Content uploaded by Paul Crits-Christoph
Author content
All content in this area was uploaded by Paul Crits-Christoph
Content may be subject to copyright.
Am JPsychiatry 149:2, February 1992 151
THE AMERICAN JOURNAL OF PSYCHIATRY
Special Articles
The Efficacy of Brief Dynamic Psychotherapy:
A Meta-Analysis
Paul Crits-Christoph, Ph.D.
Objective: Insurance companies, legislators, and funding agencies have become increasingly
concerned with efficacy and accountability in regard to psychotherapy, andpsychodynamic ther-
apy is a primary target ofconcern because it is widely practiced in outpatient settings. This paper
is a meta-analytic review ofrecent well-controlled studies ofthe efficacy ofbriefdynamic therapy.
Method: The meta-analysis included both published studies, located through an extensive corn-
puterized search of psychiatry and psychology journals, and studies reported at conferences.
Eleven studies met the inclusion criteria: use of a specific form of short-term dynamic psycho-
therapy as represented in a treatment manual or manual-like guide; comparison of brief dynamic
therapy and a waiting list control condition, nonpsychiatric treatment, alternative psychother-
apy, pharmacotherapy, or other form of dynamic therapy; provision of the information neces-
sary for calculation ofeffect sizes; at least 1 2 therapy sessions; and therapists who were trained
and experienced in briefdynamic therapy. The outcome measures compared were target syrnp-
toms, psychiatric symptoms generally, and social functioning. Results: Brief dynamic therapy
demonstrated large effects relative to waiting list conditions but only slight superiority to non-
psychiatric treatments. Its effects were about equal to those ofother psychotherapies and medi-
cation. Conclusions: These data confirm previous indications that various psychotherapies do
not differ in effectiveness, although this finding should not be generalized to all patient popula-
tions, outcome measures, and treatment types. Also, the highly controlled conditions of these
studies limit conclusions about actual practice. Future studies should address various treatment
lengths, follow-up assessments, and specific treatments, patient groups, and outcome measures.
(Am JPsychiatry 1992; 149:151-158)
Over the past decade, insurance companies, legisla-
tors, and funding agencies have become increas-
ingly concerned with efficacy and accountability in re-
Received Feb. 8, 1991; revision received June 17, 1991; accepted
July 16, 1991. From the Department of Psychiatry, School of Medi-
cine, University of Pennsylvania. Address reprint requests to Dr. Crits-
Christoph, 301 Piersol Bldg., Hospital of the University of Pennsyl-
vania, 3400 Spruce St., Philadelphia, PA 19104.
Supported in part by NIMH Clinical Research Center grant MH-
45178, grant MH-40472, and Career Development Award MH-
00756 to Dr. Crits-Christoph.
Copyright © 1992 American Psychiatric Association.
gard to psychotherapy (1 ). Psychodynamic therapy, be-
cause it is widely practiced in outpatient settings, is a
natural target for such concerns about efficacy. Although
there have been a number of reviews of the literature on
the efficacy of psychotherapy in general (2, 3) and the
efficacy of specific treatments (4, 5), few reviews have
been devoted to the efficacy of psychodynamic therapy in
particular. An obvious reason for this is the relative
dearth until recent years of well-controlled studies of the
outcome of dynamic psychotherapy. In addition, a van-
ety of methodological problems hindered the interpreta-
tion of results from earlier studies.
BRIEF DYNAMIC THERAPY
152 Am/Psychiatry 149:2, February 1992
Perhaps the greatest problem with many efficacy
studies of dynamic therapy has been the lack of control
and specification of the treatment variable. The im-
provements in outcome research due to the use of treat-
ment manuals and careful selection, supervision, and
monitoring of therapists are relatively recent (6). Many
otherwise well-controlled studies of the efficacy of dy-
namic therapy (e.g., 7-10) left unanswered questions
about the meaning of the findings because the nature of
the treatment provided by dynamic therapists was not
cleanly specified and controlled. These studies in gen-
eral allowed dynamic therapists to conduct treatment
as they usually practiced, on the assumption that differ-
ing orientations and styles among dynamic therapists
are of little concern. Although such naturalistic treat-
ment studies perhaps generalize better to the eclectic
practice of dynamic therapy as it actually occurs, nec-
ommendations about specific treatment approaches
cannot be based on such studies. In addition, recent evi-
dence (1 1 )indicates that studies which do not use treat-
ment manuals are much more prone to therapist effects
(i.e., significant differences between therapists in out-
comes) than are studies which do use treatment manu-
als. Such therapist effects create a statistical problem of
nonindependence, which can lead to highly distorted
conclusions about the efficacy of treatments when ig-
nored in the analysis of comparative outcome data (as
is typically the case) (12). Thus, many of the findings
from studies of dynamic therapy that did not use manu-
als are questionable on statistical grounds and do not
provide clear clinical recommendations because of the
eclectic treatment. Moreover, although the use of a dy-
namic therapy treatment manual does not necessarily
lead to greaten treatment benefits, the more careful se-
lection, training, and monitoring of therapists in the re-
cent comparative studies using manuals are likely to
provide the highest quality of treatment possible.
