Journal of Psychosomatic Research xxx (xxxx) xxx

Contents lists available at ScienceDirect

Journal of Psychosomatic Research

journal homepage: www.elsevier.com/locate/jpsychores

Short-term psychodynamic psychotherapy for functional somatic disorders:

A systematic review and meta-analysis of within-treatment effects
Allan Abbass a, *, Mark A. Lumley b, Joel Town a, Hannah Holmes b, c, Patrick Luyten d, e,
Angela Cooper a, Leo Russell f, Howard Schubiner g, Celine De Meulemeester c, Steve Kisely a, h
a
Centre for Emotions and Health, Dalhousie University, Halifax, Canada
b
Department of Psychology, Wayne State University, Detroit, USA
c
Department of Psychology, Appalachian State University, Boone, NC, USA
d
University Leuven, Belgium
e
University College London, UK
f
Devon Partnership NHS Trust, UK
g
Ascension Providence Hospital and Michigan State University College of Human Medicine, Southfield, MI, USA
h
University of Queensland, Australia

A R T I C L E I N F O A B S T R A C T

Keywords: Objective: A recent meta-analysis of 17 randomized, controlled trials (RCTs) showed that Short-term Psychody­
Short-term psychodynamic psychotherapy namic Psychotherapy (STPP) for functional somatic disorders (FSD) reduced somatic symptoms compared to wait
Somatoform disorders list, minimal treatment, and treatment-as-usual controls. A clinically important yet unanswered question is how
Somatization
much improvement patients experience within STPP treatment.
Somatic symptom disorders
Medically unexplained symptoms
Methods: Following a systematic search, we identified STPP trials presenting data at baseline and post-treatment/
Functional somatic disorders follow-up. Meta-analyses determined the magnitude of changes in somatic symptoms and other outcomes from
Within-treatment effects before to after STPP, and analyses examined effect sizes as a function of study, therapy, and patient variables.
Systematic review Results: We identified 37 trials (22 pre-post studies and 15 RCTs) totaling 2094 patients treated an average of
Meta-analysis 13.34 sessions for a range of FSD. Across all studies, somatic symptoms improved significantly from pre-
treatment to short-term follow-up with a large effect size (SMD = − 1.07), which was maintained at long-term
follow-up (SMD = − 0.90). After excluding two outlier studies, effects at short- and medium-term follow-up
remained significant but were somewhat reduced in magnitude (e.g., short-term SMD = − 0.73). Secondary
outcomes including anxiety, depression, disability, and interpersonal problems had medium to large effects.
Effects were larger for studies of STPP that were longer than 12 sessions or used an emotion-focused type of
STPP, and for chronic pain or gastrointestinal conditions than for functional neurological disorders.
Conclusions: STPP results in moderate to large improvements in multiple outcome domains that are sustained in
long-term follow-up. STPP is an effective treatment option for FSD and should be included in treatment
guidelines.

1. Introduction reflect the complex interaction of biological and psychosocial factors

and the integration of bodily and brain functions and dysfunctions [15].
Functional somatic disorders (FSD) have been variously labeled over Although FSD are neither purely somatic nor purely mental, psychoso­
the years and have included diagnoses such as most somatoform, psy­ cial trauma, intrapsychic conflicts, and disturbed emotion regulation are
chophysiological, psychosomatic, and somatic symptom disorders, as elevated in these disorders and believed to contribute substantially to
well as “medically unexplained” symptoms. A recent consensus defini­ them [9,29,35]. These conditions are very common in health care set­
tion views functional somatic disorders as an umbrella term that in­ tings, resulting in substantial burdens to patients and health care sys­
cludes conditions characterised by persistent and troublesome physical tems [20,41,58]. Outside of treatments like medications, rehabilitation,
symptoms that are accompanied by impairment or disability and that behavioral interventions and physical therapies, psychotherapies play a

* Corresponding author at: Rm 7508, 5909 Veteran’s Memorial Lane, Halifax, NS B3H 2E2, Canada.
E-mail address: [email protected] (A. Abbass).

https://doi.org/10.1016/j.jpsychores.2021.110473
Received 15 December 2020; Received in revised form 23 March 2021; Accepted 24 March 2021
Available online 26 March 2021
0022-3999/© 2021 Elsevier Inc. All rights reserved.

Please cite this article as: Allan Abbass, Journal of Psychosomatic Research, https://doi.org/10.1016/j.jpsychores.2021.110473
A. Abbass et al. Journal of Psychosomatic Research xxx (xxxx) xxx

prominent role in the treatment of FSD. The most commonly studied change in patients from before to after treatment (i.e., pre-post or
psychological approaches for FSD are cognitive-behavioral and related within-treatment effect sizes). Most between-condition meta-analyses of
interventions, and meta-analytic reviews of these interventions reveal RCTs use only the post-treatment or follow-up data rather than change
varying but often small effects, both for specific syndromes like chronic from baseline, due in part because of the unknown test-retest reliability
pain [78] and for mixed FSD populations [56,74]. Therefore, expansion of the outcome measures [22]. Data on change over time within treat­
of treatment options for this population is needed. ment can be extracted from the treatment arms of RCTs, and impor­
Short-term psychodynamic psychotherapy (STPP) describes a class tantly, the literature usually has numerous “pre-post,” “clinical cohort,”
of related therapies that typically last 40 or fewer sessions and share a or “uncontrolled naturalistic” studies for a given treatment. Such studies
focus on emotional and relational processes linked to development, likely reflect actual clinical practice better than RCTs, but they usually
unresolved conflicts, and past adverse experiences. Most STPP ap­ are excluded from meta-analyses.
proaches are guided by conceptual frameworks such as the 2-triangles Studies rarely report how pre-post or within-treatment effect sizes
model [54,55] and emphasize unconscious processes (thoughts, fanta­ compare to between-condition RCT effect sizes. There are three possi­
sies, and feelings) tied to adverse life events. The range of techniques bilities. First, it is certainly possible that a treatment’s pre-post effect is
used in STPP include support, interpretation, clarification of intrapsy­ larger than its controlled effect because the latter removes change that
chic patterns, challenges to defenses, and eliciting the experience and occurs due to non-treatment factors. Second, the pre-post effect may be
expression of feelings related to conflicted relationships in the past and similar in magnitude to the controlled effect when no change occurs in
present. Different types of STPP vary in their focus; for example, some the control condition, as might happen with highly stable disorders.
therapies (e.g., Psychodynamic-Interpersonal Therapy; PIT), target pri­ Finally, the pre-post effect may be smaller than the controlled effect if,
marily patient insight or understanding of intrapsychic and interper­ for example, higher quality therapy is provided in RCTs than naturalistic
sonal conflicts, whereas other therapies (e.g., Intensive Short-term studies due to better therapist training, treatment adherence, supervi­
Psychodynamic Therapy; ISTDP; Emotional Awareness and Expression sion, or even more homogeneous patient selection. Therefore, it is
Therapy; EAET) are emotion-focused, emphasizing in-session emotional important to evaluate pre-post effect sizes and compare them with
activation, experiencing, and expression. controlled, between-condition effect sizes from RCTs.
The efficacy of STPP has been studied in over 250 randomized In summary, meta-analyses of pre-post comparisons offer a comple­
controlled trials (RCTs) [47], and reviews and meta-analyses of these mentary perspective to meta-analyses of RCTs of a treatment’s effects. In
trials conclude that STPP improves numerous conditions, including this paper, we complement our review of controlled effects of STPP for
depression [24], anxiety [38], personality disorders [72], and common FSD [6] by examining the uncontrolled effects of STPP for FSD over
mental disorders [7]. We recently examined the efficacy of STPP for time—from pre-treatment to follow-up. We obtained data not only from
FSD, conducting a meta-analysis of 17 RCTs [6]. Compared to minimal the treatment arms of RCTs analyzed in Abbass et al. [6] but also from a
or no treatment (i.e., waitlist, treatment-as-usual, or minimal contact), larger number of pre-post, naturalistic, and non-randomized trials of
STPP resulted in lower somatic symptoms, with large effect sizes STPP for FSD. In addition to determining the effect size of STPP on so­
(standardized mean difference, SMD) at post-treatment (SMD = − 0.84) matic symptoms (primary outcome) and many secondary outcomes,
and long-term follow-up (6 or more months, SMD = − 1.00), and analyses also tested predictors of treatment effect sizes on somatic
generally large effects on various secondary outcomes. symptoms at short-term follow-up (where we have the greatest number
Our meta-analysis of STPP for FSD, however, examined only RCTs of studies), including whether or not the data came from an RCT, various
and used only post-intervention data, comparing STPP to controls. Such methodological features of the studies, the length of therapy, the specific
between-condition effects from RCTs (“controlled effects”) are viewed as type of STPP, whether or not therapy was emotion-focused, and the type
the gold standard for intervention meta-analyses. It is important, how­ of FSD (e.g., chronic pain, gastrointestinal, neurological). We also con­
ever, to consider what such meta-analyses do—and do not—accomplish. ducted sensitivity analyses to address concerns about unknown pre-post
Between-condition effect sizes from RCTs provide the most theoretically reliability of measures as well as other potential confounds.
and scientifically valuable information about the unique or specific ef­
fects of a treatment. They accomplish this by “subtracting” the effects of 2. Methods
the control condition and its many non-specific factors, such as repeated
assessment, trial participation, and especially naturalistic change over 2.1. Study registration
time.
There are, however, limitations to RCTs and the meta-analyses that Our research plan was published on the PROSPERO website
summarize them. Generalizability to the larger population of treatment- (PROSPERO 2017 CRD42017083235) prior to commencing this study.
seeking patients and front-line practitioners is limited due to selection We followed the Preferred Reporting Items for Systematic Reviews and
bias into RCTs (e.g., willingness to be randomized) as well as the unique Meta-Analyses (PRISMA) recommendations for the background, search
characteristics and context of those providing therapy. Data from RCTs strategy, methods, results, discussion, and conclusions [32].
may also be biased by negative reactions to being assigned to the control
condition as well as variations in what is offered to controls. For 2.2. Selection criteria
example, “waitlist controls” may artificially enhance treatment effects
[60], and “minimal contact” or “treatment-as-usual” controls vary We searched for all studies using STPP to treat adult patients with
widely among trials and among patients within trials. Yet, most meta- FSD. STPP was defined as a treatment that: 1) is verbal, in-person,
analyses of RCTs collapse these control conditions, and meta-analyses provided in either individual or group formats, and in any setting; 2)
may also include various “active controls” into their overall effect size targets psychodynamic processes and is informed by major developers of
estimates. Such variation in the control / comparison conditions com­ STPP (e.g., David Malan, Habib Davanloo, James Mann, Peter Sifneos,
plicates interpretation of the effect sizes obtained from between- among others); and 3) is 40 or fewer standard-length sessions, but more
condition meta-analyses of RCTs. than a single session. With respect to FSD, we included studies of DSM-IV
In addition, there is a question of great clinical importance that re­ somatoform disorders, pain disorders, and other conditions that would
mains unanswered by between-condition RCT meta-analyses. Patients likely meet DSM-5 criteria for a somatic symptom and related disorder,
and clinicians typically wish to know how much improvement occurs such as irritable bowel syndrome. Importantly, we excluded studies of
when receiving a given treatment, not how much improvement relative conditions with known structural pathological or disease processes such
to some control condition. This clinical question is best answered not by as cancer, cardiovascular disease, or autoimmune disease. In addition,
randomization and comparison controls, but rather by examining included studies had to provide usable data from both baseline and post-

