrence report) A written report,
Group staff to determine the acceptability of the POA or usually completed by a nurse and forwarded to risk
discuss any other area of concern. Alternatives for a Full management personnel, that describes and provides docu-
FSA submission include FSA Option 1 (attestation that mentation for any unusual problem, incident, or other
an FSA was completed, but not submitted to The Joint situation that is likely to lead to undesirable effects or that
Commission), Option 2 (on-site survey with documented varies from established policies and procedures.
findings), and Option 3 (on-site survey without docu-
mented findings). The FSA encourages organizations to indicator 1. A measure used to determine, over time,
be in continuous compliance with the Joint Commission performance of functions, processes, and outcomes.
accreditation requirements and helps them to iden- 2. A statistical value that provides an indication of the
tify and manage risk. At the time of the next full survey, condition or direction over time of performance of a
surveyors will validate that the MOS (if applicable or if defined process or achievement of a defined outcome.
submitted as part of the FSA process) was implemented
207
individual served The terms individual served, patient, and Lean Six Sigma A patient-focused health care manage-
care recipient all describe the individual, client, consumer, or ment philosophy that promotes various proven methods of
resident who actually receives health care, treatment, and/ stabilizing, standardizing, and simplifying work processes
or services. to reduce waste and improve quality of care. Lean Six
Sigma practices contribute to reducing adverse events,
Intracycle Monitoring (ICM) A process to help accred- increasing patient perception of high-quality care, raising
ited organizations at various touch points in the triennial levels of staff retention and satisfaction, and saving the
accreditation cycle with their continuous compliance organization money.
efforts. The process involves access to an ICM Profile avail-
able on the organization’s Joint Commission Connect™ licensed practical nurse (LPN) A nurse who has
extranet site. The ICM Profile identifies high-risk areas and completed a practical nursing program and is licensed by
related standards areas and displays them within a Focused a state to provide routine patient care under the direction
Standards Assessment (FSA) tool, which allows organi- of a registered nurse or a physician. Referred to as licensed
zations to conduct a self-assessment of standards to iden- vocational nurse (LVN) in some states.
tify and manage risk in the organization. See also Focused
Standards Assessment (FSA). malpractice Improper or unethical conduct or unreason-
able lack of skill by a holder of a professional or official
The Joint Commission An independent, not-for-profit position; often applied to physicians, dentists, lawyers, and
organization, The Joint Commission is dedicated to public officers to denote negligent or unskillful perfor-
improving the safety and quality of health care through mance of duties when professional skills are obligatory.
standards development, public policy initiatives, accredita-
tion, and certification. The Joint Commission accredits and measurement The process of collecting and
certifies more than 15,000 health care organizations and aggregating data.
programs in the United States.
medication Any prescription medications, sample medica-
Joint Commission International (JCI) JCI extends The tions, herbal remedies, vitamins, nutriceuticals, vaccines, or
Joint Commission’s mission worldwide. Through both over-the-counter drugs; diagnostic and contrast agents used
international consultation and accreditation, JCI helps on or administered to persons to diagnose, treat, or prevent
to improve the quality and safety of patient care in many disease or other abnormal conditions; radioactive medica-
nations. JCI has extensive international experience working tions, respiratory therapy treatments, parenteral nutrition,
with public and private health care organizations and local blood derivatives, and intravenous solutions (plain, with
governments in more than 40 countries. electrolytes and/or drugs); and any product designated by
the US Food and Drug Administration (FDA) as a drug.
Joint Commission Resources (JCR) Not-for-profit This definition of medication does not include enteral
affiliate of The Joint Commission designated by The Joint nutrition solutions (which are considered food products),
Commission to publish publications and multimedia prod- oxygen, and other medical gases.
ucts. JCR reproduces and distributes these materials under
license from The Joint Commission. Go to www.jcrinc.org. medication error A preventable event that may cause or
lead to inappropriate medication use or patient harm while
kaizen An intensive process used to improve a small, the medication is in the control of the health care profes-
focused issue in a week or less. sional, patient, or consumer. Such events may be related
to professional practice, health care products, procedures,
latent condition A condition that exists as a consequence and systems, including prescribing; order communica-
of management and organizational processes and poses tion; product labeling, packaging, and nomenclature;
the greatest danger to complex systems. Latent conditions compounding; dispensing; distribution; administration;
can be identified and corrected before they contribute education; monitoring; and use.
to mishaps.
208
Glossary
multivoting A tool used to narrow down a broad list of Plan-Do-Study-Act (PDSA) cycle A four-part method
ideas (such as those generated through brainstorming) to for discovering and correcting assignable causes to improve
those that the team members collectively find most worthy the quality of processes. Synonyms: Deming cycle, Shewhart
of immediate attention. cycle, Plan-Do-Check-Act (PDCA) cycle.
near miss A patient safety event that did not reach the policy A principle or method that is developed for the
patient; also called a close call or a good catch. purpose of guiding decisions and activities related to
governance, management, care, treatment, and services.
negligence Failure to use such care as a reasonably prudent A policy is developed by organization leadership, approved
and careful person would use under similar circumstances. by the governing body of the organization, and maintained
in writing.
no-harm event A patient safety event that did not result in
harm to the patient. practice guidelines Tools that describe a specific procedure
or processes found, through clinical trials or by consensus
occurrence report See incident report. opinion of experts, to be the most effective in evaluating
and/or treating a patient, resident, or individual served who
operational definition A description of a concept demon- has a specific symptom, condition, or diagnosis. Synonyms
strated by the process that includes a means of measurement. include clinical practice guideline, practice parameter,
An operational definition may be used during the creation of protocol, preferred practice pattern, and guideline.
a project charter, when defining a customer’s requirements,
when collecting data, or when analyzing capability. practitioner Any individual who is licensed and qualified
to practice a health care profession (for example, physician,
operative or other high-risk procedures Operative and nurse, social worker, clinical psychologist, or respiratory
other invasive and noninvasive procedures (performed in therapist) and is engaged in the provision of care, treat-
order to remedy an injury, ailment, defect, or dysfunction) ment, or services.
that place the patient at risk. The focus is on procedures
and is not meant to include medications that place the prescribing or ordering The process of a licensed inde-
patient at risk. pendent practitioner or prescriber transmitting a legal order
or prescription to the organization directing the preparing,
outcome The result of the performance (or nonperformance) dispensing, and administering of a specific medication to
of a function(s) or process(es). a specific individual. It does not include requisitions for
medication supplies.
outcome measure A tool used to assess data that indicates
the results of performance or nonperformance of a function prevention/early detection (domain) The degree to
or procedure. which appropriate services are provided for promotion,
preservation, and restoration of health and early detection
Pareto chart A vertical bar graph that displays the data of disease.
being studied in order from largest to smallest. The Pareto
chart is useful in analyzing the frequency of problems or privileging The process whereby the specific scope and
causes in a process, when wanting to focus on the most content of patient care services (that is, clinical privileges)
significant problems or concerns, when analyzing broad are authorized for a health care practitioner by a health care
cases by looking at their specific components, and when organization, based on evaluation of the individual’s creden-
communicating about data. tials and performance.
patient safety event An event, incident, or condition that