In the area of brief dynamic therapy, a number of
specific dynamic approaches have now been articulated
in the form of guides or manuals to be used in the train-
ing of therapists: Luborsky (13), Strupp and Binder (14),
Sifneos (15), Davanloo (16), Mann (17), Horowitz
(18), and others. Although not all of these books have
been labeled “manuals,” what they have in common are
a clearly laid out theory, detailed case examples, and a
set of technical recommendations for the practice of a
particular dynamic therapy (1 9). In addition, these guides
lend themselves well to use in a research context, and
studies implementing these approaches are beginning to
accumulate.
Because of these improvements in outcome research,
I was able to conduct a meta-analytic review (20) of the
evidence for the efficacy of short-term dynamic ther-
apy, attending only to studies that controlled the treat-
ment variable through the use of a manual or manual-
like guide for therapists. Three main questions were
addressed: 1) To what extent is there evidence that brief
dynamic therapy is more effective than no treatment?
2) How does brief dynamic therapy compare to regular
contact with patients that is not standard psychiatric
treatment? and 3) How does brief dynamic therapy
compare to other forms of psychiatric treatment, in-
cluding medication and nondynamic psychotherapies?
In addition to summarizing the literature on these ques-
tions, I will make recommendations about future re-
search that would advance this area of study.
METHOD
Selection of Studies
To qualify for this review, a study had to test the ef-
ficacy of a specific form of short-term dynamic therapy
as represented in a treatment manual or manual-like
guide. Studies comparing brief dynamic therapy to a
waiting list control condition, a nonpsychiatnic treat-
ment, an alternative psychotherapy, pharmacotherapy,
or another form of dynamic therapy (or some combina-
tion of these comparison conditions) were included. In
addition, only reports that included the information
necessary for the calculation of effect sizes (discussed
later) were included. The review was not restricted to
published studies but included studies that had been
presented at conferences. The published studies were
found through an extensive computerized search of
psychiatry and psychology journals.
Any study that did not involve a patient group (e.g.,
analogue study) was excluded from the review. The
shortest study accepted for the review was the 12-ses-
sion treatment described by Mann (17). Also excluded
from the review were studies that used therapists who
were not specifically trained and experienced in brief
dynamic therapy techniques. This excluded studies
which either used trainees or used experienced thera-
pists who had not received training in short-term dy-
namic therapy. This criterion was chosen because most
therapists with a psychodynamic orientation tend to
adhere to long-term treatment and are slow to adapt to
the philosophy of short-term therapies (21).
The 1 1 studies that met the inclusion criteria are
listed in table 1 with the type of patients, treatment con-
ditions, number of patients in each condition, and
length of therapy for each study. These studies included
tests of the efficacy of the brief dynamic treatments de-
scribed by Mann (17), Davanloo (16), Luborsky (13),
Horowitz (18), Pollack and Homer (32), Strupp and
Binder (14), Malan (33), and Klerman, Weissman,
Rounsaville, and Chevron (34). The interpersonal ap-
proach of Klerman et al., although less like other dy-
namic therapies in that the relationship with the thera-
pist is not generally a focus of intervention, was
included because of its historical roots in the work of
Sullivan. In addition, the NIMH Treatment of Depres-
sion Collaborative Research Program (35) identified
the interpersonal approach as generally within the psy-
chodynamic domain, and psychodynamically oriented
therapists were selected to perform the treatment in that
study.
The 11 studies typically had a training period for
PAUL CRITS-CHRISTOPH
Am JPsychiatry I 49:2, February 1992 153
TABLE 1. Studies of the Efficacy of Brief Dynamic Psychotherapy
Authors Disorder Treatment Conditions N Treatment Length
Piper et al. (22) Mixed Dynamic
Waiting list 48
57 Mean=18.6 sessions
S months
Shefler and Dasberg Mixed Dynamic 16 Mean=12 sessions
(unpublished, 1989) Waiting list 17 3 months
Elkin et al. (23) Depression Interpersonal
Cognitive
Imipramine
Placebo and clinical management
47
37
37
35
Mean=16.2 sessionsa
Mean=16.2 sessionsa
Mean=16.2 sessionsa
Mean=16.2 sessionsa
Rounsaville et al. (24) Opiate addiction Interpersonal
Low-contact 22
28 Mean=13.2
1 session/month for 6 months1’
Winston et al. (25) Personality
disorders Dynamic-Davanloo
Dynamic-Pollack and Homer
Waiting list
IS
17
I 7
40 sessions
40 sessions
S months
Thompson et al. (26) Depression Dynamic
Cognitive
Behavioral
30
31
30
16-20 sessions
16-20 sessions
16-20 sessions
Brom et al. (27) Posttraumatic
stress disorder Dynamic
Desensitization
Hypnotherapy
Waiting list
29
3 1
29
23
Mean=1 8.8 sessions
Mean=1 S sessions
Mean=14.4 sessions
4 months
DiMascio et al. (28) Depression Interpersonal
Amitriptyline
Low-contact
17
20
21
Mean=16 sessions
Mean=16 sessions
1 session/month
Carroll et al. (29) Cocaine abuse Interpersonal
Relapse prevention 21
2 1 12 sessions
I 2 sessions
Woody et al. (30) Opiate addiction Dynamic
Cognitive
Drug counseling
32
39
35
Mean=12 sessions
Mean=9.S sessions
Mean=17 sessions
Marmar Ct al. (31) Pathological grief Dynamic
Self-help 31
30 Mean=12 sessions
Mean=12 sessions
aMean number of sessions for all patients who completed the study; data were not broken down by treatment group.