2
A. Abbass et al. Journal of Psychosomatic Research xxx (xxxx) xxx

treatment/follow-up and include at least one of our study outcome Supplement Fig. 1), which notes the representativeness of the sample,
variables. ascertainment of diagnosis, comparability to any controls and mea­
surement of outcomes. Ratings were done independently by a pair of
2.3. Search strategy reviewers, and differences in ratings were discussed to reach consensus.

We updated an earlier search [3,4] that covered the published 2.6. Data analyses
literature to 2008; for the current review, we searched for all studies (no
language restriction) published from January 2006 through May 2020. Separate meta-analyses were conducted for each outcome measure
All studies from the recent review [6] were evaluated for inclusion in and at each of the three follow-up time points, when data were available
this review. We searched the following data bases: PubMed, Web of from at least two studies for the outcome and time point. Effect sizes
Science, EMBASE, the Cochrane Central Register of Controlled Trials using SMD were calculated using RevMan, WinPepi and Comprehensive
(CENTRAL) and PsycINFO. Combinations of the following terms were Meta-analysis. We defined effect sizes as small (SMD of 0.20–0.49),
used: 1) psychotherapy, psychoanalytic, psychodynamic, dynamic, or medium (SMD of 0.5–0.79) and large (SMD of ≥0.8) [37]. Significance
short-term therapy; AND 2) clinical trial, randomized controlled trial, or was assessed using 95% confidence intervals, and heterogeneity by
naturalistic study; AND 3) a long list of various FSD symptoms and using I2 statistic. A value of greater than 50% for the I2 statistic indicates
conditions (chest pain, pain, somatoform disorder, medically unex­ heterogeneity. The random-effects model was used for all the analyses
plained symptoms, psychogenic pain, conversion disorder, somatosen­ because we could not definitively exclude between-study variation even
sory disorder, urethral syndrome, fibromyalgia, functional neurological in the absence of statistical heterogeneity.
disorder, functional movement disorder, psychogenic non-epileptic We adjusted for the fact that pre-test and post-test scores are not
seizures, non-epileptic attack disorder, headache, migraine, irritable independent from each other by using the correlation between the two
bowel, dyspepsia, dermatitis, inflammatory dermatosis, laryngospasm, when original studies reported such values. Such correlations were
pharyngospasm, hysteria, hypochondriasis, tics, Tourette’s, tinnitus, rarely available, however, so we used a default correlation of r = 0.59,
temporomandibular syndrome, bruxism, abdominal pain, leg pain, foot which is the median within-group correlation reported from a meta-
pain, back pain, muscle tension, muscular disorder, muscle strain, arm analysis of 811 correlations stemming from 123 intervention trials [12].
pain, hand pain, chronic fatigue syndrome, fatigue, alexithymia, somatic Note that two of the 37 studies in this review—both conducted by
symptom disorder, somatization disorder, functional somatic symptom, one team [16,17]. Both of these studies reported very high STPP-related
functional somatic syndrome, functional somatic disorder). In addition, reductions in somatic symptoms (over 5 SDs) as well as outlying values
we searched prospective trial registries for unpublished ongoing for reductions in depression, anxiety, and general psychiatric symptoms
research (e.g., http://www.controlled-trials.com, https://clinicaltrials. at the two time points they were assessed (short-and medium-term); no
gov/) and an internet database of controlled and comparative other studies even approached such extreme values. Therefore, we
outcome studies on psychological treatments of somatic symptom dis­ present results for both the full sample of studies and also after excluding
orders (http://www.psychotherapyrcts.org). these two outlier studies from short- and medium-term follow-up. Note
that long-term follow-up effects did not change because these two
2.4. Selection and data extraction studies did not have long-term follow-up data.
We undertook subgroup analyses of studies to determine correlates
Two reviewers (PL, CD) screened titles and abstracts to confirm of effect sizes. In particular, we examined effect sizes based on several
eligibility. Full-text versions of studies were then examined for inclu­ features: whether or not the study was an RCT, had adherence ratings,
sion/exclusion by pairs of reviewers (PL/AA and JT/LR). Disagreement used video or audio review, used a manual for therapy; whether or not
between authors was discussed toward reaching consensus; when therapy was more than 12 sessions and was emotion-focused; the spe­
consensus could not be reached, a third author (SK) was consulted. cific type of STPP; and the type of FSD treated. To have a large enough
Descriptive data from selected studies were extracted and tabulated sample size for reliable inferences, these subgroup analyses were con­
by pairs of reviewers. These data included type of FSD (pain, gastroin­ ducted on only the 24 studies that assessed the primary outcome of
testinal, neurological, or mixed somatic symptoms), the study design somatic symptoms and only at the short-term follow-up.
(RCT or pre-post/uncontrolled), number and gender of patients We undertook several sensitivity analyses. Given that meta-analyses
receiving STPP, type of STPP, treatment duration, and follow-up time- of within-condition data can be biased by the fact that the correlations
points. Reviewers also recorded whether or not the therapy was video/ between pre and post measures of the outcome variable are usually
audio recorded, manualized, evaluated for adherence, and focused on unknown but are often lower than r = 0.59 (our default value), we also
emotion (vs. insight). tested a low (r = 0.2) test-retest value for the 24 studies that assessed
Raw data for effect sizes for each outcome measure were extracted somatic symptoms at short-term follow-up. We also explored any het­
separately by a reviewer (HH) who has no affiliation with STPP. Data erogeneity further through sensitivity analyses of the effect of omitting
entry was spot checked by two others (AA, SK). For the current analyses, each study in turn. When multiple measures were used for the same
pre-treatment and available post-treatment / follow-up data were outcome, we examined the effect of substituting one for the other. A few
extracted for the STPP condition only. The outcome categories were as of the studies have small overlap of patients for certain outcome mea­
follows: somatic symptoms (primary outcome), anxiety, depression, sures [2,26,46,65], so we assessed the effect of omitting each study in
general symptoms, interpersonal problems, physical function, disability, turn. Finally, we tested for publication bias for our primary outcome
quality of life, health care use, and health care cost. Post-intervention (somatic symptoms) using funnel plot asymmetry (where low p values
outcomes were categorized into three time-points: short-term (< 3 suggest publication bias).
months), medium-term (3 to 6 months), and long-term (> 6 months).
3. Results
2.5. Quality ratings
3.1. Description of included studies
Given that two different study designs were included in this meta-
analysis, two quality assessment approaches were used. For RCTs, we Our search identified 546 titles through bibliographic databases and
used the Cochrane Risk of Bias tool to rate the methodological charac­ 267 through other sources such as the ISRCTN trial registry (Online
teristics of studies. For the pre-post/case series studies, we used a Supplementary Fig. 2). After screening and full-text review, we retained
modified version of the Newcastle Ottawa Rating Scale [76] (See Online a total of 37 studies including 2094 patients receiving STPP (Table 1).

