Content uploaded by Steve Kisely
Author content
All content in this area was uploaded by Steve Kisely
Content may be subject to copyright.
Fax +41 61 306 12 34
E-Mail karger@karger.ch
www.karger.com
Special Article
Psychother Psychosom 2009;78:265–274
DOI: 10.1159/000228247
Short-Term Psychodynamic
Psychotherapy for Somatic Disorders
Systematic Review and Meta-Analysis of Clinical Trials
Allan Abbass a Stephen Kisely b–d Kurt Kroenke e
a Department of Psychiatry, Centre for Emotions and Health, and Departments of
b Community Health and
Epidemiology and
c Psychiatry, Dalhousie University, Halifax, N.S. , Canada;
d School of Medicine, Griffith University,
Meadowbrook, Qld. , Australia; e Regenstrief Institute, Indiana University School of Medicine, Indianapolis, Ind. , USA
a 54% greater treatment retention in the STPP group versus
controls. Conclusion: STPP may be effective for a range of
medical and physical conditions underscoring the role of pa-
tients’ emotional adjustment in overall health. Future re-
search should include high-quality randomized and clinical
effectiveness studies with attention to healthcare use and
costs. Copyr ight © 2009 S. Karger AG, B asel
Introduction
Half of all outpatient medical visits are related to so-
matic complaints, of which at least one third to one half are
medically unexplained
[1] . Many are individual physical
symptoms, such as pain (e.g. low back, joint, chest, abdom-
inal, headache) and nonpain (e.g. fatigue, dizziness, palpi-
tations) complaints. Others consist of a cluster of somatic
symptoms for which the etiology is poorly understood,
such as irritable bowel syndrome, fibromyalgia, chronic
fatigue syndrome, temporomandibular disorder and in-
terstitial cystitis. These functional somatic syndromes of-
ten overlap and are similar in terms of psychiatric comor-
bidity, functional impairment and family history
[2– 4] .
Distressing somatic symptoms are also increased 2- to
3-fold in patients with depressive and anxiety disorders
[5, 6] . More recently, it has also been shown that disease-
specific somatic symptoms in patients with a variety of
Key Words
Psychotherapy, short-term ⴢ Somatoform ⴢ
Psychotherapy, psychodynamic ⴢ Psychotherapy,
psychosomatic ⴢ Psychophysiologic disorders
Abstract
Background: Somatic symptom disorders are common, dis-
abling and costly. Individually provided short-term psy-
chodynamic psychotherapies (STPP) have shown promising
results. However, the effectiveness of STPP for somatic symp-
tom disorders has not been reviewed. Methods: We under-
to ok a sy ste mati c re view o f rand omi zed co ntro lle d tr ials and
controlled before and after studies. The outcomes included
psychological symptoms, physical symptoms, social-occu-
pational function, healthcare utilization and treatment con-
tinuation. Results: A total of 23 studies met the inclusion cri-
teria and covered a broad range of somatic disorders.
Thirteen were RCTs and 10 were case series with pre-post
outcome assessment. Of the included studies, 21/23 (91.3%),
11/12 (91.6%), 16/19 (76.2%) and 7/9 (77.8%) reported signifi-
cant or possible effects on physical symptoms, psychologi-
cal symptoms, social-occupational function and healthcare
utilization respectively. Meta-analysis was possible for 14
studies and revealed significant effects on physical symp-
toms, psychiatric symptoms and social adjustment which
were maintained in long-term follow-up. Random-effect
modeling attenuated some of these relationships. There was
Recei ved: February 19, 2008
Accepted after revision: July 18, 2008
Publis hed online: July 11, 2009
Dr. Al lan Abbass
5909 Vetera ns Memorial Lane, Room 8203
Abbie J. L ane Building, QEII Hea lth Sciences C enter
Hal ifax, NS, B3H 2E2 (Ca nada)
Tel. +1 902 473 2514, Fax +1 902 473 4545, E -Mail alla n.abbass@da l.ca
© 200 9 S. Karger AG, Basel
0033–3190/09/0785–0265$26.00/0
Accessible online at:
www.karger.com/pps
Abbass /Kisely /Kroenke
Psychother Psychosom 2009;78:265–274
266
medical disorders are influenced as much by psychologi-
cal factors as by the severity of the underlying medical
disorder
[7, 8] . While some patients with medically unex-
plained symptoms meet criteria for somatoform disor-
ders, the boundaries are not always clear-cut between so-
matoform symptoms and the distressing and persistent
somatic symptoms experienced by patients with func-
tional somatic disorders, depression, anxiety and even
some medical conditions
[8] .
The treatment of somatoform disorders and related
conditions manifested by poorly explained somatic symp-
toms has been covered in several recent comprehensive
reviews
[9–15] . Cumulatively, these reviews confirm that
2 of the most evidence-based treatments are cognitive-
behavioral therapy (CBT) and antidepressants. Too few
studies of other treatments were then found to lend them-
selves to a meta-analysis.
Unresolved unconscious emotional issues have long
been considered an important causal factor in a range of
physical illnesses and somatic symptom disorders
[16] . In
clinical practice, psychodynamic psychotherapies focus
on this unconscious process by which emotions translate
into somatic symptoms, somatic focus and, indeed, ob-
jectively measurable physical sequelae.
Short-term psychodynamic psychotherapies (STPP)
are a group of brief therapy methods developed over the
past 50 years by proponents including Mann, Sifneos,
Malan and Davanloo
[17] . Some STPP methods aim for
insight into various unconscious phenomena, while oth-
ers seek to address alexithymia, or difficulty identifying
and experiencing emotions. With these different goals,
technical differences have developed over time, with
some methods being more versus less focused on emo-
tional experiencing. They share the common goals of
making unconscious phenomena conscious and working
through underlying conflicts.
The efficacy of STPP across a range of common men-
tal disorders was reviewed in 2 recent meta-analyses
[18 ,
19] . There are limitations to the generalizability of these
findings to the treatment of somatic disorders. One re-
view only included a single study with somatoform dis-
orders
[18] , and the other excluded studies with formal
psychotherapy treatment controls. Both reviews were re-
stricted to RCTs of individual STPP methods. Thus, the
great majority of all STPP studies for somatic symptom
disorders have never been reviewed. The purpose of this
paper was to critically review and meta-analyze, where
appropriate, data from studies using both RCT and non-
RCT designs in order to examine the effectiveness of
STPP in patients with somatic symptom disorders.