bMean number of sessions not reported, but data on dropouts suggest low-contact patients attended a mean of approximately four sessions.
therapists before the actual outcome study, but, as in-
dicated in the selection criteria, they did not use inexpe-
rienced therapists. On the contrary, most studies used
highly experienced therapists. In four studies the aver-
age length of postgraduate clinical experience was more
than 9 years. Four other studies listed only the mini-
mum number of postgraduate years required for thera-
pists who participated in the study (typically 2 years).
One study used therapists who were not only experi-
enced clinicians but also had supervised training in the
particular form of brief dynamic therapy for 2 years
before the start of the study. Another study employed
therapists who had on the average 5.5 years of experi-
ence in working with the types of patients studied. One
study did not specify the therapist experience level other
than to say “an experienced psychiatrist.”
Outcome Measures
Rather than asking the broad question, Does brief dy-
namic therapy work? I asked three more specific ques-
tions: I ) Does short-term dynamic therapy affect the
specific target symptoms or problem areas for which
patients are seeking treatment? 2) Does short-term dy-
namic therapy change the general level of psychiatric
symptoms, and 3) Does short-term dynamic therapy
improve social functioning?
The goal of this meta-analytic review was, as much as
possible, to assess the effects of brief dynamic therapy
on specific outcome measures. Since different outcome
measures may vary in their reliability and sensitivity to
change (36), it has been recommended (37) that meta-
analyses not combine information on different outcome
measures. Although the practice of combining informa-
tion on heterogeneous outcome measures was common
in early meta-analyses (e.g., 2), more recent meta-ana-
lytic reviews (e.g., 5) have begun to focus on specific
measures.
In selecting measures that could provide answers to
my three questions, I had to chose measures that were
available in most studies. To assess specific target
symptoms or problem areas for which patients seek
treatment, I examined two forms of outcome measures.
For studies of heterogeneous patient groups, the target
complaints method (38) was selected. This measure re-
quires the patient to rate the severity of the three main
problems for which he or she is seeking treatment. For
studies that investigated dynamic therapy for a specific
patient group (e.g., cocaine abusers), the main problem
area was already defined by the nature of patient selec-
tion. In these studies, a measure of the severity of this
particular problem area or disorder (e.g., the Beck De-
pression Inventory for depressed patients) was selected
to represent change in target symptoms.
To answer the question about the effect of brief dy-
namic therapy on general levels of psychiatric symptoms,
the SCL-90 (39) was chosen. The Social Adjustment Scale
(40) was selected to assess social functioning. In each of
BRIEF DYNAMIC ThERAPY
1S4 Am JPsychiatry I 49:2, February 1992
TABLE 2. Effect Sizes for Comparisons of Brief Dynamic Psychotherapy and Other Treatment Conditions
Condition Compared With
Brief Dynamic Psychotherapy Number of
Comparisons
Average Effect Size (d)a
Target
Symptoms General Psychiatric
Symptoms Social
Adjustment
Waiting list control
Alternative nonpsychiatric treatments
Alternative psychiatric treatments
Other dynamic therapy
S
5
9
I
1.10
0.32
-0.05
-0.25
0.82
0.20
-0.01
0.05
0.81
0.09
-0.05
0.04
aCalculated according to the method of Cohen (41).
the few cases where these measures were not discussed in
the published version of the study, the original investiga-
tors were contacted and asked to provide data on the
measure if it was used in the study. In two studies the
Social Adjustment Scale was not used, but ratings on al-
ternative measures of social functioning were available,
and change on these other instruments was evaluated for
this review. If no measure of the construct was available,
no effect size was calculated.
In general, only assessments of outcome done at or
near the termination of treatment were used for this re-
view. Although it might be of interest to examine the
effects of dynamic therapy some time after termination,
since delayed effects might be expected, not enough re-
ports of follow-up data are available yet.
Effect Size Calculations
The primary defining characteristics of a meta-analy-
sis are the calculation of effect sizes and the integration
of these effect sizes across studies. For this study the
calculation of effect sizes was guided by Cohen (41) and
used his d statistic. Effect size was calculated as the
posttreatment mean of the dynamic treatment group
minus the posttreatment mean of the comparison
group, divided by the standard deviation of the corn-
parison group. In some cases, only the adjusted post-
treatment means and standard deviations (adjusted for
initial levels) were available, and these were used in-
stead of the raw means. A pooled standard deviation
(across treatment groups) was used when the standard
deviation for the comparison group was not available.