3
A. Abbass et al. Journal of Psychosomatic Research xxx (xxxx) xxx

Table 1
Description of studies.
1st Author, year Patient Group n STPP Sessions Longest RCT Adherence Video/ Manual or Emotional < 12
[reference] Model follow-up rated audio guide experience sessions
(Months) review

Abbass 2008 [2] Headache 29 ISTDP 19.7 36 No No Yes Yes Yes No

Abbass 2009 MUS in emergency 50 ISTDP 3.8 12 No Yes Yes Yes Yes Yes
[34]
Alessiani 2020 Chronic migraine 96 Unclear 12 12 No No No No No Yes
[10]
Alteri 2009 [11] Headache 13 Unclear 8 12 No No No No No –
Bassett 1985 Chronic pain 14 Unclear 12 12 Yes No Yes No No No
[13]
Burger 2016 Chronic pain 72 EAET 5 6 No No No Yes Yes Yes
[14]
Chavooshi 2016 Medically 23 ISTDP 20 3 Yes Yes Yes Yes Yes No
[16] unexplained pain
Chavooshi Medically 177 ISTDP 16 3 Yes Yes Yes Yes Yes No
2017a [17] unexplained pain
Chavooshi Medically 42 ISTDP 13 12 Yes Yes Yes Yes Yes No
2017b [18] unexplained pain
Chirco 2015 Bruxism 5 ISTDP 20 12 Yes No Yes Yes Yes No
[19]
Creed 2003 [21] Severe IBS 85 PIT 8 12 Yes Yes No Yes No Yes
Faramarzi 2015 Functional dyspepsia 24 SEP 16 12 Yes No No Yes No No
[25]
Flibotte 2012 Fibromyalgia 67 ISTDP 7.2 Post No Yes Yes Yes Yes Yes
[26]
Hamilton 2000 Chronic dyspepsia 31 PIT 8 12 Yes Yes Yes Yes No Yes
[28]
Hawkins 2004 Chronic back pain 47 ISTDP 8 12 No No No Yes Yes Yes
[30]
Hecke 2008 Psychosomatic 34 SASB 25 12 No No No Yes No No
[31]
Hinson 2006 Functional movement 10 ISTDP 12 Post No No No Yes Yes No
[33] disorders
Junkert-Tress Somatoform disorders 24 TLDP 25 60 No No Yes Yes No No
2001 [36]
Lilliengren 2020 Chronic Pain 228 ISTDP 6.1 36 No Yes Yes Yes Yes Yes
[46]
Limburg 2019 Functional vertigo and 98 Mix ~24 6 No No No No No No
[49] dizziness
Lumley 2008 Fibromyalgia 10 Mix 10 3 No No Yes Yes Yes Yes
[51]
Lumley 2017 Fibromyalgia 79 EAET 8 6 Yes Yes Yes Yes Yes Yes
[53]
Monsen 2000 Chronic pain 20 ACTM 33 12 Yes No No Yes Yes No
[57]
Petoliccho 2017 Chronic migraine 117 Unclear 8 6 No No No No No Yes
[62]
Reuber 2007 Functional 91 PIT 6 6 No No No Yes No Yes
[63] neurological
disorders
Russell 2016 Pseudoseizures 28 ISTDP 3.6 36 No Yes Yes Yes Yes Yes
[64]
Russell 2017 Functional 11 ISTDP 11.7 Post No No No Yes Yes Yes
[65] Neurological
Sattel 2012 [66] Multisomatoform 107 PIT 12 9 Yes Yes Yes Yes No No
disorder
Schaerfert 2013 MUS 170 PIT 12 9 Yes No No Yes No No
[67]
Scheidt 2013 Fibromyalgia with 24 Unclear 25 12 Yes Yes No Yes No No
[68] depression
Selders 2015 MUS 57 DIT 20 Post No No No No No No
[69]
Thakur 2017 IBS 36 EAET 3 2.5 Yes Yes Yes Yes Yes Yes
[71]
Tschuschke Somatoform disorders 50 Unclear 20 12 No No No No No No
2007 [73]
Ventegodt 2008 Somatoform disorders 31 Mix 20 Post No No No No Yes No
[75]
Williams 2018 Functional 44 BAPIT 11.9 Post No No No Yes Yes Yes
[77] neurological disorders
Yarns 2020 [79] Chronic pain 28 EAET/ 8 3 Yes Yes Yes Yes Yes Yes
ISTDP
Yasky 2016 [80] Psychosomatic 22 SEP 15 Post No No No No – No

RCT: Randomized Controlled Trial, MUS: Medically Unexplained Symptoms, ISTDP: Intensive Short-term Dynamic Psychotherapy, PIT: Psychodynamic Interpersonal
Therapy, TLDP: Time Limited Dynamic Psychotherapy, SEP: Supportive Expressive Therapy, SASB: Structural Analysis of Social Behavior, DIT: Dynamic Interpersonal

4
A. Abbass et al. Journal of Psychosomatic Research xxx (xxxx) xxx

Therapy, EAET: Emotional Awareness and Expression Therapy. ACTM: Affect Consciousness Treatment Model, BAPIT: Brief Augmented Psychodynamic Interpersonal
Therapy.