M e t h o d s
Selection of Studies
We included studies of STPP therapies in somatic symptom
disorders covering both medically explained and unexplained
symptoms without regard to the presence of a formal psychiatric
disorder to better reflect the case mix seen in general medical set-
tings. We included both RCTs as well as before and after studies
such as mirror designs of the same subjects. Studies of STPP de-
livered in either individual or group format were included.
Search Strategy
We se arched PsycIn fo from 1967 to t he pre sent, Me dli ne from
1966 to the present and the Cochrane Library from 2005 to the
present up to July 2007. Many papers had been found in a previous
broad search conducted for a Cochrane review of STPP therapies
for mental disorders
[19] . Our strategy included broad searches
with the following terms: psychotherapy, psychodynamic, dy-
namic or short-term therapy and clinical trial, naturalistic study,
or randomized trial and 37 specific terms, such as chest pain, ab-
dominal pain and headache. We searched for further trials by
scrutinizing the reference lists of initial studies identified and
other relevant review papers. We also contacted selected authors
and experts. Two reviewers (A.A. and S.K.) independently ex-
tracted data. Two reviewers collated and independently assessed
abstracts.
Study Description
The studies were reviewed for treatment characteristics, study
methodology, sample characteristics, outcome measures, and re-
ported results on primary indices under the categories psycho-
logical symptoms, somatic symptoms, social-occupational func-
tioning and healthcare utilization. We specifically noted which
studies were ma nualized, which had adherence ratings and which
had blinded ratings of outcome. For RCTs, we used the Cochrane
Collaboration Depression Anxiety and Neurosis (CCDAN) qual-
ity rating scale to numerical ly rate the study quality. This 23-item
scale includes a broad range of indicators such as allocation con-
cealment and sample size and has a maximum value of 46.
M e t a - A n a l y s i s
Where appropriate, we combined the results of the studies us-
ing meta-analysis. We used Review Manager version 4.1, a statis-
tical software package for managing and analyzing a Cochrane
Collaboration systematic review, for our analysis. We divided the
outcomes into short-term (up to 3 months), medium-term (3–9
months) and long-term ( 1 9 months), and measured effect size
(ES) using standardized mean differences (SMD). We defined ES
as small (ES = 0.20–0.49), medium (ES = 0.5–0.79) and large (ES
6 0.8) [20] . We assessed significance using 95% confidence inter-
vals (CI) and heterogeneity with the Q and I
2 statistic. A value
1 50% for the I
2 statistic indicates heterogeneity. We evaluated
publication bias using the fail-safe N statistic. This is the number
of nonsignifica nt studies that would be ne cessary to reduce t he ES
to a neglig ible value of 0.10. This w as calculated apply ing the Win-
Pepi statistical package
[21] .
Short-Term Psychodynamic
Psychotherapy for Somatic Disorders
Psychother Psychosom 2009;78:265–274
267
R e s u l t s
Study Inclusion Criteria and Characteristics
We found 1 100 citations of interest in the initial elec-
tronic searches, of which 33 papers were potentially rel-
evant and subjected to strict eligibility assessment. Of
these, we excluded 8 which did not meet our inclusion
criteria and 2 which were duplicate publications ( fig. 1 ).
The 23 eligible studies included 13 RCTs and 10 pre-post
st udie s. Ei ght een foc used on speci fi c sy mpt oms or symp-
tom clusters, while 5 studied general somatic symptoms
or clusters of disorders. Although 15 studies cited spe-
cific STPP models, only 6 described manualized treat-
ments and 6 noted adherence verification. Nine had
blinded ratings of outcome. The CCDAN quality ratings
averaged 26.5 (SD = 7.3, range 16–36), suggesting moder-
ate study quality. These studies were performed in 10 dif-
ferent countries over the past 25 years.
P a t i e n t s
There were a total of 1,870 subjects (study range = 10–
342), of which 873 (range = 10–87) received STPP and 535
(range = 22–257) served as controls. The investigations
included a mean of 77 (SD = 63) patients. The patients
averaged 41.3 years of age (SD = 10), and 57.8% (SD = 26)
were female.
C o n d i t i o n s
The sample was comprised of 13 different medical
conditions affecting various major systems including
dermatological, neurological, cardiovascular, respirato-
ry, gastrointestinal, musculoskeletal, genitourinary and
immunological systems. Six studies involved patients
with chronic pain. Some included somatic disorders,
such as irritable bowel syndrome and chronic pain, which
are known to have moderately strong associations with
psychological factors. Others considered medical condi-
tions which, though manifested by somatic symptoms,
Total papers yielded
Abstracts searched
electronically for key terms
(n >100)
Not meeting inclusion criteria
(n >70)
Possible inclusion
Abstracts scrutinized in detail
(n = 33)
Not meeting inclusion criteria
(e.g. not an intervention trial)
(n = 8)
Papers scrutinized in detail
(n = 25)
Excluded
(n = 2)
Papers in review
(n = 23)
Excluded from meta-analysis
(e.g. insufficient data)
(n = 9)
Papers in meta-
analysis
(n = 14)
Fig. 1. Number of papers yielded by search
strategy in systematic review.
Abbass /Kisely /Kroenke
Psychother Psychosom 2009;78:265–274
268
are less clearly linked to emotional dysregulation, such as
Crohn’s disease, coronary artery disease, emphysema,
bronchitis and Sjögren’s syndrome.
O u t c o m e s
The majority of all measured outcomes showed bene-
f it s i n e it her RC Ts o r p re-p os t s tu di es . Tw ent y- on e (91 .3%)
reported significant (n = 17) or possible (n = 4) symptom
benefits related to the main physical condition. Eleven of
12 (91.6%) observed significant (n = 9) or possible (n = 2)
social-occupational f unction improvements. Sixteen of 21
(76.2%) found significant (n = 13) or possible (n = 3) psy-
chological symptom benefits. Finally, 7 of 9 (77.8%) re-
ported significant (n = 6) or possible (n = 1) reductions in
healthcare utilization. An outcome possibly worse than
the control was reported in only the bronchitis/emphy-
sema study
[22] on some of the symptom measures. In this
study, more STPP patients had stopped smoking, perhaps
leading to withdrawal, anxiety or depressive symptoms.