could have resulted or did result in harm to a patient.
See also adverse event, near miss, sentinel event.
209
procedure 1. A series of steps taken to accomplish a quality improvement An approach to the continuous
desired end, as in a therapeutic, cosmetic, or surgical proce- study and improvement of the processes of providing health
dure. 2. A unit of health care, as in services and procedures. care services to meet the needs of individuals and others.
A procedure is not necessarily developed by organization Synonyms include continuous quality improvement, contin-
leadership, approved by the governing body of the organiza- uous improvement, organizationwide performance improve-
tion, and maintained in writing. ment, and total quality management.
process A goal-directed, interrelated series of actions, quality of care The degree to which care, treatment, and
events, mechanisms, or steps that transform inputs services for individuals and populations increases the like-
into outputs. lihood of desired health outcomes. Considerations include
the appropriateness, efficacy, efficiency, timeliness, accessi-
process measure An intermediate indicator of the success bility, and continuity of care; the safety of the care envi-
of an intervention. ronment; and the individual’s personal values, practices,
and beliefs.
proficiency testing The assessment of technical knowledge
and skills relating to certain occupations. referral The sending of an individual (1) from one clini-
cian to another clinician or specialist, (2) from one setting
proximate cause A system failure that naturally and or service to another, or (3) by one physician (the referring
directly produces a consequence. It is the superficial or physician) to another physician(s) or other resource, either
obvious cause for an occurrence. Treating only the symp- for consultation or care.
toms, or the proximate special cause, may lead to some
short-term improvements but does not prevent the varia- registered nurse (RN) An individual who is licensed to
tion from recurring. practice professional nursing.
pull system A system driven by the needs of the down- relations diagram A tool used to generate understanding
stream lines. By specifying value, identifying the value of how various aspects of a problem are connected,
stream, and creating flow, Lean thinking allows pull to including cause-and-effect relationships.
take place. In a Lean Six Sigma health care system, patients
pull the product along rather than having the market- relevance The applicability and/or pertinence of the indi-
place push it onto them on the organization’s timetable. cator to its users and customers. For Joint Commission
Pull offers more flexibility and accommodates changes in purposes, face validity is subsumed in this category.
customer demand.
reliability The capability of an indicator to accurately and
push system Work that is driven by the output of the consistently identify the events it is designed to identify
preceding lines. This is work that is pushed along regard- across multiple health care settings.
less of need or request. It involves providing a service or
product in anticipation of a need and is often associated resilience The degree to which a system continuously
with high inventory and the risk of errors or higher error prevents, detects, mitigates, or ameliorates hazards
rates. In a Lean Six Sigma health care system, push must be or incidents.
eliminated and replaced with a pull system to facilitate flow.
respect and caring The degree to which those providing
quality control A set of activities or techniques whose services do so with sensitivity for an individual’s needs,
purpose is to ensure that all quality requirements are being expectations, and individual differences, and the degree to
met. In order to achieve this purpose, processes are moni- which the individual or a designee is involved in his or her
tored and performance problems are solved. own care decisions.
210
Glossary
risk adjustment A statistical process for reducing, scatter diagram A illustration graphically plotting pairs of
removing, or clarifying the influences of confounding numerical data to display the possible relationship— not
factors that differ among comparison groups (for example, necessarily a cause-and-effect relationship—between one
logistic regression, stratification). variable and another. If the variables are correlated, the data
points will fall along a line or curve; the better the correla-
risk containment Immediate actions taken to safeguard tion, the more closely the points will adhere to the line.
individuals from a repetition of an unwanted occurrence.
Actions may involve removing and sequestering drug stocks scientific method The systematic process of determining
from pharmacy shelves, checking or replacing oxygen what is known about a problem, deciding what needs
supplies or specific medical devices, or cordoning off an icy to be changed, forming a hypothesis for implementing
section of a walkway. change, testing the hypothesis, and evaluating the result.
A successful result would lead to implementation of the
risk management activities Clinical and administrative change; an unsuccessful result would lead to restarting
activities undertaken to identify, evaluate, and reduce the the process.
risk of injury to patients, staff, and visitors and the risk of
loss to the organization itself. seclusion The involuntary confinement of an individual in
a room alone, for any period of time, from which the indi-
risk points Specific points in a process that are susceptible vidual is physically prevented from leaving. Seclusion does
to failure or system breakdown. They generally result from not include involuntary confinement for legally mandated
a flaw in the initial process design, a high degree of depen- but nonclinical purposes, such as the confinement of a
dence on communication, nonstandardized processes, and/ person who is facing serious criminal charges or who is
or failure or absence of backup. serving a criminal sentence.
root cause A fundamental reason for the failure or sentinel event 1. As defined by The Joint Commission:
inefficiency of a process. A patient safety event (not primarily related to the natural
course of the patient’s illness or underlying condition)
root cause analysis A process for identifying the basic that reaches a patient and results in death, permanent
or causal factor(s) underlying variation in performance, harm, or severe temporary harm. Sentinel events are
including the occurrence or possible occurrence of a a subcategory of adverse events. 2. As defined by Joint
sentinel event. Commission International: An unanticipated occurrence
involving death or major permanent loss of function
run chart A tool for measuring variation in the perfor- unrelated to the natural course of the patient’s illness or
mance of a given task or process. underlying condition.
safety The degree to which the risk of an intervention (for services Structural divisions of an organization, its medical
example, use of a drug or a procedure) and risk in the care staff, or its licensed independent practitioner staff.
environment are reduced for a patient and other persons,
including health care practitioners. Safety risks may arise SIPOC process map A high-level process map that depicts
from the performance of tasks, from the structure of the suppliers, inputs, process, outputs, and customers.
physical environment, or from situations beyond the
organization’s control (such as weather). Six Sigma The measure of variation that achieves
3.4 defects per million opportunities, or 99.99966%
safety management Activities selected and imple- acceptability.
mented by the organization to assess and control the
impact of environmental risk, and to improve general special-cause variation See variation.
environmental safety.
211
staff As appropriate to their roles and responsibilities, all survey A key component in the accreditation process,
people who provide care, treatment, and services in the whereby a surveyor(s) conducts an on-site evaluation of an
organization, including those receiving pay (for example, organization’s compliance with Joint Commission or Joint
permanent, temporary, and part-time personnel, as well Commission International accreditation requirements.
as contract employees), volunteers, and health profession
students. The definition of staff does not include licensed surveyor For purposes of Joint Commission accreditation,
independent practitioners who are not paid staff or who are a licensed physician, surgeon, podiatrist, dentist, nurse,
not contract employees. physician assistant, administrator, social worker, psychol-
ogist, behavioral health care professional, or any other
staffing effectiveness The number, competence, and skill health care professional who meets The Joint Commission’s
mix of staff as related to the provision of needed care, treat- surveyor selection criteria, evaluates compliance with
ment, and services. accreditation requirements, and provides education and
consultation regarding compliance with accreditation
stakeholder analysis A tool used to ascertain the level of requirements to surveyed organizations or systems.
commitment from key people involved in a process change.
tracer methodology A process surveyors use during the
standard A principle of patient safety and quality of care on-site survey to analyze an organization’s systems, with
that a well-run organization meets. A standard defines the particular attention to identified priority focus areas, by
performance expectations, structures, or processes that must following individual patients through the organization’s
be substantially in place in an organization to enhance the care process in the sequence experienced by each indi-
quality of care, treatment, or services. vidual. Depending on the setting, this process may require
surveyors to visit multiple care programs and services
standard deviation A measure of variability that indicates within an organization or within a single program or service
the spread of a set of observations around the mean. to “trace” the care rendered.
standard work A tool used to create a logical work flow underlying cause The systems or process cause that allows
with a minimum of waste. for the proximate cause of an event to occur. Underlying
causes may involve special-cause variation, common-cause
suicide The act of ending one’s own life voluntarily and variation, or both and may or may not be root causes.
intentionally.
utility systems Building systems that provide support
surveillance Systematic method of collecting, consoli- to the environment of care, including electrical distribu-
dating, and analyzing data concerning the frequency or tion and emergency power; vertical and horizontal trans-
pattern of, and causes or factors associated with, a given port; heating, ventilating, and air-conditioning (HVAC);
disease, injury, or other health condition. Data analysis is plumbing, boiler, and steam; piped gases; vacuum
followed by the dissemination of that information to those systems; and communication systems, including data
who can improve outcomes. Examples of surveillance data exchange systems.
can include ventilator-associated pneumonia, antibiotic
prophylaxis, hemodialysis catheter infections, implant value stream mapping A tool used to help health care
infections, surgical site infections, hand hygiene, multi- organizations operate successfully by eliminating, or at least
drug-resistant organisms (MRSA, VRE), equipment sterile minimizing, non-value-added activities—that is, any steps
processing, vaccinations, urinary tract infections, and health in a process that do not contribute to a patient’s experience
care worker immunization. of value.
212
Glossary
variation The differences in results obtained in measuring by interactions of variables of that process inherent in
the same phenomenon more than once. Excessive variation all processes, not a disturbance in the process. It can be
frequently leads to waste and loss, such as the occurrence removed only by making basic changes in the process.
of undesirable patient health outcomes and increased cost Special-cause variation (also called exogenous-cause variation
of health services. The sources of variation in a process or extrasystemic cause variation) in performance results from
over time can be grouped into two major classes: common assignable causes. Special-cause variation is intermittent,
causes and special causes. Common-cause variation (also unpredictable, and unstable. It is not inherently present in
called endogenous-cause variation or systemic-cause variation) a system; rather, it arises from causes that are not part of the
in a process is due to the process itself and is produced system as designed.
213
214
Index
A B
Accountability, 7 Bar-coded sponge system, 201
Accreditation Beds
sentinel event requirements in, 27–28, 40 alarm systems for, 68, 72
survey process in, 40, 212 falls from, 7, 68, 72, 81
Action plans, 2, 8–12, 91–131 guard rails on, 54
acceptability of, 24, 29, 108 Benchmarking, 115, 205
definition of, 205 Benefits of root cause analysis, 7
framework for, 12, 13–15, 24 Blame
on geriatric adverse events in VHA, 178, 180–184 and creating no-blame environment, 48
implementation of, x, 12, 108–117, 183 fear of, 53, 54
what, how, when, who, and where issues in, 104–108 Brainstorming, 59, 135, 138–139
Active failures, 5 definition of, 205
Adverse drug events, 205 on improvement actions, 102, 112, 113
Adverse drug reactions, 205 on problem definition, 50, 52
Adverse events, 20, 205 on proximate causes, 68, 69, 71,
communication with patient and family on, 38, 40 on surgical error, 139
documentation of, 37–38 Breakaway compliant fixtures, 55
in geriatric VHA patients, 175–184 Buffers, built-in, in risk reduction strategies, 93, 94
international incidence of, vii
reducing risk of recurrence, 7–8 C
Affinity diagrams, 71, 135, 136–137, 205 Capability charts, 135, 140, 205
Aggregate, definition of, 205 Care, x, 205
Aggregate data, 73, 205 for patients affected by sentinel event, 37
and continuous variable indicators, 73–74 quality of, 210
on improvement actions, 115 respect in, 210
and rate-based indicators, 73 Care recipient, x
on root causes and risk reduction strategies, 26 Case studies, x, 175–202
AIDS, incorrect ordering of diagnostic tests in, 186–192 on geriatric adverse events in VHA, 175–184
American Society for Healthcare Risk Management, 37 on incorrect ordering of HIV tests, 186–192
Attendance to team meetings, 47 on retained surgical item, 194–202
Audio recordings, 54 Cause-and-effect diagrams, 149–150, 206–207.
Automation, in risk reduction strategies, 93 See also Fishbone diagrams
Change analysis, 59, 68, 135, 141, 205
Change management, 84, 111, 135, 142, 205
215
Change process, 105, 108, 113 in retained surgical item, 202