Effect size calculated this way can be readily inter-
preted as the distance, in standard deviation units, that
the average brief dynamic therapy patient was from the
average comparison group patient. An effect size of
zero indicates that the dynamic therapy and compari-
son groups were equal in outcome. All calculations
were made so that a positive effect size indicates supe-
riority of brief dynamic therapy and a negative effect
size indicates superiority of the comparison condition.
Thus, an effect size of 0.50 would indicate that the im-
provement of the average dynamic therapy patient was
one-half of the standard deviation more than that of the
average comparison group patient. Because the effect
size is in units of standard deviations, it can be con-
verted directly into a percentage representing nonover-
lap of the distributions (41 ). For example, an effect size
of 1 .00 would indicate that the average dynamic ther-
apy patient was better off than 84% of the comparison
group patients.
As an aid in interpreting effect sizes, Cohen (41 ) pro-
vided benchmarks for his d statistic. A value of 0.2 rep-
resents a small effect for the behavioral sciences, 0.5 a
medium effect, and 0.8 a large effect. I will present al-
ternative views of the meaning of different effect sizes
in the Conclusions section.
RESULTS
Table 2 presents the effect sizes from the comparisons
of brief dynamic therapy and the other conditions tested.
Waiting List
Five comparisons, from four studies (22, 25, 27, She-
fler and Dasberg), are included in the results shown in
table 2. The effect size for target symptoms is especially
large, has been reported in previous studies (e.g., 42),
indicating that these measures are particularly sensitive
to change.
When the effect sizes are translated into percentages,
the data indicate that in terms of target symptoms the
average brief dynamic therapy patient was better off
than 86% of the waiting list patients. In terms of gen-
eral symptoms (SCL-90), the average brief dynamic
therapy patient was better off than 79% of the waiting
list patients. For social adjustment, the average brief dy-
namic therapy patient was also better off than 79% of
the waiting list patients.
Nonpsychiatric Treatments
Five studies compared brief dynamic therapy to some
form of comparison condition that was not a standard
psychiatric treatment modality (i.e., not medication or
an alternative psychotherapy). The expectation in all of
these studies was that dynamic therapy would produce
greater effects than the comparison conditions. These
comparison conditions ranged from the placebo and
clinical management condition of the NIMH Treatment
of Depression Collaborative Research Program study
(23) to standard drug counseling (30), mutual self-help
groups (31), and two studies employing a low-contact
treatment (24, 28).
Overall, the average effect sizes indicate little difference
between the outcomes of brief dynamic therapy and these
PAUL CRITS-CHRISTOPH
Am JPsychiatry 149:2, February 1992 155
comparison conditions. The largest average effect was for
target symptoms, indicating that the average dynamic
therapy patient was better off than 62% of the compari-
son group patients. The average effects, however, may be
misleading, as there is considerable variation among these
studies, perhaps as a result of the various types of com-
parison conditions, patient groups, etc. The effect sizes
were fairly large in two studies, i.e., a comparison of the
effects of dynamic therapy and drug counseling (30) on
general psychiatric symptoms (effect size=O.74) and a
comparison of the effects of interpersonal therapy and
low-contact therapy (28) on target symptoms (effect
size=O.67). In addition, in three of the five studies there
was a much higher dropout rate from the alternative treat-
ment than from dynamic therapy.
Other Psychiatric Treatments
Six studies (23, 26-30) yielded nine comparisons of
brief dynamic therapy and other standard psychiatric
treatments. The alternative treatments included medi-
cation and various cognitive and behavioral therapies.
The predominant trend is clear: only small differences
between brief dynamic therapy and other treatments
are generally apparent. No individual study effect size
was greater than 0.30.
The single comparison (25) of one form of dynamic
therapy (Davanloo approach) with another (Pollack
and Homer’s brief adaptation therapy) yielded no evi-
dence for the superiority of one treatment.
CONCLUSIONS
Meaning of Effect Sizes
The effect sizes shown by this analysis can be de-
scribed as large according to Cohen’s definition (41).
The translation to percentages communicates even
more directly the extent to which a patient is much bet-
ter off receiving dynamic therapy than no treatment.
These effects are somewhat larger than the effect size
(0.68) presented by Smith and Glass (2) for the general
effects of psychotherapy. The heterogeneous treatments
and control conditions used by Smith and Glass, how-
ever, may account in part for this difference. Alterna-
tively, it is possible that the more recent studies re-
viewed here, which controlled the delivery of treatment
to a greater extent, demonstrated larger changes than
did the naturalistic assessments of dynamic therapy in
the past.