Long-term follow-up in this set of studies was the
norm. Nineteen (82.6%) had follow-up of the treated cas-
es. The average duration of follow-up was 19.6 months
(SD = 16) with a range of 1.5 to 60 months.
M e t a - A n a l y s e s
Fourteen studies provided usable data for meta-analy-
ses. We did not include data from a 15th study
[23] as this
was a report on peptic ulcer from 1983, i.e. before the in-
troduction of triple therapy for the eradication of Helico-
bacter pylor i . The remainder either did not have outcomes
fitting our categories or did not present data in a useable
format. The numbers for individual studies vary accord-
ing to the outcome (e.g. depression, anxiety, somatic and
general psychiatric symptoms) and length of follow-up
(e.g. short-, medium- and long-term).
With respect to short-term outcome (0–3 months),
the fixed-effect model showed moderate improvements
(ES = 0.58–0.78) relative to controls for general psychiat-
ric symptoms, depression, anxiety and somatic symp-
toms ( fig. 2 ). All these results were significant. The ran-
dom-effect model yielded similar results except for so-
matic symptoms, where the difference marginally failed
to reach significance (SMD = –0.79, 95% CI = –1.69 to
+0.18; Z = 1.94, p = 0.051).
There were significant differences of at least moderate
magnitude in the medium-term outcome for general psy-
chiatric symptoms (SMD = –0.56, 95% CI = –0.81 to
–0.31; Z = 4.35, p ! 0.0001), depression (SMD = –0.84,
95% CI = –1.34 to –0.35; Z = 3.31, p ! 0.001), anxiety
(SMD = –1.00, 95% CI = –1.51 to –0.50; Z = 3.89, p =
0.0001) and somatic symptoms (SMD = –0.87, 95% CI =
–1.37 to –0.38; Z = 3.45, p ! 0.001) using the fixed-effect
model. The random-effect model produced similar re-
sults for all outcomes.
The difference between intervention and control
groups was maintained in the long-term follow-up ( 1 9
months) for the fixed-effect model ( fig. 3 ). There were also
significant differences using the random-effect model for
general psychiatric symptoms (SMD = –1.45, 95% CI =
–2.87 to –0.03; Z = 2.00, p = 0.05). However, there were no
significant differences after 9 months between interven-
tion and control groups using the random-effect model
for depression (SMD = –1.48, 95% CI = –3.57 to 0.61; Z =
1.32, p = 0.19), anxiety (SMD = –1.53, 95% CI = –3.42 to
0.37; Z = 1.47, p = 0.14) or somatic symptoms (SMD =
–2.21, 95% CI = –5.49 to 1.07; Z = 1.32, p = 0.19).
Only 3 studies considered social adjustment or dis-
ability and the fixed-effect model showed modest, sig-
nificant improvements relative to controls in the short-
term (SMD = –0.65, 95% CI = –0.91 to –0.40; Z = 3.96,
p ! 0.001) and long-term outcomes (SMD = –0.69, 95%
CI = –0.96 to –0.43; Z = 3.60, p ! 0.001). The random-
effect model produced identical results.
Ten studies provided data for dropout from STPP
treatment versus control conditions. The rates of dropout
were significantly higher in the control groups (OR =
1.54, 95% CI = 1.06–2.25; Z = 2.25, p = 0.02), suggesting
STPP patients were 54% more likely to stay in treat-
ment.
H e t e r o g e n e i t y
Although the number of studies that reported any giv-
en outcome was small, we calculated formal tests of het-
erogeneity. These were significant in the majority of all
our meta-analyses. They were only nonsignificant for
medium-term outcomes and social adjustment. Similar-
ly, the I
2 statistic was consistently 1 50% for both short-
and long-term outcomes, although less so for medium-
term ones. The results of our meta-analyses should there-
fore be interpreted with caution.
Sensitivity Analyses
We conducted sensitivity analyses of the effect of only
including RCTs
[24–26, 29, 32–35, 37] . The fixed-effect
model results remained significant for all outcomes. Us-
ing the random-effect model, the medium-term out-
comes were unaltered, but the results were no longer sig-
nificant for any of the short- or long-term outcomes.
Restricting the analyses to studies with high CCDAN
scores, defined as a value greater than the midpoint of the
Short-Term Psychodynamic
Psychotherapy for Somatic Disorders
Psychother Psychosom 2009;78:265–274
269
scale ( 1 18), gave identical results to considering only
RCTs.
We also conducted a sensitivity analysis to examine
the effects of emotion-focused versus insight-based, or
interpersonally focused, approaches, by meta-analyzing
studies that emphasized emotional experiencing in their
technical description
[26, 28, 30, 36, 37] . The effects using
both the fixed- and random-effect models were signifi-
cant with medium to large ES on all measures in the
short-term (fixed-effect sizes = 0.60–1.10) and medium-
Study
or subcategory
n Mean STPP n Mean
control
SMD (fixed)
95% CI
Weight
%
SMD (fixed)
General psychiatric symptom measures: short-term
Svedlund et al. [24], 1983 50 7.96 (0.62) 50 11.34 (0.99) 1.99 –4.06 [–4.76, –3.37]
Hamilton et al. [25], 2000 37 0.67 (0.48) 31 0.67 (0.52) 4.23 0.00 [–0.48, 0.48]