resistance to, 108, 128 safety in, 202
stages of, 105, 108, 113 Czech Republic, mandatory reporting system in, 29
Check sheets, 135, 143–144, 205
Clinical errors, 205–206 D
Clinical practice guidelines, 209 Data collection. See Information and data collection
Close calls. See Near-miss events Decision making in team, 47
Comment questions, 57 Defensiveness in interviews, 56, 136
Common-cause variations, 6, 82, 213 Define, measure, analyze, improve, and control (DMAIC)
Communication process, 110–112, 133, 206
as causal factor, 84 Deming cycle, 109–110, 209
checklist on, 85 Denmark, mandatory reporting system in, 29
errors related to, 69 Direct causes, 68. See also Proximate causes
on improvement actions, 116–117 Disclosable information, 24, 36
on investigation progress, 52–53 Discovery, definition of, 206
in medication errors, 9 DMAIC (define, measure, analyze, improve, and control)
with patients and families, 40, 42, 43, 197 process, 110–112, 133, 206
in retained surgical item, 200, 204 Documentation
in team approach, 50–51 as evidence, 58
in VHA, and geriatric adverse events, 177, 180, 182, incident reports in, 37–39, 207
183, 185
Competency assessment, 11 E
Complexity, 206 Early detection/prevention, definition of, 209
risk for system failures in, 50 Elopement, 44, 206
Complication, definition of, 206 improvement actions in, 101
Comprehensive Accreditation Manual, 19 information collected on, 55
Comprehensive systematic analysis, vii, 1, 206 literature review on, 59
requirement for, 19, 23, 24 reporting mechanism for, 53
Computerized provider order entry (CPOE) systems, 191 team investigation of, 47, 53
Confidentiality issues, 22–23, 33 Emergency, definition of, 206
contractual agreements on, 23, 30–31 Endogenous-cause (common cause) variations, 6, 82, 213
developing policies on, 36 Environmental factors, 12, 14, 69, 82, 84
Contemplation stage in change process, 105, 108 checklist on, 85
Continuous variable indicators, 73–74 identification of, 79–80
Contractual agreements, 23, 30–31 in medication errors, 12
Contributing factors, 7, 206 questions in analysis of, 83
differentiated from root cause, 86 Equipment factors, 12, 69, 71
identification of, 68–71 categories of, 84
in near-miss events, 18, 50 in geriatric adverse events in VHA, 179, 180–181
ranking of, 50 identification of, 79
Control charts, 59, 135, 145–146, 206 in medication errors, 12
on improvement actions, 113, 115 Errors, vii, 21
patterns in, 146 clinical, 205–206
Corrective actions, 98. See also Improvement actions of commission, 19, 206
Coupled systems, 50, 206 communication with patient and family on, 38, 40, 41,
Culture, organizational, 36–37 195
hierarchies in, 50
216
Index
medical, compared to sentinel events, 34 Flowcharts, 59, 68, 135, 151–152