In addition to Cohen’s definitions, the magnitude of
the effects of dynamic therapy can be understood
through comparison with other types of interventions
in medicine. Rosenthal (43) has performed such com-
parisons using the Smith and Glass meta-analysis (2).
He demonstrated that the 0.68 effect size found by
Smith and Glass was larger than the effect sizes for sev-
eral well-known clinical trials in medicine which were
halted before completion because it was decided the ef-
fect was large enough that it would be unethical to
withhold from some patients a clearly efficacious treat-
ment. As mentioned, the effect sizes for dynamic ther-
apy versus waiting list control conditions found here
were even larger than the effect size reported by Smith
and Glass.
Obviously, with aspects of patient care that involve
death or serious illness, even a small treatment effect
might be deemed to be especially important. Neverthe-
less, it is informative to see that the effects of brief dy-
namic therapy (and psychotherapies in general) are
strong in comparison to the effects of such medical
treatments.
Other forms of psychotherapy (besides dynamic)
were not compared to waiting list control conditions in
this investigation. Dobson (5) reported an average ef-
fect size of 2.15 (based on Beck Depression Inventory
scores) for comparisons of Beck’s cognitive therapy to
a waiting list or no treatment for depression. Although
this value is substantially larger than the effect sizes re-
ported here for comparisons of brief dynamic therapy
and waiting list conditions, there is as yet no direct evi-
dence for superiority of cognitive therapy, because the
two studies that directly compared a manual-based dy-
namic or interpersonal therapy to cognitive therapy for
the treatment of depression (23, 26) showed no differ-
ences between these treatments. It is likely that depres-
sion is a condition that responds well to treatments and
that this accounts for the large effect size reported by
Dobson (5) for comparisons of cognitive therapy to a
waiting list and to no treatment.
The data reported here on dynamic therapy’s lack of
superiority to other standard psychiatric treatments are
perhaps not surprising in light of previous reviews of
the psychotherapy literature which have shown that
various treatments do not differ in their effectiveness (2,
3). The fact that these conclusions hold up in studies
where the treatment variable has been standardized and
controlled through the use of treatment manuals is,
however, still noteworthy. Although this finding ap-
pears robust, it is probably unwarranted to generalize
this conclusion across all patient populations, outcome
measures, and treatment types. For example, brief dy-
namic therapy has yet to be compared with exposure
and response prevention in the treatment of obsessive-
compulsive disorder. It is conceivable that, given the
documented success of this behavioral treatment, expo-
sure and response prevention would be shown to be su-
perior to dynamic therapy on measures of compulsive
behaviors. In brief, these unresearched questions are
still unanswered.
Despite the strong effects in the comparisons of brief
dynamic therapy and waiting list control conditions,
there was less evidence for brief dynamic therapy’s su-
periority to other treatment conditions that were not
standard psychiatric treatments. Some studies, how-
ever, did yield evidence of superior efficacy. Moreover,
in several studies there were higher dropout rates with
the alternative treatments than with dynamic therapy.
However, given the variation in the findings and the
BRIEF DYNAMIC THERAPY
156 Am /Psychiatry 149:2, February 1992
diverse nature of the control conditions and patient
groups, it is difficult to draw meaningful conclusions
from this small set of studies. Clearly, more research is
needed to assess whether brief dynamic therapy is more
beneficial than drug counseling, self-help groups, and
the like.
The rigorous requirements (such as careful selection,
training, supervision, and monitoring of therapists and
the use of treatment manuals) that have been imple-
mented in psychotherapy outcome research and were
used as criteria for selection of the studies reviewed here
may limit the generalizability of the findings to the out-
patient practice of dynamic therapy. On the one hand,
the careful selection and supervision of therapists is
likely to ensure that a high-quality service is delivered;
on the other hand, therapists in practice are not bound
by a particular treatment guide or manual. Therapists
in practice would therefore have more flexibility in their
clinical work, which could conceivably lead to more fa-
vorable outcomes. Thus, what is gained in knowledge
about the use of specific treatments under controlled
conditions might also be a loss in terms of generalizabil-
ity to the actual current practice of dynamic psycho-
therapy. Not only were the studies reviewed here per-
formed under highly controlled conditions, but some of
the forms of treatment studied may not reflect dynamic
therapy as it is typically practiced. Interpersonal ther-
apy (34), for example, was investigated in four of the
1 1 studies, but this approach may be quite distant from
the psychoanalytically oriented forms of dynamic ther-
apy more commonly practiced. This aspect of the data
base for this review further limits its generalizability.
What Is Missing?
An examination of what is not contained in this re-
view can help further illuminate its limitations and
point the way for future research. Lack of information
on certain treatments, patient groups, outcome meas-
ures, various treatment lengths, follow-up assessments,
and interaction hypotheses will be discussed in turn.