Monsen and Monsen [26], 2000 20 0.40 (0.26) 20 0.66 (0.44) 2.35 –0.71 [–1.35, –0.06]
Junkert-Tress et al. [27], 2001 60 0.62 (0.52) 63 1.01 (0.59) 7.26 –0.70 [–1.06, –0.33]
Abbass [28], 2002 23 30.65 (24.40) 23 73.52 (36.68) 2.31 –1.35 [–2.00, –0.71]
Creed et al. [29], 2003 65 0.77 (0.48) 70 0.85 (0.50) 8.43 –0.16 [–0.50, 0.18]
Hinson et al. [30], 2006 9 30.60 (9.20) 9 37.70 (6.10) 1.01 –0.87 [–1.84, 0.11]
Tschuschke et al. [31], 2007 50 0.71 (0.44) 49 0.93 (0.44) 6.02 –0.50 [–0.90, –0.10]
Subtotal 314 315 33.60 –0.69 [–0.86, –0.52]
Test for heterogeneity: 2 = 112.75, d.f. = 7 (p = 0.00001), I2 = 93.8%
Test for overall effect: Z = 7.97 (p < 0.00001)
Depression: short-term
Svedlund et al. [24], 1983 50 2.40 (0.32) 50 3.55 (0.48) 3.11 –2.80 [–3.35, –2.24]
Guthrie et al. [32], 1993 50 8.18 (8.08) 47 13.60 (10.14) 5.82 –0.59 [–1.00, –0.18]
Jantschek et al. [33], 1998 52 7.80 (8.10) 27 7.80 (7.20) 4.46 0.00 [–0.46, 0.46]
Monsen and Monsen [26], 2000 20 0.49 (0.45) 20 0.83 (0.53) 2.36 –0.68 [–1.32, –0.04]
Abbass [28], 2002 28 7.11 (8.16) 29 17.45 (8.41) 2.97 –1.23 [–1.80, –0.66]
Hinson et al. [30], 2006 9 3.90 (2.10) 9 14.80 (7.10) 0.69 –1.98 [–3.16, –0.80]
Subtotal 209 182 19.41 –0.97 [–1.19, –0.74]
Test for heterogeneity: 2 = 65.90, d.f. = 5 (p = 0.00001), I2 = 92.4%
Test for overall effect: Z = 8.50 (p < 0.00001)
Anxiety: short-term
Svedlund et al. [24], 1983 50 4.03 (0.33) 50 5.54 (0.39) 1.94 –4.15 [–4.85, –3.44]
Bassett and Pilowsky [34], 1995 5 8.00 (2.70) 3 9.30 (0.60) 0.44 –0.51 [–1.98, 0.97]
Jantschek et al. [33], 1998 50 39.00 (11.81) 27 39.60 (10.70) 4.40 –0.05 [–0.52, 0.42]
Monsen and Monsen [26], 2000 20 0.31 (0.31) 20 0.60 (0.64) 2.40 –0.57 [–1.20, 0.07]
Linnet and Jemec [35], 2001 15 39.94 (8.29) 13 37.08 (9.10) 1.72 0.32 [–0.43, 1.07]
Abbass [28], 2002 25 7.56 (7.52) 26 20.35 (9.41) 2.47 –1.48 [–2.10, –0.85]
Hawkins [36], 2003 47 10.91 (3.89) 47 11.34 (4.77) 5.89 –0.10 [–0.50, 0.31]
Hinson et al. [30], 2006 9 4.90 (2.40) 9 19.70 (10.20) 0.72 –1.90 [–3.06, –0.74]
Subtotal 221 195 19.98 –0.74 [–0.96, –0.52]
Test for heterogeneity: 2 = 124.89, d.f. = 7 (p = 0.00001), I2 = 94.4%
Test for overall effect: Z = 6.64 (p < 0.00001)
Somatic symptoms: short-term
Svedlund et al. [24], 1983 50 9.72 (0.74) 50 12.68 (0.82) 2.21 –3.76 [–4.42, –3.10]
Bassett and Pilowsky [34], 1995 5 7.20 (1.30) 3 7.00 (1.00) 0.47 0.14 [–1.29, 1.58]
Hamilton et al. [25], 2000 37 10.90 (6.40) 31 12.40 (5.50) 4.20 –0.25 [–0.73, 0.23]
Monsen and Monsen [26], 2000 20 1.95 (1.50) 20 3.50 (2.19) 2.30 –0.81 [–1.46, –0.16]
Linnet and Jemec [35], 2001 14 28.59 (23.18) 13 21.44 (16.84) 1.66 0.34 [–0.42, 1.10]
Creed et al. [29], 2003 74 51.70 (28.38) 79 55.30 (27.38) 9.57 –0.13 [–0.45, 0.19]
Hawkins [36], 2003 47 35.98 (22.51) 47 48.60 (23.10) 5.68 –0.55 [–0.96, –0.14]
Hinson et al. [30], 2006 9 29.00 (20.60) 9 71.21 (42.50) 0.92 –1.20 [–2.23, –0.18]
Subtotal (95% CI) 256 252 27.00 –0.59 [–0.78, –0.40]
Test for heterogeneity: 2 = 107.17, d.f. = 7 (p = 0.00001), I2 = 93.5%
Test for overall effect: Z = 6.15 (p < 0.00001) –10 –5 0 5 10
Fig. 2. Meta-analysis of short-term out comes. Figures in parentheses are SD and values in square brackets rep-
resent 95% confidence limits.
Abbass /Kisely /Kroenke
Psychother Psychosom 2009;78:265–274
270
term results (fixed-effect sizes = 0.81–1.31). There were
insufficient studies to undertake meta-analyses of the
long-term outcomes.
Finally, we conducted sensitivity analyses of the effect
of only including studies with evaluation of therapy ad-
herence
[25–29, 33] . The fixed-effect model results re-
mained significant for all outcomes. Using the random-
effect model, the results were significant for general psy-
chiatric symptoms (SMD = –0.54, 95% CI = –0.96 to
–0.12; Z = 2.53, p = 0.01) and depression (SMD = –0.60,
95% CI = –1.09 to –0.11; Z = 2.42, p = 0.02) but not anxi-
ety or somatic symptoms in the short-term. There were
insufficient studies to undertake meta-analyses of the
medium- and long-term outcomes.
Publication Bias
The fail-safe N for short-term effectiveness ranged be-
tween 41 and 56, depending on the outcome, suggesting
that these findings were reasonably robust against publi-
cation bias. For medium-term outcomes, the fail-safe N
was between 16 and 19, indicating that these results were
more subject to publication bias. In the long-term follow-
up, our findings for depression and anxiety (fail-safe Ns
of 42 and 44, respectively) were more robust against pub-
lication bias than those for general psychiatric and so-
matic symptoms (fail-safe Ns of 14 and 12, respectively).
When we calculated the fail-safe N for our sensitivity
analyses, the numbers were reduced for the short-term
outcomes, but there was little effect on the medium- to
long-term outcomes where meta-analyses were possible.
For instance, the fail-safe N for short-term effectiveness
from RCTs ranged between 29 and 38.