medication. See Medication errors definition of, 207
of omission, 19, 197, 198, 206 on improvement actions, 102, 113
retained surgical item in, 194–202 on medication errors, 99, 152
Swiss chess model of, 3, 4, 92 on process problems, 83–84
Ethical issues in disclosure of errors, 38 underlying causes identified in, 82
Evaluation checklist on root cause analysis, 8, 10–11 Focused Standards Assessment (FSA), 207
Event investigation. See Problem investigation Follow-up activities, 26, 33
Evidence preservation in sentinel events, 37, 57–58 Follow-up questions, 57
Exogenous-cause (special-cause) variations, 6, 82, 213 Forcing functions, 92
Expectations in improvement actions, 104, 122 Form design for laboratory tests, errors related to, 189, 190,
Exploratory questions, 57 191
Extrasystemic cause (special-cause) variations, 6, 82, 213 Frequency/severity matrix on safety-related problems, 3
Function, definition of, 101, 207
F
Fail-safe design, 95 G
Failure mode, effects, and criticality analysis (FMECA), 97, Gantt charts, 52, 59, 75, 135, 153
147–148, 206 definition of, 207
Failure mode and effects analysis (FMEA), ix, 2–3, 97–98, on improvement actions, 104–105, 106–107, 153
135, 147–148 on suicide investigation, 75, 76
definition of, 206 Geriatric adverse events in VHA, 175–184
effective use of, 147 Governance, Leadership, and Direction (GLD) Standards,
frequency/severity matrix in, 3 34–35
of potential improvement actions, 102 Group interviews, 55–56, 57
steps in, 98, 147, 148 Guidelines, practice, 209
Failures, vii
active, 5 H
contributing factors in, 7 Hand hygiene, 82, 144
definition of, 206 Hazardous conditions, 18, 19, 207
to follow process, 69 High-reliability organizations, 3, 92, 93, 95
process design reducing risk of, 93 High-risk processes, identification of, 50
Falls Histograms, 113, 115, 135, 154, 207
from bed, 7, 68, 72, 81 HIV infection, incorrect ordering of diagnostic tests in,
multiple root causes in, 7, 87 186–192
of older patients in VHA, 175–184 HTLV-I/II tests, incorrect ordering of, 186–192
Families affected by sentinel event, communication with, Human factors, 12, 69
38, 40 categories of, 84
Fear of blame, 53, 54 definition of, 207
Fishbone diagrams, 59, 68, 71, 135, 149–150 identification of, 79
definition of, 206–207 Human resources issues, 13, 82
effective use of, 149 checklist on, 85
on improvement actions, 102, 113 questions in analysis of, 83
on process problems, 83–84 in suicide event, 83
on suicide event, 150 Human Subjects Research Programs (HRP) standards, 35
Five whys, definition of, 207 Human T-cell lymphotropic virus (HTLV), inappropriate
Flow, in process, 134, 172 testing for, 186–192
217
I Institute of Medicine, vii, 191