Fifteen (40.5%) of the 37 studies were RCTs of STPP, whereas the other
Table 2
22 (59.5%) were pre-post or naturalistic cohort studies. Several of these
Meta-analyses of studies examining the effects of STPP for functional somatic
22 studies also presented data from a separate, non-randomized com­
disorders.
parison group, but the STPP treatments were essentially a cohort study
and are treated as such in these analyses. Comparison # n SMD [95% CI] Significance
Studies
As shown in Tables 1, 18 studies (48.6%) were of pain-related con­
ditions (chronic pain, fibromyalgia, head pain), 9 studies (24.3%) were Pre to < 3 months
Post-tx
of mixed somatic symptom conditions, 6 studies (16.2%) were of func­
Somatic symptoms 24 1059 − 1.07 [− 1.40, <0.0001
tional neurological disorders, and 4 studies (10.8%) were of functional − 0.74]
gastrointestinal disorders. Most of the studies (k = 28, 75.7%) followed a Depression 16 766 − 1.25 [− 1.72, <0.0001
specific STPP model: 12 studies (32.4%) tested Intensive Short-term − 0.78]
Dynamic Psychotherapy (ISTDP) [8,23]; 5 studies (13.5%) used Anxiety 15 560 − 0.64 [− 0.93, 0.0001
0.35]
Psychodynamic-Interpersonal Therapy (PIT) [34], 4 studies (10.8%)
−
General symptoms 19 866 − 0.85 [− 1.20, <0.0001
tested Emotional Awareness and Expression Therapy (EAET) [52], 2 − 0.50]
studies (5.4%) implemented Supportive Expressive Therapy [50], and 5 Physical 6 235 − 0.98 [− 1.51, <0.0001
studies (13.5%) used other STPP approaches. Nine studies (24.3%) were dysfunction − 0.45]
Interpersonal 6 172 0.66 [− 0.88,
of short-term psychodynamic models that were either linked to multiple − <0.0001
problems − 0.44]
STPP theorists or not linked to a specific STPP theorist but otherwise met Disability 6 176 − 1.07 [− 1.50, <0.0001
criteria as STPP. − 0.64]
Treatments averaged 13.3 sessions (SD = 7.2, range: 3–33), and 22 of Global dysfunction 3 90 − 1.32 [− 1.81, <0.0001
the studies (59.5%) had 12 or fewer therapy sessions, whereas 15 studies − 0.83]
Pre to 3–6 months
(40.5%) had therapy longer than 12 sessions. Most studies (k = 30; Post-tx
80.1%) had follow-up evaluations beyond post treatment, and among Somatic symptoms 13 809 − 0.92 [− 1.27, <0.0001
these studies, the longest follow-up assessment averaged 13.3 months − 0.57]
(SD 12.6, range 2.5–60). Most studies (k = 34, 91.9%) took place in Depression 9 725 − 1.66 [− 2.35, <0.0001
0.97]
outpatient settings, but 3 studies (8.1%) were on inpatient units. Thirty- −
Anxiety 7 455 − 0.64 [− 0.93, <0.0001
one studies (83.8%) provided individual therapy whereas the other 6 − 0.35]
(16.2%) provided group or combination interventions; 20 studies General symptoms 9 559 − 1.05 [− 1.76, 0.003
(54.0%) used an emotion-focused form of STPP. − 0.34]
Physical 4 263 − 0.30 [− 0.44, <0.0001
dysfunction 0.16] (a)
3.2. Study quality
−
Disability 6 176 − 0.96 [− 1.31, <0.0001
− 0.61]
The overall quality of the RCT studies was moderate as determined Global dysfunction 3 90 − 0.81 [− 1.16, <0.0001
by pairwise, independent Cochrane Risk of Bias ratings [20]. Eight of the − 0.46] (a)
Pre to > 6 months
15 RCT studies (53.3%) had blinded measurement of some outcomes (6
Post-tx
did not, 1 unclear), 10 (66.7%) had adequate allocation concealment (4 Somatic symptoms 10 534 − 0.90 [− 1.23, <0.0001
unclear, 1 did not), 11 (73.3%) had random sequence generation (2 did − 0.57]
not, 2 were unclear), and 12 (80.0%) had complete outcome data or Depression 5 341 − 0.66 [− 0.91, <0.0001
adjustments to correct for missing data such as intention to treat − 0.41]
Anxiety 5 341 0.88 [− 1.23, 0.0001
methods (1 did not, 1 unclear). It was not possible to determine if
−
− 0.53]
outcome reporting was complete due to lack of published protocols General symptoms 8 431 − 0.59 [− 0.81, <0.0001
except for 3 studies that did appear complete. Blinding of either thera­ − 0.37]
pists or patients is not possible in psychotherapy research so this was Physical 4 195 − 0.91 [− 1.24, 0.0003
dysfunction 0.58]
rated as absent in each case (Online Supplement Table 1). For the 22 pre- −
Disability 3 62 − 1.71 [− 2.53, <0.0001
post/cohort studies, based on a modified Newcastle Ottawa Rating − 0.89]
System, 16 studies (72.7%) were rated low quality, 4 were rated good, Health care contacts 4 241 − 0.39 [− 0.62, 0.0007
and 2 were rated fair (Online Supplement Table 2). Other measures − 0.17]
revealed variability of study rigour. Most of the 37 studies (k = 28; Health care costs 3 78 − 0.32 [− 0.44, <0.0001
0.20] (a)
75.6%) used a manual to guide therapy, 17 studies (45.9%) had audio/
−

video review, and 14 studies (37.8%) were rated for adherence. Note: (a) I2 ≤ 50%; Negative values of effect estimates favor STPP.

3.3. Effects of STPP on somatic symptoms 3.4. Effects of STPP on secondary outcomes

As shown in Table 2 (and Supplementary Fig. 3), reductions in so­ As shown in Table 2 (and Supplementary Fig. 3), STPP led to sig­
matic symptoms from before to after STPP were large in magnitude at all nificant reductions in depression, anxiety, and general psychiatric
three post-treatment time points: short-term (i.e., less than 3 months, k symptoms at all three follow-up time points. For the full set of studies,
= 24; SMD = − 1.07), medium-term (3 to 6 months, k = 13; SMD = these effects were typically large in magnitude at each time. Removing
− 0.92), and long-term (over 6 months: k = 10; SMD = − 0.90). the two outlier studies reduced the short-term effects to medium in
Excluding the two outlier studies reduced the effect sizes at short-term magnitude on depression (SMD = − 0.74, 95% CI: − 0.96, − 0.53; p <
(SMD = − 0.73, 95% CI = − 0.90, − 0.56, p < .0001) and at medium- .0001), anxiety (SMD = − 0.50, 95% CI: − 0.75, − 0.26; p < .0001), and
term (SMD = − 0.61, 95% CI = − 0.76, − 0.46, p < .0001). The large general symptoms (SMD = − 0.59, 95% CI: − 0.70, − 0.48; p < .0001).
long-term effect remained unchanged at SMD = − 0.90 because the two Similarly, the effects on these outcomes at medium-term follow-up were
outlier studies did not contribute data at long-term.