Study
or subcategory
n Mean STPP n Mean control SMD (fixed)
95% CI
Weight
%
SMD (fixed)
General psychiatric symptom measures: long-term
Svedlund et al. [34], 1983 49 7.90 (0.73) 50 11.74 (0.93) 4.03 –4.55 [–5.31, –3.79]
Junkert-Tress et al. [27], 2001 46 0.64 (0.47) 63 1.01 (0.59) 15.13 –0.68 [–1.07, –0.29]
Creed et al. [29], 2003 68 0.78 (0.49) 71 0.72 (0.51) 20.88 0.12 [–0.21, 0.45]
Tschuschke et al. [31], 2007 35 0.56 (0.41) 49 0.93 (0.44) 11.24 –0.86 [–1.31, –0.40]
Subtotal 198 233 51.28 –0.70 [–0.91, –0.48]
Test for heterogeneity: 2 = 123.06, d.f. = 3 (p = 0.00001), I2 = 97.6%
Test for overall effect: Z = 6.43 (p < 0.00001)
Depression: long-term
Svedlund et al. [23], 1983 49 2.12 (0.35) 50 3.47 (0.39) 5.52 –3.61 [–4.26, –2.97]
Baldoni et al. [37], 1995 11 7.12 (4.12) 21 9.85 (6.01) 4.22 –0.49 [–1.23, 0.25]
Subtotal 60 71 9.74 –2.26 [–2.75, –1.77]
Test for heterogeneity: 2 = 38.78, d.f. = 1 (p = 0.00001), I2 = 97.4%
Test for overall effect: Z = 9.09 (p < 0.00001)
Anxiety: short-term
Svedlund et al. [24], 1983 49 4.11 (0.38) 50 5.53 (0.44) 5.90 –3.42 [–4.05, –2.80]
Baldoni et al. [37], 1995 11 6.62 (5.26) 21 10.09 (5.30) 4.13 –0.64 [–1.39, 0.11]
Subtotal 60 71 10.03 –2.28 [–2.76, –1.80]
Test for heterogeneity: 2 = 31.28, d.f. = 1 (p = 0.00001), I2 = 96.8%
Test for overall effect: Z = 9.30 (p < 0.00001)
Somatic symptoms: long-term
Svedlund et al. [24], 1983 49 8.05 (0.75) 50 13.57 (0.90) 2.23 –6.61 [–7.62, –5.59]
Baldoni et al. [37], 1995 11 9.75 (3.13) 21 10.57 (4.97) 4.33 –0.18 [–0.91, 0.55]
Creed et al. [29], 2003 72 52.80 (30.12) 77 51.10 (27.99) 22.39 0.06 [–0.26, 0.38]
Subtotal 132 148 28.95 –0.49 [–0.77, –0.21]
Test for heterogeneity: 2 = 150.43, d.f. = 2 (p = 0.00001), I2 = 98.7%
Test for overall effect: Z = 3.40 (p < 0.0007) –10 –5 0
Favors treatment Favors control
510
Fig. 3. Meta-analysis of long-te rm outcomes. Figures in parentheses are SD and values in square brackets rep-
resent 95% confidence limits.
Short-Term Psychodynamic
Psychotherapy for Somatic Disorders
Psychother Psychosom 2009;78:265–274
271
Other Studies and Findings
Nine studies did not meet criteria for inclusion in the
meta-analysis, yet provided preliminary evidence sup-
porting STPP for a range of conditions ( table 1 ).
Despite known bacteriological causes of ulcer, Sjodin
[23] found that STPP brought sustained gains compared
to medical treatment as usual in ulcer patients.
Bassler et al.
[45] studied a 12-week inpatient treat-
ment program for chronic ‘psychogenic’ pain that includ-
ed individual and group STPP. Sixty percent of the pa-
tients reported amelioration of pain symptoms. Those
who intellectualized and rationalized more had less re-
sponse to treatment, highlighting the purported role of
emotional experiencing in bringing symptom ameliora-
tion.
Case series for physical symptoms yielded improve-
ment rates of 76–9 0%
[42–44] . Ventegodt et al. [47] , using
a combination of STPP and ‘body work’, found signifi-
cant symptom improvements in a mixed group of physi-
cally ill patients, although a large portion of the sample
was lost to follow-up.
Two studies [38, 41] examined the impact of STPP on
alexithymia. Beresnevaite
[41] found that reductions in
alexithymia were associated with fewer cardiac events in
2-y ear follow-up of pat ients wit h coronary arte ry dis eas e.
While the treatment and control groups did not signifi-
cantly differ in degree of alexithymia at posttreatment,
the STPP group had no hospitalizations plus reduced re-
ports of angina whereas the controls had 4 hospitaliza-
tions for angina. Poulsen
[38] found that rheumatoid ar-
thritis and Sjögrens’ syndrome patients treated with
group STPP had lower alexithymia ratings compared to
controls at posttreatment, but they did not have pretreat-
ment measurements.
Reduced hospitalization rates compared to controls
were reported in studies of STPP for coronary artery dis-
ease
[41] ( 2 , p ! 0.01), Crohn’s disease [39] (p = 0.03) and
chronic respiratory disease
[22] ( 2 , p ! 0.001). Two trials
reported trends toward reduced surgical procedures in
ulcer disease
[23] (p = 0.07) and in Crohn’s disease [33] ,
where 15% of the STPP patients vs. 26% of the controls
required surgery (p = 0.27).
Discussion
Within the limitations of study quality and the effects
of heterogeneity on statistical interpretability, the evi-
de nce fro m this r evie w sug gest s that STPP me thods show
promise as adjunctive or solo treatments for a range of
somatic problems
[48] . In addition to reducing physical
and psychological symptoms, these brief treatments ap-
pe ar ed to i mpr ov e t re at me nt c omp li ance a s wel l a s s oc ia l-
occupational function and reduce healthcare utilization.
These improvements were noted in the majority of stud-
ies as measured by blinded clinicians, unblinded clini-
cians, and patient self-ratings.
These results compare favorably to a similarly con-
ducted review of CBT for somatic disorders
[10] . This tri-
al included 29 RCT and 2 non-RCT studies of diverse
conditions with a variety of CBT methods and group and
individual form ats. In t his revie w, only 9 stated t hey used
manuals and 7 had adherence ratings. They found defi-
n ite or po ssib le sy mp to m b en ef it s i n 82 %, f un ct io na l b en -
efits in 73% and psychological benefits in 46% of the in-
cluded studies.