Most notably missing from studies of brief dynamic
therapy are evaluations of Sifneos’s well-known ap-
proach (15). Although two studies comparing this ap-
proach to a waiting list control (44, 45) and a study
involving Sifneos’s approach and two other brief dy-
namic therapies (46) have been conducted, data from
these studies were not presented in a manner amenable
to a meta-analysis. Means and standard deviations (or
significance tests) were not presented for these studies,
and the types of outcome measures used here were not
employed. In addition, two studies (44, 45) used train-
ees as therapists. These studies, however, are important
contributions to the literature on brief dynamic ther-
apy. Unlike many of the studies used for the meta-
analysis reported here, these three studies of Sifneos’s
approach involved patients who were carefully selected
on the basis of explicit criteria for short-term dynamic
therapy. Also, the investigators in these studies at-
tempted to operationalize psychodynamic change,
rather than examining only symptomatic change. All
three studies showed considerable positive evidence for
the effectiveness of Sifneos’s approach. More research
is needed to document the size of the effect on different
forms of outcome.
Another limitation of the studies I have reviewed is
that many of the patient groups were types that clini-
cians believe are particularly difficult to treat. Included
were two studies of opiate addicts (24, 30), one of co-
caine abusers (29), and one of patients with personality
disorders (25). Although brief dynamic therapy may be
most indicated for adjustment disorders, some anxiety
disorders, and relationship conflicts, studies focusing
on these problems have not been performed, with the
exception of one study on posttraumatic stress disorder
(27). A larger effect size for brief dynamic therapy
might be expected if these “ideal” patient types were
chosen for research.
The outcome measures of target symptoms, general
symptoms (SCL-90), and social adjustment (Social Ad-
justment Scale) were selected because these measures
are commonly used and it is preferable to calculate ef-
fect sizes within outcome domains rather than combin-
ing results across domains and measures. These meas-
ures, however, do not capture the specific areas where
short-term dynamic therapy theoretically should be
most successful. Studies of more theoretically relevant
measures-dynamic conflicts, transference themes, and
relationship patterns (e.g., 47)-are beginning to ap-
pear, but these measures have not yet been used in effi-
cacy studies. With the exception of Klerman’s interper-
sonal therapy, which focuses on manifest social
adjustment, the other brief dynamic therapies were
evaluated in terms of outcome measures that were not
the central focus of treatment. Therapies such as medi-
cation, cognitive therapy, and behavioral treatments,
which do focus directly on overt symptoms, might be
expected to achieve larger changes on these measures.
Larger effects for brief dynamic therapy may be evident
once theoretically relevant measures are studied. Al-
though previous research has often failed to detect
specificity of the effects of different treatments in theo-
retically important outcome domains, the lack of such
findings is likely to be a function of the difficulty in
measuring these domains.
In searching for changes in measures that are theo-
retically important for brief dynamic therapy, it is es-
sential that patients receive an adequate “dose” of
treatment. Howard et al. (48) pooled data from many
studies of eclectic therapy to determine the relationship
between “dose” of psychotherapy and symptomatic
outcome. Their results indicated that 75% of patients
achieve an effective “dose” of therapy (i.e., symptom
relief) within 26 sessions. In psychodynamic therapy,
important therapeutic work (the working through
process) is likely to take place after the initial symptom
relief, and an adequate period for such working through
is probably necessary for dynamic change to occur.
Thus, studies designed to test the efficacy of dynamic
therapy should not only measure the theoretically im-
PAUL CRITS-CHRISTOPH
Am /Psychiatry 149:2, February 1992 157
portant variables but should also allow enough time for
change in these variables to occur.
It may not be necessary, of course, for the working
through to occur only during psychotherapy sessions.
A very brief (12-session) dynamic therapy might begin
the working through, and then the successful patient
might be able to continue the process after the termina-
tion of treatment (18). Research on this process would
have to include assessment of therapeutic benefits some
time after the treatment formally ends. As mentioned
earlier, few follow-up results are available. Ideally, the
follow-up period should be free of further treatment,
since this additional treatment would confound inter-
pretation of the benefits apparent at the follow-up as-
sessment. No studies performed to date have attempted
to control the follow-up period. With patients in need
of further treatment, however, the ethical problems of
withholding treatment would have to be considered.
A final issue not taken up by this review is the role of
other factors that might moderate the relationship be-
tween treatment type and outcome. Investigators would
do well to look for interactions between patient charac-
teristics and treatment types in evaluating efficacy data.
The study by Piper et al. (22) on changes associated
with dynamic therapy and a waiting list control condi-
tion for patients with high and low “quality of object
relations” is an excellent example of the examination of
interactions. To a certain extent this issue reduces to the
problem discussed earlier, defining the appropriate pa-
tient group for dynamic therapy. It is unlikely, however,
that only patients who seem theoretically suitable for
dynamic therapy will be selected in future studies, espe-
cially considering the current diagnostic emphasis in
psychiatry and the resulting constraints on federally
funded research. Rather than select patients on the basis
of suitability for dynamic therapy, researchers are more
likely to measure variations in patient characteristics
and relate these variables to outcome within and across
treatment modalities.