5
A. Abbass et al. Journal of Psychosomatic Research xxx (xxxx) xxx

somewhat reduced after exclusion of the two outliers: depression (SMD Table 3
= − 0.51, 95% CI: − 0.71, − 0.31; p < .0001), anxiety (SMD = − 0.41, Subgroup analyses of STPP effects on short-term somatic symptoms (including
95% CI: − 0.55, − 0.28; p < .0001), and general symptoms (SMD = and excluding two outlier studies).
− 0.38, 95% CI: − 0.48, − 0.28; p < .0001). Variable # Studies Effect Estimate Significance
Also as shown in Table 2, the effects of STPP on reducing physical Data from RCT?
dysfunction were significant and large at short-term and long-term Yes 13 − 1.52 [− 2.15, <0.0001
follow-up, although small at medium-term. Effects on disability were Yes (no outliers) 11 − 0.89] <0.0001
large at all three time points, and effects on global dysfunction were No 11 − 0.80 [− 1.07, <0.0001
0.52]
large at short- and medium-term follow-ups. Interpersonal problems −
− 0.67 [− 0.90,
were assessed in 6 studies at short-term follow-up only, and the reduc­ − 0.43]
tion was medium in magnitude. Finally, a handful of studies assessed Adherence rated?
STPP effects on health care contacts and costs at long-term follow-up Yes 11 − 1.59 [− 2.23, <0.0001
only; there were significant, small magnitude effects on these outcomes. Yes (no outliers) 9 − 0.94] <0.0001
− 0.77 [− 1.04,
− 0.51]
3.5. Subgroup analyses No 13 − 0.69 [− 0.94, <0.0001
− 0.45]
Meta-analyses of predictors of effect size—analyses of subgroups of Audio/video used?
Yes 12 1.42 [− 2.04,
studies—were conducted only on the primary outcome of somatic − <0.0001
Yes (no outliers) 10 − 0.80] <0.0001
symptoms and only at the short-term follow-up, where the largest − 0.68 [− 0.93,
number of studies were found (k = 24). The results of these subgroup − 0.43]
meta-analyses are shown in Table 3, which presents results both No 12 − 0.78 [− 1.02, <0.0001
including and excluding the two outlier studies. Of note, studies with − 0.53]
Therapy manual used?
and studies without all of the subgroup features had effect sizes that Yes 20 − 1.19 [− 1.57, <0.0001
were significant and typically at least medium in magnitude. Prior to Yes (no outliers) 18 − 0.81] <0.0001
exclusion of the two outliers, studies that were adherence-rated or over No 4 − 0.75 [− 0.94, 0.0004
12 sessions in duration had significantly larger effects than studies − 0.57]
0.57 [− 1.10,
without these features, studies of chronic pain had larger effects than −
− 0.04]
those of neurological conditions, and studies of ISTDP had larger effects Emotion-focused STPP?
than those of PIT. Studies that were RCTs, conducted audio/video re­ Yes 15 − 1.40 [− 1.92, <0.0001
view, used a therapy manual, or were emotion-focused, had numerically Yes (no outliers) 13 − 0.87] <0.0001
larger effects than studies without these features, but not significantly No 8 − 0.79 [− 1.02, <0.0001
0.56]
so, due primarily to the substantial heterogeneity. When the subgroup
−
− 0.65 [− 0.95,
analyses were repeated excluding the two outliers, there were no sig­ − 0.35]
nificant differences as a function of subgroup. However, when based on Therapy >12 sessions?
a clinically meaningful effect that is at least small (> 0.20 SD), therapy Yes 9 − 1.88 [− 2.85, <0.0001
Yes (no outliers) 7 0.90]
longer than 12 sessions was clinically more effective than shorter ther­ − <0.0001
No 15 − 0.98 [− 1.47, <0.0001
apies, studies of chronic pain or gastrointestinal disorders were clini­ − 0.48]
cally larger than those of functional neurological disorders, and both − 0.70 [− 0.89,
ISTDP and EAET yielded clinically larger effects than PIT. − 0.50]
Type of FSD
Chronic pain 13 1.51 [− 2.09,
3.6. Sensitivity analyses, heterogeneity, and publication bias
− <0.0001
Chronic pain (no 11 − 0.93] <0.0001
outliers) 3 − 0.71 [− 0.87, <0.0001
Sensitivity analyses using a correlation of 0.2 between pre and post Gastrointestinal 4 − 0.55] <0.0001
measures of somatic symptoms for the 24 studies at short-term follow-up Neurological 4 − 0.88 [− 1.10, 0.01
Mixed 0.65]
indicated that the effect size changed only slightly, from − 1.07 to − 1.05 −
− 0.48 [− 0.86,
(95% CI: − 1.36, − 071; p < .0001), suggesting the obtained effect sizes − 0.10]
are minimally biased by assuming a correlation of r = 0.59. Sensitivity − 0.35 [− 0.49,
analyses also examined the effect of substituting one measure for − 0.21]
another when multiple instruments were used for the same outcome. Type of therapy
ISTDP 9 1.98 [− 3.03,
The results show that this made little difference to the findings. Omitting
− <0.0001
ISTDP (no outliers) 7 − 0.92] <0.0001
any study with patient overlap with another study also made little dif­ EAET 4 − 0.81 [− 1.07, <0.0001
ference to the findings. PIT 3 − 0.56] 0.0006
There was evidence of heterogeneity (I2 > 50%) in 20 of the 23 − 0.80 [− 1.13,
0.46]
analyses presented in Table 2. When we explored this further through −
− 0.41 [− 0.64,
sensitivity analyses of excluding the two outlier studies [16,17], het­ − 0.18]
erogeneity was no longer significant for medium-term anxiety (I2 =
ISTDP = Intensive Short-term Dynamic Psychotherapy; EAET = Emotional
46%). Similarly, the results for long-term depression were no longer
awareness and Expression Therapy; PIT = Psychodynamic-Interpersonal
heterogeneous (I2 = 0%) on excluding a third study [18]. Removal of
Therapy.
other single studies did not affect heterogeneity.
Finally, we used a funnel plot to assess possible effects of publication
bias on our primary outcome. Egger’s regression asymmetry test was
positive (intercept − 4.55, 90% C.I., − 6.69 to − 2.41, p = .01), indicating non-significant in the case of anxiety (− 2.50, 90% C.I., − 4.23 to − 0.76,
possible publication bias. We did not use trim and fill given this method p = .076) and general psychiatric symptoms (− 4.08, 90% C.I., − 6.68 to
performs poorly in the setting of heterogeneity [32]. We found similar − 1.47, p = .051). When excluding the two outlier studies [16,17], the
results for Egger’s regression asymmetry test in the case of depression test for publication bias was no longer significant for somatic symptoms
(− 7.11, 90% C.I., − 10.41 to − 4.11, p = .006). However, the test was (intercept − 2.00, 90% C.I., − 3.35 to − 0.65, p = .062) and continued to