Likewise, 91.3% of the STPP studies showed at least
some benefit (on 6 1 parameter) for these patient popula-
tions, compared to 69% of the antidepressant research in
a s ys te ma ti c r ev ie w o f 9 4 r a nd om iz ed t ri a ls
[9] . Moreover,
the antidepressant studies were only short-term, with a
median duration of 9 weeks, compared to the long-term
follow-up in the majority of the STPP studies. Recent lit-
erature syntheses confirm that CBT and antidepressants
are among the most evidence-based treatments for so-
matic symptom disorders
[13 , 14] . Our findings suggest
STPP may be another valuable therapeutic option.
Emotional factors, including reduced alexithymia,
building awareness of unconscious processes and emo-
tional experiencing are possible or probable treatment
factors rendering these therapies effective. This notion is
bolstered by our subanalysis showing strong effects when
studying the more emotion-focused STPP models. This
finding concurs with a recent meta-analysis of 10 STPP
studies of diverse conditions according to which outcome
correlated with emotional focus
[49] . Comparative evalu-
ations of the more emotion-focused versus insight-based
models are warranted to test the hypothesis that emo-
tional experiencing has a central healing effect in these
somatic disorders, as this research suggests.
The somewhat positive results of this review should be
interpreted within the following limitations. First, the in-
cluded studies were of variable methodological quality,
conducted with a broad range of scientific rigor. Second,
there is a high probability of selection bias in some of the
studies, although in the 13 RCTs the use of randomiza-
tion should have mitigated between-group differences as
a confounder. Third, there is possible reporting bias,
where striking positive (stopping smoking) or negative
events (vagotomy surgery) would be more likely reported
Abbass /Kisely /Kroenke
Psychother Psychosom 2009;78:265–274
272
Tab le 1. STPP study designs and outcomes
Authors Subjects at
baseline
Age
years
Female
%
STPP
model
Number
of
sessions
Follow-up
months
Control Outcomes (somatic/
S-O function/psychologi-
cal/healthcare utilization)
total STPP
Randomized trials
Rosser et al. [22], 1983
Chronic bronchitis and emphysema
33 16 66 38 Malan 8 6 medical treatment +/0/– +/0 0/– +
Svedlund et al. [24], 1983
Irritable bowel syndrome
101 50 24 70 Malan ≤10 15 medication,
anxiolytics
++++
Sjodin [23], 1983
Ulcer disease
103 50 45.5 39 Malan ≤10 3, 12 medication + + + 0
Basssett and Pilowsky [34], 1985
Chronic pain
22 14 40.8 17 not
defined
12 6, 12 supportive
cognitive therapy
0 +/0 0
Poulsen [36], 1991
Rheumatoid arthritis, Sjögren’s
syndrome
46 23 51.9 90 group-
analytic
12 9 no treatment + +/0
Guthrie et al. [32], 1993
Refractory irritable bowel syndrome
102 53 47 86 Hobson 7 3 supportive
therapy
++++
Baldoni et al. [37], 1995
Urethral syndrome/pelvic pain
36 13 00 Davanloo,
Malan
14–16 6, 48 treatment as usual + + +
Jantschek et al. [33], 1998
Keller et al. [39], 2004
Deter et al. [40], 2007
Crohn’s disease
108 71 16–55 NA Luborsky,
relaxation
10+ 6, 12, 24 medical treatment 0 0 0 +
Hamilton et al. [25], 2000
Chronic functional dyspepsia
77 37 40 59.5 Hobson 8 12 supportive therapy + +/0 0
Monsen and Monsen [26], 2000
Chronic pain
40 20 45.5 35 affect-
focused
33 12 treatment as usual + + +
Beresnevaite [41], 2000
Coronary heart disease
40 20 51.8 5 group 16 6, 12, 24 education group +/0 + +/0
Linnet and Jemec [35], 2001
Atopic dermatitis
32 16 28.3 23 Malan 15.5 12 medical
treatment
00
Creed et al. [29], 2003
Irritable bowel syndrome
257 85 18–65 NA Hobson 8 12 treatment as usual +/0 + +/0 +
Pre-post studies
Barnat [42], 1981
Refractory headache
79 36 75.9 individual
STPP
5++
Sifneos [43], 1973
Physical symptoms
14 Sifneos +
Nielsen et al. [44], 1988
Physical symptoms
10 33.4 70 Sifneos,
Malan
22 24 + +
Bassler et al. [45], 1994
Chronic pain
50 36.2 68 individual,
group
12 1.5 + +
Junkert-Tress et al. [27], 2001
Somatoform mixed
87 36 55 Strupp,
Binder
60 + +
Abbass [28], 2002, [46], 2003
Somatoform mixed
33 40.6 63.6 Davanloo 18.6 12, 36 + + + +
Hawkins [36], 2003
Chronic pain
47 46 64 Davanloo,
group
8+0
Hinson et al. [30], 2006
Movement disorders
10 31 77.7 Davanloo 12 + + +
Tschuschke et al. [31], 2007
Somatoform disorders
50 42.5 62 STPP Group 20 6, 12 + +
Ventegodt et al. [47], 2007
Chronic pain
31 STPP, body
work
20 12 + + +
S-O = Social-occupationa l. +, O and – denote STPP superior to, equal to or inferior to control or pretreatment primary measures. Blank spaces de-
note no data provided or not applicable.
Short-Term Psychodynamic
Psychotherapy for Somatic Disorders
Psychother Psychosom 2009;78:265–274
273
in o nly s ome studie s. Fo ur th, mo st o f the treat ment s we re
neither manualized nor adherence rated to ensure treat-
ment standardization. Fifth, only 4 and 5 of the studies
in the meta-analysis had medium- and long-term follow-
ups, respectively. Finally, the heterogeneity in most meta-
analyses, the loss of sig nif icance in some cases using ran-
dom-effect modeling and the inclusion of only 14 studies
suggest that the meta-analysis results need to be inter-
preted with caution.
This heterogeneity may have arisen from both clinica l
or methodological diversity, or both, among the trials
[47] . In this study clinical variation could be explained by
the diversity of the psychotherapeutic interventions that
were included in the review (e.g. group versus individual
therapy formats), as well as of subjects in terms of diag-
noses (e.g. Crohn’s disease versus movement disorders)
and socio-demographics (e.g. age and gender). Method-
ological diversity could be explained by differences be-
tween the studies in terms of design (e.g. randomized ver-
sus nonrandomized designs) or in the way the outcomes
were defined and measured. We attempted to minimize
heterogeneity in several ways. Firstly, we did not report
combined ES for short-, medium- and long-term out-
comes but reported the results for depression, anxiety
and somatic symptoms separately. Secondly, we under-
took sensitivity analyses restricted to randomized con-
trolled trials, higher-quality studies and adherence-rated
therapies. Thirdly, we also reported random-effect meta-
analyses, which incorporate heterogeneity in their calcu-
lation.