Immediate causes. See Proximate causes Intentional act causing sentinel event, 86
Improvement actions, 91–131 Interviews, 54, 55–57
categories of, 98 defensiveness in, 56, 136
change process in, 105, 108, 113 with group, 55–56, 57
communication of results in, 116–117 open-ended questions in, 56, 57
defining goals, scope, and activities in, 104, 121 recording methods, 54
definition of, 207 stages pf, 56–57
DMAIC model in, 110–112 Intracycle Monitoring (ICM), 208
expectations in, 104, 122 Investigation of event. See Problem investigation
in immediate changes, 74–75 Ireland, mandatory reporting system in, 29
implementation of, 108–117 Ishikawa, Kaoru, 134
individuals responsible for, 105, 124–125 Ishikawa diagrams, 134, 149–150, 206–207.
internal comparisons of, 115 See also Fishbone diagrams
location of, 105, 126
measurements on, 72, 102, 104, 113–115, 122 J
mistake proofing in, 98–102 Japan, mandatory reporting system in, 29
performance targets in, 115, 131 The Joint Commission, 208
pilot test of, 108 acceptance of action plan by, 24, 108
in plan-do-study-act cycle, 109–110, 111 reporting of sentinel events to, 20, 22–26
prioritizing of, 102, 119 Sentinel Event Database of, 22, 26, 28, 50, 86
problems and remedies in, 115–116 sentinel event definition of, 1, 17, 211
proposed, evaluation of, 102–104, 119–120 Sentinel Event Hotline of, 1, 20, 22
risk reduction strategies in, 91–98 Sentinel Event Policy of, ix, 1, 3–4, 8, 9, 17–28, 32–33
scientific method in, 102, 109 Universal Protocol for Preventing Wrong Site, Wrong
time frame for, 104–105, 106–107, 123 Procedure, Wrong Person Surgery, 75, 94
what, how, when, who, and where issues in, 104–108, web site resources of, 9, 23, 24
127 The Joint Commission Connect extranet site, 19, 22, 23
Incident reports, 37–38, 207 Joint Commission International (JCI), 208
Indicators, 72–74 acceptance of action plan by, 29, 108
definition of, 72, 73, 207 sentinel event definition of, 1, 211
selection of, 114 Sentinel Event Policy of, ix, 4–5, 8, 17, 28–35
Individual served, x, 208 The Joint Commission Perspectives, 50
Infant abductions, 87, 97 Joint Commission Resources (JCR), 208
Information and data collection
on improvement actions, 114–115 K
legal issues in, 54, 55 Kaizen, 135, 155–156, 208
measurements in, 71–74, 114–115
in problem investigation, 53–59. L
See also Problem investigation Labeling of medications, 10
on proximate and underlying causes, 71–74 Laboratory testing
recording methods in, 54, 55 affinity diagram on, 137
Information management issues, 14, 69, 82 computerized provider order entry systems in, 191
checklist on, 85 for HIV infection, incorrect orders in, 186–192
questions in analysis of, 83 value stream mapping on, 173
in suicide event, 83 Language barriers, 50
Institute for Safe Medication Practices, 97 Latent conditions, 5, 8, 87, 195, 208
218
Index
Leadership, 14, 83 flowcharts on, 99, 152

checklist on, 85 improvement actions in, 102, 114
in improvement actions, 105, 116 information collected on, 55, 114
and organizational culture, 36–37 literature review on, 59
questions in analysis of, 83 mistake proofing in prevention of, 100
reports provided to, 116–117 multiple root causes in, 87
of team, 45, 47, 48–49 and near misses, 18
Leadership Standards, 4, 27–28, 34–35 operational definition of, 159
Lean Six Sigma, 133–134, 208 physical evidence on, 57
Leape, Lucian L., 93 preservation of evidence in, 37
Legal issues proximate and root causes of, 7
contractual agreements on, 23, 30–31 questions in investigation of, 9–12
in protection of information, 54, 55 reporting mechanism for, 53
in retained surgical item, 194–195 risk containment in, 37
in sentinel events, 22–23, 25, 26, 36, 38, 40, 54 risk points and risk reduction strategies in, 94, 95, 97,
Licensed practical nurse (LPN), 208 99, 100
Literature review, 59 scatter diagram on, 165
Location of improvement actions, 105, 126 team investigation of, 47, 53
in VHA, 175, 182, 183
M Meetings of team members, 47–49
Malpractice claims, 38, 40, 208 Mistake proofing, 98–102
in retained surgical item, 194–195 Motorola, 133
Measurement, 71–74 Multivoting, 112, 113, 135, 157, 209
checklist on, 74, 114
data in, 205 N
definition of, 71, 113, 208 National Academy for State Health Policy, 25
of geriatric adverse events in VHA, 178 National Center for Patient Safety, 97, 101, 175–184
of improvement actions, 72, 102, 104, 113–115, 122 National Patient Safety Goals, 28, 36, 207
indicators used in, 72–74, 114 Near-miss events, 18, 209
key questions on, 74, 114, 129–130 as learning opportunity, 19
in plan-do-study-act cycle, 114 problem statement in, 49–50
plan for, 114, 129–130 root cause analysis of, 2
purposes of, 72 Negligence, definition of, 209
Media No-blame environment, 52
notification of event, 38 No-harm events, 18, 19, 213
reports on patient safety events in, viii, 2, 20, 23, 29, 50 Non-value-added activities, 212
Medical mistakes and errors Note taking during interviews, 54
compared to sentinel events, 34
definition of, 41 O
physician disclosure of, 40, 41, 195 Observations and statements of witnesses, 55–57
Medication, definition of, 208 Occurrence (incident) reports, 37–39, 207
Medication errors, 2, 5, 44 Older patients in VHA, serious adverse events among,
case example of, 2 175–184
in children, viii On-site visits, 24, 26
definition of, 208 Open-ended interview questions, 56, 57
detailed description of, 67 Operating suite, risk reduction strategies in, 96
documentary evidence on, 58 Operational definitions, 135, 158, 209
219
Operative or other high-risk procedures, 209 Privileging, definition of, 209