The construct of “suitability for dynamic therapy”
itself needs further research. Although there are clinical
indicators of this quality, no reliable and valid assess-
ment device has yet been created that can be used to test
these clinical hypotheses. Actually, it may be of interest
to ask not only what types of patients are in general
suitable for dynamic therapy but also what types of pa-
tients do best and worst in brief dynamic therapy com-
pared to an alternative therapy. For example, when one
compares the process of dynamic therapy and cognitive
therapy, a hypothesis that comes to mind is that pa-
tients who have more interest in examining the subtle,
complex meanings of events and interpersonal transac-
tions are a better match (and therefore would have bet-
ter outcomes) with dynamic therapy and that patients
who are more task oriented and see things in a more
cut-and-dried fashion would probably prefer a straight-
forward, logical therapy, such as cognitive therapy,
which gives a clear, direct explanation of their problems
and proceeds in steps with regular homework assign-
ments. By asking these more and more refined ques-
tions, researchers stand a better chance of uncovering
the conditions under which dynamic therapies and
other therapies are most effective.
REFERENCES
1. Garfield SL, Bergin AE: Introduction and historical overview, in
Handbook of Psychotherapy and Behavior Change: An Empiri-
cal Analysis, 3rd ed. Edited by Garfield SL, Bergin AE. New
York, John Wiley & Sons, I 986
2. Smith ML, Glass GV: Meta-analysis of psychotherapy outcome
studies. Am Psychol I 977; 32:752-760
3. Luborsky L, Singer B, Luborsky L: Comparative studies of psy-
chotherapy: is it true that “everybody has won and all must have
prizes”? Arch Gen Psychiatry 1975; 32:995-1008
4. Miller RC, Berman JS: The efficacy of cognitive behavior thera-
pies: a quantitative review of the research evidence. Psychol Bull
1983; 94:39-53
S. Dobson KS: A meta-analysis of the efficacy of cognitive therapy
for depression. JConsult Clin Psychol 1989; 57:414-419
6. Luborsky L, DeRubeis R: The use of psychotherapy treatment
manuals: a small revolution in psychotherapy research style. Clin
Psychol Rev 1984; 4:5-14
7. Strupp HH, Hadley SW: Specific versus non-specific factors in
psychotherapy. Arch Gen Psychiatry 1979; 36:1 12-1 126
8. Sloane RB, Staples FR, Cristol AH, Yorkston NJ, Whipple K:
Psychotherapy Versus Behavior Therapy. Cambridge, Mass,
Harvard University Press, I 975
9. Pilkonis PA, Imber SD, Lewis P, Rubinsky P: A comparative out-
come study of individual, group, and conjoint psychotherapy.
Arch Gen Psychiatry 1984; 41:431-437
10. Piper WE, Debbane EG, Bienvenu J-P, Garant J: A comparative
study of four forms of psychotherapy. JConsult Clin Psychol
1984; 52:268-279
1 1. Crits-Christoph P, Baranackie K, Kurcias J, Beck A, Carroll K,
Perry K, Luborsky L, McLellan AT, Woody GE, Thompson L,
Gallagher D, Zitrin C: Meta-analysis of therapist effects in psy-
chotherapy outcome studies. Psychotherapy Research (in press)