6
A. Abbass et al. Journal of Psychosomatic Research xxx (xxxx) xxx

not be significant for anxiety and general symptoms: it was only sig­ yielded substantially larger effects on somatic symptoms than non-
nificant for depression. (See Online Supplement Figs. 4–7). Thus the emotion-focused therapies, although this difference was due primarily
bulk of all markers of possible publication bias disappeared with to two outlier studies of a highly emotion-focused therapy, ISTDP. When
removal of 2 outliers. specific types of STPP were compared, however, and after excluding
outliers, the emotion-focused therapies of ISTDP and EAET yielded
4. Discussion clinically larger effects (by 0.40 SD) than did the mostly insight-focused
PIT. The differential effect of these treatment models is consistent with
This meta-analysis of 37 trials of STPP for FSD indicates that STPP findings of a Cochrane review of STPP for common mental disorders [7].
leads to large reductions in somatic symptoms following treatment as Such findings are consistent with a growing literature attesting to the
well as medium or large improvements in most other secondary out­ value of emotional processing [44,45,59] and meta-analyses showing
comes, including depression, anxiety, general psychiatric symptoms, that patients’ emotional expression is a strong predictor of positive
disability, and physical function. Notably, these effects are durable, therapy outcomes [61].
lasting beyond 6 months with no signs of decrement or reversal. FSDs The improved symptoms and functioning following STPP were
commonly result in chronic functional impairment and long-term excess maintained through follow-up beyond 6 months. Meta-analyses of STPP
costs to patients and health and insurance providers. For this reason, the for other disorders have also found sustained or increased benefits in
findings of significant, sustained reductions in health care cost, follow-up [5,7,24,48]. STPP may yield important relational and per­
disability and physical dysfunction are also important. Further, such sonality changes that prevent relapse after treatment [70]. There is some
measures go beyond subjective patient symptom reports, strengthening evidence sustained benefits may be seen in other treatments, including
the evidence in support of STPP. The findings of this meta-analysis CBT [27,40,74].
provide valuable information regarding the clinically relevant ques­ There are limitations of the literature and our review of it. First, the
tion of expected effects from engaging in STPP: patients and providers quality of studies was often subpar, a wide range of outcome measures
can predict substantial and lasting improvements in somatic symptoms was used, most analyses had heterogeneity, and the findings may have
and other outcomes. been influenced by publication bias. Although we tried to minimize
Relatively little is known about how effect sizes from before to after a some of these limitations via sensitivity analyses, results should be
treatment compare to effect sizes obtained from comparisons of treat­ interpreted with caution. Second, we found evidence of possible publi­
ment to randomized control conditions, which is the gold standard meta- cation bias. However, it is important to note that the test for publication
analytic approach. The current analyses found that the large reduction bias was no longer significant on somatic measures after removing the
in somatic symptoms within-STPP is comparable to that obtained when two outlier studies. This is because such outliers inflate the amount of
STPP is compared to no-treatment conditions (waiting list, minimal residual heterogeneity in the meta-analytic distribution. The resulting
contact, or treatment as usual) in a recent meta-analysis [6]. Similarly, increased heterogeneity can be mistakenly attributed to publication bias
the current subgroup meta-analysis found that (after removing two [39]. Third, the current analyses were pre-post and not compared to
outliers) the pre-post effects of STPP when conducted in an RCT were control conditions, which limits conclusions about the specific benefits
similar to the STPP effects from pre-post/naturalistic cohort studies. One of STPP. However, our internal analyses comparing data from natural­
might have expected that the within-treatment effects from uncontrolled istic studies versus RCTs, as well as the separate between-condition
studies would be larger than the between-condition effects from meta-analysis of STPP [6], strengthens the conclusion that STPP has
controlled studies, given that the latter remove improvements due to the large benefits beyond several non-specific factors. Comparisons with
passage of time and other nonspecific factors. Such differences between active controls—not just with no or minimal treatment—are needed to
controlled and uncontrolled effect sizes are expected for many disorders further test the specificity of STPP. Fourth, STPP has several variants, the
that show improvement without treatment. This lack of differences in boundaries distinguishing STPP from other therapies are not definitive,
between-condition and within-STPP effects may be due to FSD being and it is not optimal to classify treatments based on theorists or the brief
relatively stable over several months without treatment, meaning that descriptions of the therapies that are provided in articles. Finally, FSD is
patients in no-treatment control conditions would improve little or not a heterogeneous category, which limits conclusions for any specific
at all. syndrome or disorder. It should be noted, however, that co-morbidity,
FSDs are sometimes subclassified according to the primary organ chronic overlapping conditions, and multiple somatic symptoms are
system or somatic symptom, and our review found that over half of the extraordinarily common [1,42,81], suggesting that an umbrella cate­
FSD populations were of chronic pain, with smaller numbers of gastro­ gory such as FSD has validity and utility.
intestinal, neurological, and mixed presentations. Analyses of these FSD
subtypes suggested that STPP for chronic pain or gastrointestinal dis­ 5. Conclusion
orders had large effect sizes, whereas STPP had only medium effects for
neurological disorders. Although the small number of studies of This systematic review and meta-analysis offer further evidence that
gastrointestinal and neurological disorders suggests caution in inter­ STPP is both an effective and efficacious treatment for diverse functional
pretation, this finding is consistent with clinical observations that and somatic symptom disorders, yielding large magnitude, durable ef­
functional neurological disorders are particularly challenging to treat in fects from before treatment to follow-up beyond 6 months. The effects of
brief therapies, as such patients have significant difficulties regulating STPP for FSD compare quite favorably to effects of cognitive-behavioral
emotions, experiencing cognitive-perceptual disruption with emotional and related interventions for FSD [56,74,78], suggesting that STPP
activation [64,65]. should be included in treatment guidelines for these common clinical
STPP varies in several ways, including duration of treatment and the presentations, and maybe the preferred treatment approach for some
focus of therapy. Consistent with other reviews [43], the current ana­ patients, especially those with chronic pain or functional gastrointes­
lyses suggest that longer therapy—operationalized here as over 12 ses­ tinal disorders. Future research should directly compare STPP to other
sions—yields somewhat larger effects than shorter therapies, although evidence-based approaches for FSD, test individual differences as pre­
even shorter treatments had significant, medium/large magnitude dictors or moderators of treatment outcomes, and examine therapeutic
benefits. Regarding the focus of therapy, an earlier meta-analysis of mechanisms of various treatments for FSD.
STPP for FSD [3,4] suggested that STPPs that focus on emotional acti­
vation and expression yielded larger benefits than STPPs that target Conflicts of interest
primarily insight. The current, much larger meta-analysis finds some
support for this proposal. Therapies rated as relatively emotion-focused None