Ordering, 209 Proactive risk reduction, viii, ix
computerized provider order entry systems in, 191 failure mode and effects analysis in, ix, 2–3, 97–98
of laboratory tests for HIV diagnosis, 186–192 Problem definition, 49–53
Outcome, 209 brainstorming on, 50, 52
measures of, 73, 209 detailed description in, 67–68, 77
identification of high-risk processes in, 50
P worksheet on, 61
Pareto charts, 113, 135, 159–160, 209 Problem investigation, 53–59
Patients in early response, 37–40
affected by sentinel event, 37, 38, 40 evidence in, 37, 57–58
falls of. See Falls fear of blame in, 53, 54
identification of, 9 flowcharts in, 68
rights of, 28 interviews in, 54, 55–57
use of term, x legal issues in, 54, 55
Patient Safety and Quality Improvement Act (2005), 25 literature review in, 59
Patient safety events, vii, 1 recording methods in, 54, 55
definition of, 18, 209 team approach to, 44–49
examples of, viii witness statements and observations in, 55–57
multiple, 5 work plan in, 52, 53
and near-miss events, 18 written responses in, 57, 63–65
no-harm, 18 Procedure
as sentinel events, 17–18, 19, 23 definition of, 210
types of, 18 failures related to, 69, 70
Performance measures, 72 and geriatric adverse events in VHA, 179, 180
Photographs for recording information, 54 standardization of, 93, 101, 180
Physicians Process
disclosure of medical mistakes to patient and family, 40, checklist in analysis of, 84, 85
41 contributing factors in, 68–69, 78
on performance improvement teams, 46 definition of, 101, 210
stages in change process, 105, 108 failure to follow, 69
Pilot test of improvement actions, 108 flawed, 69, 101
Plan-Do-Check-Act (PDCA) cycle, 109, 209 flowcharts and fishbone diagrams on, 83–84
Plan-Do-Study-Act (PDSA) cycle, 75, 109–110, 111, 209 flow in, 134, 172
components of, 110 high-risk, 50
measurement in, 114 improvement actions affecting, 103
questions considered during, 111 measures of, 72, 73, 210
Policy, definition of, 209 need for change in, 8
Practice guidelines or parameters, 209 pull in, 134
Practitioner, definition of, 209 risk points in, 94–97
Precontemplation stage in change process, 105, 108 risk reduction strategies on, 93–97
Preferred practice patterns, 209 Proficiency testing, 210
Preoperative procedures, 94, 96 Protocols (practice guidelines), 209
Preparation for root cause analysis, ix–x, 43–65 Proximate causes, x, 5, 6–7, 67–80
Prescribing, definition of, 209 definition of, 68, 210
Prevention/early detection, definition of, 209 differentiated from root cause, 86
220
Index
underlying causes of, 81 time allowed for RCA in, 52

Pull system, 134, 172, 210 Risk adjustment, 211
Push system, 172, 210 Risk containment, 37, 211
Risk management activities, 211
Q Risk points, 94–97, 211
Quality control, 210 Risk reduction strategies, 7–8, 15, 91–98
Quality improvement, vii, 33–34, 210 failure mode and effects analysis in, ix, 2–3, 97–98
Quality Improvement and Patient Safety (QPS) Standards, forcing functions in, 92
4, 33–34 of high-reliability organizations, 92, 93
Quality of care, 210 identification of, 96
Questioning technique in interviews, 56, 57 immediate changes in, 74–75
proactive, vii, ix, 2–3, 97–98
R systems approach in, 3, 94–97
Radiofrequency identification of sponges, 201, 202 Root cause
Rate-based indicators, 73 definition of, 1–2, 211
Recording information, methods of, 54, 55 differentiated from contributing cause, 86
Redundancy in risk reduction strategies, 93, 94 identification of, x, 81–89
Referrals, 210 multiple, interrelationships of, 86–98
Registered nurse (RN), 210 pruning list of, 84, 86
Relations diagrams, 135, 161–162, 210 Root cause analysis
Relevance, definition of, 210 and action plan. See Action plans
Reliability, 210 benefits of, 7
and high-reliability organizations, 3, 92, 93, 95 case studies on, x, 175–202
Reporting of sentinel events, 19, 33, 37–40 completion dates in, 24, 29, 52, 53
confidentiality in, 22–23, 33, 36 definition of, vii, 1–2, 211
developing policies on, 35–36 effective, strategies for, 8
disclosable information in, 24, 36 evaluation checklist on, 7, 9–10
documentation for, 37–38 and failure mode and effects analysis, 99, 100
by electronic submission, 24 framework for, 12, 12–15, 24
on hotline, 20, 24 immediate changes made during, 74–75
international differences in, 29 JCI policy on, 29
in investigation process, 52–53 maximizing value of, 7–8
to The Joint Commission, 20, 22–26 overview of, ix, 1–15
to Joint Commission International, 29 preparing for, ix–x, 43–65
legal issues in, 22–23, 25, 26, 36, 38, 40 proactive, 97–98
methods of, 22, 24, 26 problem defined in, 49–53. See also Problem definition
notification checklist on, 38, 39 problem investigation in, 53–59.
in on-site interactions, 24, 26 See also Problem investigation
state requirements on, 25 proximate causes in, x, 67–80. See also Proximate causes
Research programs with human subjects, 35 reluctance in, 88
Resilience, viii, 210 steps in, ix–x, 11, 43, 67, 81, 91
Resources required for improvement actions, 103, 104, 122 team approach to, 44–49. See also Team approach to
Respect, 210 root cause analysis
Retained surgical item, 194–202 time allowed for, 24, 29, 52
Reviewable sentinel events, 7, 18–19 tools and techniques in. See Tools and techniques
JCI policy on, 29 Run charts, 135, 163, 211
required response to, 24–28
221
S reviewable. See Reviewable sentinel events