12. Crits-Christoph P, Mintz J: Implications of therapist effects for
the design and analysis of comparative studies of psychothera-
pies. JConsult Clin Psychol 1991; 59:20-26
13. Luborsky L: Principles ofPsychoanalytic Psychotherapy: A Man-
ual for Supportive-Expressive Treatment. New York, Basic
Books, 1984
14. Strupp HH, Binder J: Psychotherapy in a New Key. New York,
Basic Books, I 984
15. Sifneos PE: Short-Term Psychotherapy and Emotional Crisis.
Cambridge, Mass, Harvard University Press, I 972
16. Davanloo H: Short-Term Psychotherapy. New York, Jason Ar-
onson, 1980
17. Mann J: Time-Limited Psychotherapy. Cambridge, Mass, Har-
yard University Press, I 973
18. Horowitz MJ: Stress Response Syndromes. New York,Jason Ar-
onson, I 976
19. Crits-Christoph P, Barber J(eds): Handbook of Short-Term Dy-
namic Therapies. New York, Basic Books, 1991
20. Glass GV, McGaw B, Smith ML: Meta-Analysis in Social Re-
search. Beverly Hills, Calif, Sage Publications, 1981
21. Budman SH: Introduction, in Forms of BriefTherapy. Edited by
Budman SH. New York, Guilford Press, 1981
22. Piper WE, Azim HF, McCallum M, Joyce AS: Patient suitability
and outcome in short-term individual psychotherapy. JConsult
Clin Psychol 1990; 58:475-481
23. Elkin I, Shea MT, Watkins JT, Imber SD, Sotsky SM, CollinsJF,
Glass DR, Pilkonis PA, Leber WR, Docherty JP, Fiester SJ, Par-
loff MB: National Institute of Mental Health Treatment of Dc-
pression Collaborative Research Program: general effectiveness
of treatments. Arch Gen Psychiatry 1989; 46:971-982
24. Rounsaville BJ, Glazer W, Wilbur CH, Weissman MM, Kleber
HD: Short-term interpersonal psychotherapy in methadone-
maintained opiate addicts. Arch Gen Psychiatry 1983; 40:629-
636
BRIEF DYNAMIC ThERAPY
1S8 Am/ Psychiatry 149:2, February 1992
25. Winston A, Pollack J, McCullough L, Flegenheimer W, Kesten-
baum R, Trujillo M: Brief dynamic psychotherapy of personality
disorders. JNerv Ment Dis 1991; 179:188-193
26. Thompson LW, Gallagher D, Breckenridge JS: Comparative ef-
fectiveness of psychotherapies for depressed elders. JConsult
Clin Psychol 1987; 55:385-390
27. Brom D, Kieber RJ, Defares PB: Brief psychotherapy for posttrau-
matic stress disorders. JConsult Clin Psychol 1989; 57:607-612
28. DiMascio A, Weissman MM, Prusoff BA, Neu C, Zwilling M,
Klerman GL: Differential symptom reduction by drugs and psy-
chotherapy in acute depression. Arch Gen Psychiatry 1979; 36:
1450-1456
29. Carroll KM, Rounsaville BJ, Gawin FH: A comparative trial of
psychotherapies for ambulatory cocaine abusers: relapse preven-
tion and interpersonal psychotherapy. Am JDrug Alcohol Abuse
1991; 17:229-247
30. Woody GE, Luborsky L, McLellan AT, O’Brien CP: Corrections
and revised analyses for psychotherapy in methadone mainte-
nance patients. Arch Gen Psychiatry 1990; 47:788-789
31. Marmar CR, Horowitz MJ, Weiss DS, Wilner NR, Kaltreider
NB: A controlled trial of brief psychotherapy and mutual-help
group treatment of conjugal bereavement. Am JPsychiatry
1988; 145:203-209
32. Pollack J, Homer A: Brief adaptation-oriented psychotherapy, in
Clinical and Research Issues in Short-Term Dynamic Psycho-
therapy. Edited by Winston A. Washington, DC, American Psy-
chiatric Press, 1985
33. Malan DH: The Frontier of Brief Psychotherapy. New York, Ple-
num, 1976
34. Klerman GL, Weissman MM, Rounsaville B, Chevron ES: Inter-
personal Psychotherapy of Depression. New York, Basic Books,
1984
35. Elkin I, Parloff MB, Hadley SW, Autry JH: NIMH Treatment of
Depression Collaborative Research Program: background and
research plan. Arch Gen Psychiatry 1985; 42:305-3 16
36. Lambert MJ, Hatch DR, Kingston MD, Edwards BC: Zung,
Beck, and Hamilton Rating Scales as measures of treatment out-
come: a meta-analytic comparison. JConsult Clin Psychol 1986;
54:54-5 9
37. Rosenthal R, Rubin DB: Meta-analytic procedures for combin-
ing studies with multiple effect sizes. Psychol Bull 1986; 99:400-
406
38. Battle CC, Imber SD, Hoehn-Saric R, Stone AR, Nash EH, Frank
J: Target complaints as a criteria of improvement. Am JPsycho-
ther 1966; 20:184-1 92
39. Derogatis LR, Lipman RS, Covi L: SCL-90: an outpatient psychi-
atric rating scale-preliminary report. Psychopharmacol Bull
1973; 9:13-28
40. Weissman MM, Bothwell 5: Assessment of social adjustment by
patient self-report. Arch Gen Psychiatry 1976; 33:1111-111S
41 .Cohen J: Statistical Power Analysis for the Behavioral Sciences.
New York, Academic Press, 1977
42. Luborsky L, Crits-Christoph P, Mintz J, Auerbach A: Who Will
Benefit From Psychotherapy? Predicting Therapeutic Outcomes.
New York, Basic Books, 1988
43. Rosenthal R: How are we doing in soft psychology? Am Psychol
1990; 45:775-777
44. Sifneos PE: Short-Term Dynamic Psychotherapy: Evaluation and
Technique, 2nd ed. New York, Plenum, 1987
45. Sifneos PE: Learning to solve emotional problems: a controlled
study of short-term anxiety-provoking psychotherapy, in The
Role of Learning in Psychotherapy. Edited by Porter R. Boston,
Little, Brown, I 968
46. Barth K, Nielsen G, Haver B, Havik OE, Molstad E, Rogge H,
Skatun M: Comprehensive assessment of change in patients
treated with short-term dynamic psychotherapy: an overview.
Psychother Psychosom 1988; 50:141-150
47. Luborsky L, Crits-Christoph P: Understanding Transference:
The Core Conflictual Relationship Theme Method. New York,
Basic Books, 1990
48. Howard KI, Kopta SM, Krause MS, Orlinsky D: The dose-effect
relationship in psychotherapy. Am Psychol 1986; 41:159-164