7
A. Abbass et al. Journal of Psychosomatic Research xxx (xxxx) xxx

Acknowledgements [21] F. Creed, L. Fernandes, E. Guthrie, S. Palmer, J. Ratcliffe, N. Read, et al., The cost-
effectiveness of psychotherapy and paroxetine for severe irritable bowel syndrome,
Gastroenterology 124 (2003) 303–317.
This research was supported by the Dalhousie University Department [22] P. Cuijpers, E. Weitz, I.A. Cristea, J. Twisk, Pre-post effect sizes should be avoided
of Psychiatry and the Nova Scotia Department of Health and Wellness. in meta-analyses, Epidemiol. Psychiatr. Sci. 26 (2017) 364–368.
[23] H. Davanloo, Intensive Short-Term Dynamic Psychotherapy: Selected Papers of
Habib Davanloo, John Wiley, New York, 2000.
Appendix A. Supplementary data [24] E. Driessen, L.M. Hegelmaier, A.A. Abbass, J.P. Barber, J.J. Dekker, H.L. Van, et al.,
The efficacy of short-term psychodynamic psychotherapy for depression: a meta-
Supplementary data to this article can be found online at https://doi. analysis update, Clin. Psychol. Rev. 42 (2015) 1–15.
[25] M. Faramarzi, P. Azadfallah, H.E. Book, K.R. Tabatabai, H. Taherim, M. Kashifard,
org/10.1016/j.jpsychores.2021.110473. The effect of psychotherapy in improving physical and psychiatric symptoms in
patients with functional dyspepsia, Iran. J. Psychiatry 1 (2015) 43–49.
References [26] O. Flibotte, A Crossover Randomized Controlled Trial to Compare the Effectiveness
of Intensive Short-Term Dynamic Psychotherapy in Addition to Treatment as Usual
Versus Treatment as Usual Alone for Treating Fibromyalgia Patients, Honours
[1] L.A. Aaron, D. Buchwald, A review of the evidence for overlap among unexplained
Dissertation, St Mary’s University, Canada, 2012.
clinical conditions, Ann. Intern. Med. 134 (2001) 868–881.
[27] C. Flückiger, A.C. Del Re, The sleeper effect between psychotherapy orientations: a
[2] A. Abbass, D. Lovas, A. Purdy, Direct diagnosis and management of emotional
strategic argument of sustainability of treatment effects at follow-up, Epidemiol.
factors in chronic headache patients, Cephalalgia 28 (2008) 1305–1314.
Psychiatr. Sci. 26 (2017) 442–444.
[3] A. Abbass, S. Campbell, K. Magee, R. Tarzwell, Intensive short-term dynamic
[28] J. Hamilton, E. Guthrie, F. Creed, D. Thompson, B. Tomenson, R. Bennett, et al.,
psychotherapy to reduce rates of emergency department return visits for patients
A randomized controlled trial of psychotherapy in patients with chronic functional
with medically unexplained symptoms: preliminary evidence from a pre-post
dyspepsia, Gastroenterology 119 (2000) 661–669.
intervention study, Can. J. Emerg. Med. 11 (2009) 529–534.
[29] W. Häuser, M. Kosseva, N. Uceyler, P. Klose, C. Sommer, Emotional, physical, and
[4] A. Abbass, S. Kisely, K. Kroenke, Short-term psychodynamic psychotherapy for
sexual abuse in fibromyalgia syndrome: a systematic review with meta-analysis,
somatic disorders: a systematic review and meta-analysis, Psychother. Psychosom.
Arthritis Care Res. 63 (2011) 808–820.
78 (2009) 265–274.
[30] J.R. Hawkins, The Role of Emotional Repression in Chronic Back Pain Patients
[5] A.A. Abbass, S. Rabung, F. Leichsenring, J.S. Refseth, N. Midgley, Psychodynamic
Undergoing Psychodynamically Orientated Group Psychotherapy as Treatment for
psychotherapy for children and adolescents: a meta-analysis of short-term
their Pain, New York University School of Education, NY, 2004.
psychodynamic models, J. Am. Acad. Child Adolesc. Psychiatry 52 (2013)
[31] D. Hecke, J. Hardt, W. Tress, Effectiveness and clinical significance of brief
863–875.
dynamic psychotherapy: the Duesseldorf short-term dynamic psychotherapy
[6] A. Abbass, J. Town, H. Holmes, P. Luyten, A. Cooper, L. Russell, D. Bernier, Short-
project (since 1991), Z. Psychosom. Med. Psychother. 54 (2008) 107–131.
term psychodynamic psychotherapy for functional somatic disorders: a meta-
[32] J.P. Higgins, S. Green, Cochrane Handbook for Systematic Reviews of
analysis of randomized controlled trials, Psychother. Psychosom. 89 (6) (2020)
Interventions, John Wiley & Sons, Hoboken, NJ, 2011.
363–370.
[33] V.K. Hinson, S. Weinstein, B. Bernard, S.E. Leurgans, C.G. Goetz, Single-blind
[7] A.A. Abbass, S.R. Kisely, J.M. Town, F. Leichsenring, E. Driessen, S. De Maat, et al.,
clinical trial of psychotherapy for treatment of psychogenic movement disorders,
Short-term psychodynamic psychotherapies for common mental disorders,
Parkinsonism Relat. Disord. 12 (2006) 177–180.
Cochrane Database of Syst. Rev. 7 (2014), CD004687.
[34] R.F. Hobson, Forms of Feeling: The Heart of Psychotherapy, Tavistock, London,
[8] A. Abbass, Reaching through Resistance: Advanced Psychotherapy Techniques,
1985.
First ed., Seven Leaves Press, Kansas City, MO, 2015.
[35] G.T. Jones, C. Power, G.J. Macfarlane, Adverse events in childhood and chronic
[9] N. Afari, S.M. Ahumada, L.J. Wright, S. Mostoufi, G. Golnari, V. Reis, J.G. Cuneo,
widespread pain in adult life: results from the 1958 British birth cohort study, Pain
Psychological trauma and functional somatic syndromes: a systematic review and
143 (2009) 92–96.
meta-analysis, Psychosom. Med. 76 (2014) 2–11.
[36] B. Junkert-Tress, U. Schnierda, N. Hartkamp, N. Schmitz, W. Tress, Effects of short-
[10] M. Alessiani, B. Petolicchio, R. Di Sanctis, R. Di Giambattista, M. Puma,
term dynamic psychotherapy for neurotic, somatoform, and personality disorders:
C. Franzese, V. Di Piero, A propensity score matching study on the effect of
a prospective 1-year follow-up study, Psychother. Res. 11 (2001) 187–200.
OnabotulinumtoxinA alone versus short-term psychodynamic psychotherapy plus
[37] L.E. Kazis, J.J. Anderson, R.F. Meenan, Effect sizes for interpreting changes in
drug-of-choice as preventive therapy in chronic migraine: effects and predictive
health status, Med. Care 27 (1989) S178–S189.
factors, Res. Square (2020), https://doi.org/10.21203/rs.3.rs-21658/v1.
[38] J.R. Keefe, K.S. McCarthy, U. Dinger, S. Zilcha-Mano, J.P. Barber, A meta-analytic
[11] M. Altieri, R. Di Giambattista, L. Di Clemente, D. Fagiolo, E. Tarolla, A. Mercurio,
review of psychodynamic therapies for anxiety disorders, Clin. Psychol. Rev. 34
et al., Combined pharmacological and short-term psychodynamic psychotherapy
(2014) 309–323.
for probable medication overuse headache: a pilot study, Cephalalgia 29 (2009)
[39] S. Kepes, M.A. Thomas, Assessing the robustness of meta-analytic results in
293–299.
information systems: publication bias and outliers, Eur. J. Inf. Syst. 27 (1) (2018)
[12] E.M. Balk, A. Earley, K. Patel, T.A. Trikalinos, I.J. Dahabreh, Empirical Assessment
90–123.
of Within-Arm Correlation Imputation in Trials of Continuous Outcomes. Methods
[40] D.M. Kivlighan 3rd, S.B. Goldberg, M. Abbas, B.T. Pace, N.E. Yulish, J.G. Thomas,
Research Report. (Prepared by the Tufts Evidence-based Practice Center under
M.M. Cullen, C. Fluckiger, B.E. Wampold, The enduring effects of psychodynamic
Contract No. 290–2007-10055-I.) AHRQ Publication No. 12(13)-EHC141-EF,
treatments vis-a-vis alternative treatments: A multilevel longitudinal meta-
Agency for Healthcare Research and Quality, Rockville, MD, 2012. www.effecti
analysis, Clin. Psychol. Rev. 40 (2015) 1–14.
vehealthcare.ahrq.gov/reports/final.cfm.
[41] K. Kroenke, A practical and evidence-based approach to common symptoms, Ann.
[13] D.L. Bassett, I. Pilowsky, A study of brief psychotherapy for chronic pain,
Intern. Med. 161 (2014) 579–586.
J. Psychosom. Res. 29 (1985) 259–264.
[42] K. Kroenke, J.G. Rosmalen, Symptoms, syndromes, and the value of psychiatric
[14] A.J. Burger, M.A. Lumley, J.N. Carty, D.V. Latsch, E.R. Thakur, M.E. Hyde-Nolan,
diagnostics in patients who have functional somatic disorders, Med. Clin. North
A.M. Hijazi, H. Schubiner, The effects of a novel psychological attribution and
Am. 90 (2006) 603–626.
emotional awareness and expression therapy for chronic musculoskeletal pain: a
[43] M.J. Lambert, The efficacy and effectiveness of psychotherapy, in: M.J. Lambert
preliminary, uncontrolled trial, J. Psychosom. Res. 81 (2016) 1–8.
(Ed.), Bergin and Garfield’s handbook of psychotherapy and behavior change, 6th
[15] C. Burton, P. Fink, P. Henningsen, B. Löwe, W. Rief, EURONET-SOMA Group.
ed., Wiley, New York, 2013, pp. 169–218.
Functional somatic disorders: discussion paper for a new common classification for
[44] R.D. Lane, C. Subic-Wrana, L. Greenberg, I. Yovel, The role of enhanced emotional
research and clinical use, BMC Med. 18 (1) (2020) 34.
awareness in promoting change across psychotherapy modalities, J. Psychother.
[16] B. Chavooshi, P. Mohammadkhani, B. Dolatshahee, Efficacy of intensive short-term
Integr. (2021), https://doi.org/10.1037/int0000244.
dynamic psychotherapy for medically unexplained pain: a pilot three-armed
[45] R.D. Lane, L. Ryan, L. Nadel, L. Greenberg, Memory reconsolidation, emotional
randomized controlled trial comparison with mindfulness-based stress reduction,
arousal, and the process of change in psychotherapy: new insights from brain
Psychother. Psychosom. 85 (2016) 123–125.
science, Behav. Brain Sci. 38 (2015), e1, https://doi.org/10.1017/
[17] B. Chavooshi, M. Saberi, S.A. Tavallaie, H. Sahraei, Psychotherapy for medically
S0140525X14000041.
unexplained pain: a randomized clinical trial comparing intensive short-term
[46] P. Lilliengren, A. Cooper, J.M. Town, S. Kisely, A. Abbass, Clinical- and cost-
dynamic psychotherapy and cognitive-behavior therapy, Psychosomatics 58
effectiveness of intensive short-term dynamic psychotherapy for chronic pain in a
(2017) 506–518.
tertiary psychotherapy service, Aust. Psychiatry (2020), https://doi.org/10.1177/
[18] B. Chavooshi, P. Mohammadkhani, B. Dolatshahee, Telemedicine vs. in-person
1039856220901478.
delivery of intensive short-term dynamic psychotherapy for patients with
[47] P. Lilliengren, Comprehensive Compilation of Randomized Controlled Trials
medically unexplained pain: a 12-month randomized, controlled trial, J. Telemed.
(RCTs) Involving Psychodynamic Treatments and Interventions. https://www.
Telecare 1 (2017) 133–141.
researchgate.net/publication/317335876, 2019.
[19] S.M. Chirco, A. Bargnani, Bruxismo, Ansia e Psicoterapie Brevi Ricerca e Risultati,
[48] P. Lilliengren, R. Johansson, K. Lindqvist, J. Mechler, G. Andersson, Efficacy of
Psyche Nova 2015-2016 (2015) 33–43.
experiential dynamic therapy for psychiatric conditions: a meta-analysis of
[20] H. Cramer, R. Lauche, G. Dobos, Somatoform disorders and medically unexplained
randomized controlled trials, Psychotherapy 53 (2016) 90–104.
symptoms in primary care: a systematic review and meta-analysis of prevalence,
[49] K. Limburg, G. Schmid-Mühlbauer, H. Sattel, A. Dinkel, K. Radziej, M. Gonzales,
Dtsch. Arztebl. Int. 112 (2015) 279–287.
J. Ronel, C. Lahmann, Potential effects of multimodal psychosomatic inpatient
treatment for patients with functional vertigo and dizziness symptoms: a pilot trial,
Psychol. Psychother. Theory Res. Pract. 92 (2019) 57–73.