One strength of the reviewed research is the diversity
of study centers and the inclusion of both RCTs as well as
case series and naturalistic studies. The latter studies of-
fer evidence that some patients with this range of condi-
tions can benefit in real-world settings with improve-
ments in psychological functioning, physical symptoms
and healthcare utilization. The finding that a broad range
of conditions may benefit from this treatment suggests
STPP may provide a general health benefit.
Greater retention rates with STPP and reduced health-
care utilization are important findings. Conditions such
as movement disorders, chronic pain and headache are
often treated with medications and physical procedures
as first-line agents. Given the availability of brief psycho-
therapeutic interventions, STPP therapy might be one
option clinicians could consider before embarking on
more invasive or long-term alternatives.
Within the limitations of methodological and other
problems within this group of studies, STPP may provide
benefits across a range of physical and somatic symptom
disorders. Future research should include more rigorous
methods and study specific conditions while using treat-
ment manuals with adherence ratings. Combinations of
RCT and naturalistic studies measuring healthcare utili-
zation and mortality rates are also warranted. STPP can
be considered as a solo treatment for some somatic condi-
tions and an adjunct for other physical conditions that
may improve treatment retention and outcome.
A c k n o w l e g m e n t s
This research was supported by the Dalhousie University
Department of Psychiatry, Capital Health and the Nova Scotia
Department of Health.
References
1 Kroenke K: Patients presenting with somatic
complaints: epidemiology, psychiatric co-
morbidity and management. Int J Methods
Psychiatr Res 2003;
12: 34–43.
2 Aaron LA, Buchwald D: A review of the evi-
dence for overlap among unexplained clini-
cal conditions. Ann Intern Med 2001;
134:
868–881.
3 Henningsen P, Zimmermann T, Sattel H:
Medically unexplained physical symptoms,
anxiety, and depression: a meta-analy tic re-
view. Psychosom Med 2003;
65: 528–533.
4 Kroenke K, Rosmalen JG: Symptoms, syn-
dromes, and the value of psychiatric diag-
nostics in patients who have functional so-
matic disorders. Med Clin North Am 2006;
90: 603–626.
5 Kroenke K: The interface between physical
and psychological symptoms. Prim Care
Companion J Clin Psychiatr y 2003;
5(suppl
7):11–18.
6 Sha MC, Callahan CM, Counsell SR, West-
moreland GR, Stump TE, Kroenke K: Physi-
cal symptoms as a predictor of health care
use and mortality among older adults. Am J
Med 2005;
118: 301–306.
7 Katon W, Lin EH, Kroenke K: The associa-
tion of depression and anxiety with medical
symptom burden in patients with chronic
medical illness. Gen Hosp Psychiatry 2007;
29: 147–155.
8 Kroenke K, Sharpe M, Sykes R: Revising the
classification of somatoform disorders: key
questions and preliminar y recommenda-
tions. Psychosomatics 2007;
48: 277–285.
9 O’Malley P G, Jackson JL , Santoro J, Tomkins
G, Balden E, Kroenke K: Antidepressant
therapy for unexplained symptoms and
symptom sy ndromes. J Fam Pract 1999;
48:
980–990.
Abbass /Kisely /Kroenke
Psychother Psychosom 2009;78:265–274
274
10 Kroenke K , Swindle R: C ognitive-beh avioral
therapy for somatization and sy mptom syn-
dromes: a critical review of controlled clini-
cal trials. Psychother Psychosom 2000;
69:
205–215.
11 Allen LA, Escobar JI, Lehrer PM, Gara MA,
Woolfolk RL: Psychosocial treatments for
multiple unexplained physical symptoms: a
review of t he literature. Psychosom Med
2002;
64: 939–950.
12 Raine R, Haines A, Sensky T, Hutchings A,
L ar k in K, Bl a ck N: Sy st e ma ti c r e vi e w o f m en -
tal health interventions for patients with
common somatic symptoms: can research
evidence f rom secondary care be extrapolat-
ed to primary care? BMJ 2002;
325: 1082.
13 Jackson JL , O’Malley PG, Kroenke K: Anti-
depressants and cognitive-behavioral thera-
py for symptom syndromes. CNS Spectr
2006;
11: 212–222.
14 Kroenke K: Efficac y of treatment for so-
matoform dis orders: a review of ra ndomized
controlled trials. Psychosom Med 2007;
69:
881–888.
15 Woolfolk R , Allen L, Tiu J: New d irections in
the treatment of somatization. Psychiatr
Clin North Am 2007;
30: 621–644.
16 Alexander F (ed): Psychosomatic Medicine.
Its Principles and Applications. New York,
Norton & Co, 1950.
17 Davanloo H (ed): Short-Term Dynamic Psy-
chotherapy. New York, Aronson, 1980.
18 Leichsenring F, Rabung S, Leibing E: The ef-
ficac y of short-term psychod ynamic psycho -
therapy in specific psychiatric disorders: a
meta-ana lysis. Arc h Gen Psychiatr y 2004;
61:
1208–1216.
19 Abbass AA, Hancock JT, Henderson J,
Kisely S: Short-term psychodynamic psy-
chotherapies for common mental disor-
ders. Cochrane Database Syst Rev 2006;
4:
CD004687.
20 Cohen J: Statis tical power an alysis for the be-
havioral sciences, ed 2. Hillsdale, Erlbaum
Associates, 1988.
21 Abramson JH: WINPEPI (PEPI-for-Win-
dows): computer programs for epidemiolo-
gists. Epidemiol Perspect Innov 2004;
1: 6.
22 Rosser R, Denford J, Heslop A, Kinston W,
Macklin D, Minty K, Moynihan C, Muir B,
Rein L, Guz A: Breathlessness and psychiat-
ric morbidity in chronic bronchitis and em-
physema: a study of psychot herapeutic man-
agement. Psychol Med 1983;
13: 93–110.