Safety, 211 risk containment in, 37
culture of, 202 standards related to, 27–28, 32, 33–35
institutionalized, 93 underlying cause of, 5, 68, 81–82, 89, 212
Safety Assessment Code Matrix, 176, 177 variation as factor in, 6, 82, 213
Safety management, 211 work plan in, 52, 53
Scatter diagrams, 113, 135, 164–165, 211 Services, 211
Scientific method, 102, 109, 211 Sharp-end errors, 195
Seclusion, 211 Shewhart cycle (plan-do-study-act cycle), 75, 109–110,
errors in, 19 111, 114, 210
Sentinel Event Alerts, 26, 28, 50, 51 Single-event indicators, 73
Sentinel Event Database, 22, 26, 28, 50, 86 SIPOC (suppliers, inputs, process, outputs, customers)
Sentinel Event Hotline, 9, 20, 22 process map, 135, 166–167, 211
Sentinel Event Measure of Success (SE MOS), 26 Six Sigma, 133–134, 211
Sentinel Event Policies, 3–5 lean, 133–134, 208
action plan in, 8, 29 Slips, in sponge counting, 197
comparison of, 32–33 Special-cause variations, 6, 82, 213
goals of, 19–20, 28–29, 32 Sponge-counting protocol, 194–202
of The Joint Commission, ix, 1, 3–4, 3–5, 8, 9, 17–28, Stable process, 6
32–33 Staff, 9, 11, 212
of Joint Commission International, ix, 4–5, 8, 17, and geriatric adverse events in VHA, 179, 180
28–35 interviews with, 55–57
steps in development of, 35–36 leadership role of, 28
Sentinel events, ix, 3–5, 17–41 on team organized for root cause analysis, 45
accreditation requirements on, 27–28, 40 Staffing effectiveness, 212
appropriate patient care in, 37 Stakeholder analysis, 52, 135, 168–169, 212
communication and disclosure on, 38–40 Standard deviation, 212
compared to medical errors, 34 Standards, 212
comprehensive systematic analysis of, vii, 1, 19, 23, 25, on sentinel events, 27–28, 32, 33–35
206 Standard work, 135, 170–171, 212
confidentiality issues in, 22–23, 33, 36 State reporting requirements, 25
criterion for, 6, 17, 23, 29 Storage of medications, 12
definitions of, vii, 1, 16, 23, 32, 34, 35, 211 Suicide, 44, 212
documentation of, 37–38 documentary evidence on, 58
examples of, 8–9, 20, 44 fishbone diagram on, 150
follow-up activities on, 26, 33 Gantt chart on, 75, 76
framework for response to, 12, 13–16 immediate changes after, 74–75
immediate changes after, 74–75 improvement actions in, 101, 114
investigation of, 53–59. See also Problem investigation information collected on, 55
leadership and organizational culture in, 38–39 literature review on, 59
legal issues in, 22–23, 25, 26, 36, 38, 40, 54 physical evidence on, 58
preservation of evidence in, 37 proximate causes in, 69
problem definition in, 49–53 reporting mechanism for, 52
process flow in, 19, 21 risk points and risk reduction strategies in, 96, 101
proximate cause of. See Proximate causes root causes in, 82–83, 86
reporting of. See Reporting of sentinel events team investigation of, 46–47, 52
response to, 18, 24–28, 33, 36–40
222
Index
Suppliers, inputs, process, outputs, customers (SIPOC) control charts, 59, 113, 115, 135, 145–146, 206
process map, 135, 166–167, 211 DMAIC model, 110–112, 133, 206
Surgical procedures failure mode and effects analysis, 147–148.
brainstorming on, 139 See also Failure mode and effects analysis
and geriatric adverse events in VHA, 182, 183 fishbone diagrams, 149–150. See also Fishbone diagrams
retained surgical item in, 194–202 flowcharts, 151–152. See also Flowcharts
risk points in, 94 Gantt charts, 153. See also Gantt charts
universal protocol on, 75, 94 histograms, 113, 115, 135, 154, 207
wrong site in. See Wrong-site surgery kaizen, 135, 155–156, 208
Surveillance, 212 Lean Six Sigma, 133–134, 208
Survey, 40, 212 multivoting, 112, 113, 135, 157, 209
Surveyors, 40, 212 operational definitions, 135, 158, 209
Swiss cheese model on errors, 3, 4, 92 Pareto charts, 113, 135, 159–160, 209
Systemic-cause (common-cause) variation, 6, 82, 213 relations diagrams, 135, 161–162, 210
Systems approach in risk reduction strategies, 3, 94–97 run charts, 135, 163, 211
scatter diagrams, 113, 135, 164–165, 211
T SIPOC process maps, 135, 166–167, 211
Team approach to root cause analysis, 44–49 stakeholder analysis, 52, 135, 168–169, 212
ad hoc members in, 45, 47, 60 standard work, 135, 170–171, 212
advantages of, 45 value stream mapping, 135, 172–173, 212
communication in, 49–49 Toyota Production System, 133, 134
composition of team in, 45–47, 60 Tracer methodology, 212
core members in, 45, 46, 60 Training, 11
identification of underlying causes in, 81–82 as causal factor, 84
leaders in, 45, 47, 48–49 failures related to, 69, 70
meetings in, 47–49 and geriatric adverse events in VHA, 178, 180, 183, 184
mission of, 47 in retained surgical item, 202
no-blame environment in, 48 in risk reduction strategies, 93
in retained surgical item, 195–196 Treatment delay, 44
rules in, 47–48, 49 improvement actions in, 101
in VHA, 177 information collected on, 55
Team Strategies and Tools to Enhance Performance and literature review on, 59
Patient Safety (TeamSTEPPS), 198, 200 multiple root causes in, 87
Telephone interviews, 57 reporting mechanism for, 53
Threat vs opportunities matrix, 135, 212 team investigation of, 47, 53
Time frame
for improvement actions, 104–105, 106–107, 123 U
for root cause analysis, 24, 29, 52 Uncontrollable factors, 14, 82, 84
To Err is Human: Building a Safety Health System, vii, 191 questions in analysis of, 83
Tools and techniques, x, 59, 133–173 Underlying causes, 5, 68, 212
affinity diagrams, 71, 135, 136–137, 205 identification of, 81–82, 89
brainstorming, 138–139. See also Brainstorming Universal Protocol for Preventing Wrong Site, Wrong
capability charts, 135, 140, 205 Procedure, Wrong Person Surgery, 75, 94
change analysis, 59, 68, 135, 141, 205 Unsafe or hazardous conditions, 18, 19, 207
change management, 84, 113, 135, 142, 205 Utility systems, 46, 58, 212
check sheets, 135, 143–144, 205
comparison of, 135
223
V
Vaginal sponge, as retained surgical item, 194–202
Value stream mapping, 135, 172–173, 212
Variations, viii, 6, 213
common-cause, 6, 82, 213
special-cause, 6, 82, 213
Veterans Health Administration (VHA)
geriatric adverse events in, 175–184
National Center for Patient Safety, 97, 101, 175–184
Video recordings, 54
W
Witness statements and observations, 55–57
Womack, James, 133
Work of organization, functions and processes in, 101
Work plan for sentinel event investigation, 52, 53
completion dates in, 52, 53
worksheet on, 62
Written responses to questions on event, 57, 63–65
Wrong-site surgery
immediate changes after, 75
multiple root causes in, 87
Pareto chart on, 160
physical evidence on, 58
risk points and risk reduction strategies in, 96
universal protocol for preventing, 75, 94
WWW (who, what, when) tool, 9
224