8
A. Abbass et al. Journal of Psychosomatic Research xxx (xxxx) xxx

[50] L. Luborsky, Principles of Psychoanalytic Psychotherapy: A Manual for Supportive- [67] R. Schaefert, C. Kaufmann, B. Wild, D. Schellberg, R. Boelter, R. Faber, et al.,
Expressive Treatment, Basic Books, New York, 1984. Specific collaborative group intervention for patients with medically unexplained
[51] M.A. Lumley, J.L. Cohen, R.L. Stout, L.C. Neely, L.M. Sander, A.J. Burger, An symptoms in general practice: a cluster randomized controlled trial, Psychother.
emotional exposure-based treatment of traumatic stress for people with chronic Psychosom. 82 (2013) 106–119.
pain: preliminary results for fibromyalgia syndrome, Psychother. Theory Res. [68] C.E. Scheidt, E. Waller, K. Endorf, S. Schmidt, R. König, A. Zeeck, et al., Is brief
Pract. Train. 45 (2008) 165–172. psychodynamic psychotherapy in primary fibromyalgia syndrome with concurrent
[52] M.A. Lumley, H. Schubiner, Emotional awareness and expression therapy for depression an effective treatment? A randomized controlled trial, Gen. Hosp.
chronic pain: rationale, principles and techniques, evidence, and critical review, Psychiatry 2 (2013) 160–167.
Curr. Rheumatol. Rep. 21 (2019) 8. [69] M. Selders, R. Visser, W. van Rooij, G. Delfstra, J.A. Koelen, The development of a
[53] M.A. Lumley, H. Schubiner, N.A. Lockhart, K.M. Kidwell, S.E. Harte, D.J. Clauw, et brief group intervention (dynamic interpersonal therapy) for patients with
al., Emotional awareness and expression therapy, cognitive behavioral therapy, medically unexplained somatic symptoms: a pilot study, Psychoanal. Psychother.
and education for fibromyalgia: a cluster-randomized controlled trial, Pain 158 29 (2015) 182–198.
(2017) 2354–2363. [70] J. Shedler, The efficacy of psychodynamic psychotherapy, Am. Psychol. 65 (2010)
[54] D.H. Malan, Individual Psychotherapy and the Science of Psychodynamics, 98–109.
Butterworths, London, Boston, 1979. [71] E.R. Thakur, H.J. Holmes, N.A. Lockhart, J.N. Carty, M.S. Ziadni, H.K. Doherty, et
[55] L. McCullough, M. Magill, Affect-focused short-term dynamic therapy: Empirically al., Emotional awareness and expression training improves irritable bowel
supported strategies for resolving affect phobias (249-277), in: R.A. Levy, J. syndrome: a randomized controlled trial, Neurogastroenterol. Motil. 29 (2017),
S. Ablon (Eds.), Handbook of Evidence-Based Psychodynamic Psychotherapy, e13143.
Humana Press, New York, NY, 2009. [72] J.M. Town, A. Abbass, G. Hardy, Short-term psychodynamic psychotherapy for
[56] V. Menon, T.M. Rajan, P.P. Kuppili, S. Sarkar, Cognitive behavior therapy for personality disorders: a critical review of randomized controlled trials, J. Personal.
medically unexplained symptoms: a systematic review and meta-analysis of Disord. 25 (2011) 723–740.
published controlled trials, Indian J. Psychol. Med. 39 (2017) 399–406. [73] V. Tschuschke, T. Anbeh, P. Kiencke, Evaluation of long-term analytic outpatient
[57] K. Monsen, J.T. Monsen, Chronic pain and psychodynamic body therapy: a group therapies, Group Anal. 40 (2007) 140–159.
controlled outcome study, Psychotherapy 37 (2000) 257–269. [74] N. van Dessel, M. den Boeft, J.C. van der Wouden, M. Kleinstauber, S.S. Leone,
[58] C. Nimnuan, M. Hotopf, S. Wessely, Medically unexplained symptoms: an B. Terluin, et al., Non-pharmacological interventions for somatoform disorders and
epidemiological study in seven specialities, J. Psychosom. Res. 51 (2001) 361–367. medically unexplained physical symptoms (MUPS) in adults, Cochrane Database
[59] A. Pascual-Leone, L.S. Greenberg, Emotional processing in experiential therapy: Syst. Rev. 11 (2014). Cd011142.
why“ the only way out is through.”, J. Consult. Clin. Psychol. 75 (2007) 875–887. [75] S. Ventegodt, I. Kandel, J. Merrick, A study of experienced chronic pain in the
[60] B. Patterson, M.H. Boyle, M. Kivlenieks, M. Van Ameringen, The use of waitlists as holistic medicine clinic using mindful psychodynamic short time psychotherapy
control conditions in anxiety disorders research, J. Psychiatr. Res. 83 (2016) complemented with bodywork, J. Pain Manag. 1 (2008) 55–62.
112–120. [76] G. Wells, B. Shea, D. O’Connell, J. Peterson, V. Welch, M. Losos, P. Tugwell, The
[61] P.R. Peluso, R.R. Freund, Therapist and client emotional expression and Newcastle–Ottawa Scale (NOS) for Assessing the Quality of Non-Randomized
psychotherapy outcomes: a meta-analysis, Psychotherapy 55 (2018) 461–472. Studies in Meta-Analysis, 2000.
[62] B. Petolicchio, A. Vigano, F. Turchi, M. Squitieri, R. Di Giambattista, M. Altieri, et [77] I.A. Williams, S. Howlett, L. Levita, M. Reuber, Changes in emotion processing
al., Non-pharmacological preventive therapy to chronic migraine in the real world: following brief augmented psychodynamic interpersonal therapy for functional
a short-term psychodynamic psychotherapy approach, J. Headache Pain 18 (2017) neurological symptoms, Behav. Cogn. Psychother. 2018 (2018) 1–17.
111. [78] Williams ACDC, E. Fisher, L. Hearn, C. Eccleston, Psychological therapies for the
[63] M. Reuber, C. Burness, S. Howlett, J. Brazier, R. Grunewald, Tailored management of chronic pain (excluding headache) in adults, Cochrane Database
psychotherapy for patients with functional neurological symptoms: a pilot study, Syst. Rev. 8 (2020), https://doi.org/10.1002/14651858.CD007407.pub4.
J. Psychosom. Res. 63 (2007) 625–632. [79] B.C. Yarns, M.A. Lumley, J.T. Cassidy, W.N. Steers, S. Osato, H. Scubiner, D.
[64] L.A. Russell, A.A. Abbass, S.J. Allder, S. Kisely, B. Pohlmann-Eden, J.M. Town, L. Sultzer, Emotional awareness and expression therapy (EAET) achieves greater
A pilot study of reduction in healthcare costs following the application of intensive pain reduction than cognitive behavioral therapy (CBT) in older adults with
short-term dynamic psychotherapy for psychogenic nonepileptic seizures, Epilepsy chronic musculoskeletal pain: a preliminary randomized comparison trial, Pain
Behav. 63 (2016) 17–19. Med. (2020), https://doi.org/10.1093/pm/pnaa145.
[65] L.A. Russell, A.C. Turner, P.J. Yates, A preliminary evaluation of intensive short- [80] J. Yasky, R. King, T. O’Brien, Resistance, early engagement and outcome in
term dynamic psychotherapy within a functional neurological symptoms service, psychoanalytic psychotherapy of patients with psychosomatic disorders, Couns.
Neuropsychologist (2017) 25–44. Psychother. Res. 16 (2016) 266–276.
[66] H. Sattel, C. Lahmann, H. Gundel, E. Guthrie, J. Kruse, M. Noll-Hussong, et al., [81] M.B. Yunus, Fibromyalgia and overlapping disorders: the unifying concept of
Brief psychodynamic interpersonal psychotherapy for patients with central sensitivity syndromes, Semin. Arthritis Rheum. 36 (2007) 339–356.
multisomatoform disorder: randomised controlled trial, Br. J. Psychiatry 200
(2012) 60–67.