23 Sjodin I: Psychotherapy in peptic ulcer dis-
ease. Acta Psychiatr Scand 1983;
307: 1–90.
2 4 Sv edl und J, Sj odi n I, Ott oso n JO, Dot eva ll G :
Controlled study of psychotherapy in irrita-
ble bowel syndrome. Lancet 1983;
10: 589–
592.
25 Hamilton J, Guthrie E, Creed F, Thompson
D, Tomenson B, Bennett R, Moriart y K, Ste-
phens W, Liston R: A random ized controlled
trial of psychotherapy in patients with
chronic functional dyspepsia. Gastroenter-
ology 2000;
119: 661–669.
26 Monsen K, Monsen JT: Chronic pain and
psychodynamic body therapy: a controlled
outcome study. Psychotherapy 2000;
37: 257–
269.
27 Junkert-Tress B, Schnierda U, Hartkamp N,
Schmitz N, Tress W: Effects of short-term
dynamic psychotherapy for neurotic, so-
matoform, and personality disorders: a pro-
spective 1-year follow-up study. Psychother
Res 2001;
11: 187–20 0.
28 Abbass A: Intensive short-term dynamic
psychothe rapy in a private psyc hiatric off ice:
clinical and cost effectiveness. Am J Psycho-
ther 2002;
56: 225–232.
29 Creed F, Fernandes L, Guthrie E, Palmer S,
Ratcliffe J, Read N, R igby C, Thompson D,
Tomenson B: The cost effectiveness of psy-
chotherapy and paroxetine for severe irrita-
ble bowel synd rome. Gastroenterolog y 2003;
124: 303–317.
30 Hinson VK, Weinstein S, Bernard B, Leur-
gans SE, Goetz CG: Single-blind clinical tri-
al of psychot herapy for treatment of psycho-
genic movement disorders. Parkinsonism
Relat Disord 2006;
12: 177–180.
31 Tschuschke V, Weber R, Horn E, Kiencke
P, Tress W: Ambulante psychodynamische
Kurzgruppenpsychotherapie bei Patienten
mit somatoformen Störungen. Z Psychiatr
Psycho Psychot her 2007;
55: 87–95.
32 Gut hrie E , Cre ed F, Daw son D, Tomens on B:
A randomized controlled trial of psycho-
therapy in patients with refractory irritable
bowel syndrome. Br J Psychiatry 1993;
163:
315–321.
33 Jantschek G, Zeitz M, Pritsch M, Wirschin
M, Klor HU, St udt HH: Effect of psyc hother-
apy on the course of Crohn’s disease. Scand J
Gastroenterol 1998;
33: 1289–1296.
34 Bassett DL, Pilowsky I: A study of brief psy-
chotherapy for chronic pain. J Psychosom
Res 1985;
29: 259–264.
35 Linnet J, Jemec GBE: Anxiety level and se-
verity of sk in condition predicts outcome of
psychotherapy in atopic dermatitis patients.
Int J Dermatol 2001;
40: 632–636.
36 Hawkins J: The Role of Emotional Repres-
sion in Chron ic Back Pain. A St udy of Chron-
ic Back Pain Patients Undergoing Group
Psychodynamic Psychotherapy as Treat-
ment for Their Pain (published PhD disser-
tation). New York, New York University,
2003.
37 Baldoni F, Baldaro B, Trombini G: Psycho-
therapeutic perspectives in urethral syn-
drome. Stress Med 1995;
11: 79–84.
38 Poulsen A: Psychodynamic, time-limited
group therapy in rheumatic disease – a con-
trolled study with special reference to alexi-
thymia. Psychother Psychosom 1991;
56: 12–
23.
39 Keller W, Pritsch M, von Wietersheim J, Pe-
ter S, Osborn W, Balck F, Dilg R, Schmelz-
Schumacher E, Doppl W, Jantschek G, Deter
HC: The effects of psychotherapy and relax-
ation on the ps ychosocial a nd somatic course
of Crohn’s disease: main resu lts of the Ger-
man Prosp ective Multicente r Psychotherapy
Treatment Study on Crohn’s disease. J Psy-
chosom Res 20 04;
56: 687–696.
40 Deter HC, Keller W, von Wietersheim J,
Jantschek G , Duchmann R , Zeitz M, Germ an
Study Group on Psychosocial Intervention
in Crohn’s Disease: Psychological treatment
may reduce the need for healthcare in pa-
tients with Crohn’s disease. Inf lamm Bowel
Dis 2007;
13: 745 –7 52 .
41 Beresnevaite M: Exploring the benef its of
group psychotherapy in reducing alexi-
thymia in coronar y heart disease patients: a
preliminary study. Psychother Psychosom
2000;
69: 117–122 .
42 Barnat MR: Short-term psychotherapy and
refractory headache. Headache 1981;
21: 257–
260.
43 Sifneos PE: The preva lence of ‘alexithymic’
characteristics in psychosomatic patients.
Psychother Psychosom 1973;
22: 255–262.
44 Nielsen G, Barth K, Haver B, Havik O: Brief
dynam ic psychotherapy for pat ients present-
ing physica l symptoms. Psychother Psycho-
som 1988;
50: 35–41.
45 Bassler M, Krauthauser H, Hoffmann SO:
Inpatient psychotherapy with chronic psy-
chogenic pain patients. Psychother Psycho-
som Med Psychol 1994;
44: 299–307.
46 Abbass A: The cost-effectiveness of short-
term dynamic psychotherapy. J Pharmaco-
economics Outcomes Res 2003;
3: 535–539.
47 Ventegodt S, Thegler S, Andreasen T, Struve
F, Enevoldsen L, Bassaine L, Torp M, Mer-
rick J: Clinical holistic medicine (mindful,
short-term psychodynamic psychotherapy
complemented with bodywork) in the treat-
ment of experienced physical illness and
chronic pain. Sci World J 2007;
7: 310 –316.
48 Higgins JPT, Green S (ed): Cochrane Hand-
book for Systematic Reviews of Interven-
tions Version 5.0.0. Coc hrane Colla boration,
2008. w ww.cochrane-handbook.org (updat-
ed February 2008).
49 Diener MJ, Hilsenroth MJ, Weinberger J:
Therapist affect focus and patient outcomes
in psychody namic psychotherapy: a meta-
analysis. Am J Psychiatry 2007;
164: 936–
941.