Root Cause Analysis in Health Care

Uploaded by

Copyright:

Available Formats

Root Cause Analysis in Health Care

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

Root Cause Analysis in Health Care

Uploaded by

Copyright:

Available Formats

Includes Flash Drive!

Root Cause Analysis in

CHAPTER 1: Root Cause Analysis: An Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

CHAPTER 2: Addressing Sentinel Events in Policy and Strategy . . . . . . . . . . . . . . . . . 17

CHAPTER 3: Preparing for Root Cause Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

CHAPTER 4: Determining Proximate Causes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

CHAPTER 5: Identifying Root Causes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

CHAPTER 6: Designing and Implementing an Action Plan for Improvement . . . . . . . 91

CHAPTER 7: Tools and Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133

CHAPTER 8: Root Cause Analysis Case Studies from the Field . . . . . . . . . . . . . . . . . 175

7. San Jose Mercury News. State Fines Marin General Hospital

Investigating Patient Safety Events: sentinel event as an unanticipated occurrence involving

an RCA is on a positive, preventive approach to system Sidebar 1-1.

product problems before they occur. It provides a look not

Figure 1-1. Frequency/Severity Matrix for Prioritizing Safety-Related Problems

Figure 1-2. Swiss Cheese Model

Sidebar 1-2. those events or possible events with a significant negative

1. Illegible handwritten prescriptions Figure 1-3. The Relationships Between a

To meet these challenges, organizations must under-

Conducting an RCA has significant resource implications.

Sidebar 1-3. reduction activities. This information may benefit organiza-

Figure 1-4. Root Cause Analysis Evaluation Checklist

Root Cause Analysis Evaluation Checklist

Root Cause Analysis Team Participants:

Evaluation Level 1: Acceptability

Evaluation Level 2: Thoroughness

Figure 1-4. Root Cause Analysis Evaluation Checklist (continued)

Evaluation Level 2: Thoroughness (continued)

Evaluation Level 3: Credibility

TIP Sidebar 1-4. Get Support

A Framework for a Root Cause Analysis and Action Plan

Equipment factors How did the equipment

Other Are there any other factors

A Framework for a Root Cause Analysis and Action Plan

A Framework for a Root Cause Analysis and Action Plan

A Framework for a Root Cause Analysis and Action Plan

If after consideration of such a finding, a decision Action Item #2:

Consider whether pilot testing of a planned Action Item #4:

Action Item #6:

Action Item #7:

Action Item #8:

• Understand the differences among types of patient safety events

R oot cause analysis (RCA) plays a key role in the

• Corrective actions did not reach the patient

The Joint Commission’s Sentinel Event

Sidebar 2-2. Examples of Sentinel Events

Figure 2-1. Sentinel Event Process Flow

2) Create incident report

b. Events meeting reviewable criteria

6) SE MOS reviewed for sustainability of the improvement efforts

7) Information de-identified and entered into the Sentinel Event Database

Voluntary Reporting of Sentinel Events to

Sidebar 2-4. State and Federal Reporting Requirements

California New Hampshire

NASHP made the following findings:

the first tenet. However, among other reasons, The Joint

Figure 2-2. Sentinel Event Reporting Tool

Requirements Related to Sentinel Events

Figure 2-3. Contractual Agreement Example

Figure 2-3. Contractual Agreement Example (continued)

b. Events meeting reviewable criteria

6) SE MOS reviewed for sustainability of the improvement efforts