Evidence-Based Psychodynamic Treatments for Anxiety Disorders: A Review

Abstract

Anxiety disorders are the most prevalent mental disorders in the United States, are the most ­frequently encountered disorders in primary mental health services, and are hypothesized to be the most common presenting problems in psychodynamic therapeutic practices [1]. In fact, when combining both national and international data, the best estimate for lifetime prevalence of anxiety disorders is 16.6% [2]. These disorders include generalized anxiety disorder (GAD), panic disorder (PD), post-traumatic stress disorder (PTSD), obsessive–compulsive disorder (OCD), social phobia, agoraphobia, and specific phobias [2].

Full text

117
R.A. Levy et al. (eds.), Psychodynamic Psychotherapy Research: Evidence-Based Practice
and Practice-Based Evidence, Current Clinical Psychiatry, DOI 10.1007/978-1-60761-792-1_7,
© Springer Science+Business Media, LLC 2012
J. Slavin-Mulford, Ph.D. (*)
Department of Psychiatry , Massachusetts General Hospital,
Harvard Medical School , Boston , MA , USA
e-mail: jenelle.slavin@gmail.com
M. J. Hilsenroth, Ph.D.
Derner Institute of Advanced Psychological Studies ,
Adelphi University , Garden City , NY , USA
e-mail: hilsenro@adelphi.edu
Keywords Anxiety disorders • Effect size • GAD • Panic • Psychodynamic psychotherapy • PTSD
Prevalence, Course, and Comorbidity of Anxiety Disorders
Anxiety disorders are the most prevalent mental disorders in the United States, are the most frequently
encountered disorders in primary mental health services, and are hypothesized to be the most com-
mon presenting problems in psychodynamic therapeutic practices [ 1 ] . In fact, when combining both
national and international data, the best estimate for lifetime prevalence of anxiety disorders is
16.6% [ 2 ] . These disorders include generalized anxiety disorder (GAD), panic disorder (PD), post-
traumatic stress disorder (PTSD), obsessive–compulsive disorder (OCD), social phobia, agorapho-
bia, and specifi c phobias [ 2 ] .
Importantly, however, there is a high percentage of comorbidity among these disorders. In a
large-scale survey of mental health, Andrews et al. [ 3 ] found that panic disorder, social phobia,
GAD, and PTSD all had signifi cant odds ratios (ORs) of occurrence with one another. Anxiety dis-
orders have also been found to be highly comorbid with other disorders. For example, Andrews
et al. [ 3 ] reported that 27.8% of patients with an anxiety disorder also had a comorbid personality
disorder and 8.7% had a comorbid substance disorder. Moreover, in a prospective longitudinal cohort
study, Moffi tt et al. [ 4 ] found that of lifetime anxiety cases, 72% had lifetime major depressive dis-
order. This high rate of comorbidity is an important issue given that ample research has shown that
comorbidity is associated with higher levels of severity, greater service utilization, and a poorer
prognosis [ 3– 6 ] .
Chapter 7
Evidence-Based Psychodynamic Treatments
for Anxiety Disorders: A Review
Jenelle Slavin-Mulford and Mark J. Hilsenroth

118 J. Slavin-Mulford and M.J. Hilsenroth
The Need for Empirical Research on Treatments for Anxiety Disorders
Anxiety disorders are associated with severe impairments in functioning and have signifi cant emo-
tional and fi nancial costs both on personal and societal levels [ 7– 9 ] . For example, patients with panic
disorder have higher rates of morbidity and health care utilization than patients both with and with-
out other psychiatric disorders [ 10 ] leading them to account for 20% of all emergency room visits
[ 11 ] . Similarly, when compared with people suffering from 25 other mental disorders or common
physical conditions, people with a diagnosis of GAD report missing the most work [ 7 ] . Thus, given
the high prevalence, impairment, and cost of anxiety disorders, it is essential to continue to develop
and test treatments for anxiety.
Although pharmacological and CBT treatments are widely used to treat anxiety [ 12 ] and there is
evidence that both are effective (e.g. [ 13, 14 ] ), there are limitations with both. Concerns associated
with pharmacological treatments include: frequent relapse when medication is discontinued (e.g.
[ 14 ] ), unwanted side-effects (e.g. [ 15, 16 ] ), and potential dependency of some drugs [ 14, 17 ] . In
addition, some patients fail to respond to CBT, continue to experience symptoms, terminate early, or
relapse after treatment (e.g. [ 18, 19 ] ). Further, a few studies suggest that, at times, CBT may actually
worsen the anxiety of some patients [ 20 ] . Thus, while pharmacological and CBT interventions have
been successful in treating anxiety disorders, like all forms of treatment, they are limited, have prob-
lems, and are not helpful for all patients. Thus, it is important to test other types of treatment. Given
that psychodynamic therapy is widely used to treat anxiety [ 12 ] , its effi cacy and effectiveness is
important to test. Although there is not as extensive a research base for psychodynamic treatments
for anxiety as there is for pharmacological interventions and CBT, the work done thus far is promis-
ing (e.g. [ 21 ] ). This empirical literature will be reviewed later in the chapter after a brief review of
psychodynamic theory of anxiety.
A Brief Account of the Psychodynamic Understanding of Anxiety
All of the major psychodynamic schools (e.g., Freudian, object relations, self-psychology) have
written about anxiety with writers from Freud through Klein, Fairbairn, Sullivan, and Kohut, and all
have made important contributions [ 22 ] . Clearly, a short summary for this chapter cannot capture the
many essential differences that exist among psychodynamic theorists. However, a review of some of
the most signifi cant contributions made by psychodynamic theory in the understanding of anxiety is
important as these concepts have implications for treatment. These contributions include: the under-
lying meaning of symptoms, unconscious confl ict, defense, and the impact of object and interper-
sonal relationships.
The meaning of symptoms, including both unconscious confl ict and defense, originated with
Freud and continues to be at the heart of most psychodynamic theories. Freud’s [ 23 ] signal theory
suggested that anxiety, which has arisen from a perceived psychological danger, signals an antici-
pated threat to the ego. As a result, defenses are activated so that a dangerous intrapsychic situation
does not become traumatic. If the signal anxiety and/or defenses fail to work properly, an anxiety
attack may ensue leaving the individual vulnerable [ 24 ] . Freud suggested that treatment should bring
the threat into consciousness so that the patient can understand that the danger is not as great as origi-
nally perceived. Some of the threats he discussed were apprehension about object loss, castration
anxiety, and superego anxiety [ 22 ] . For Freud, object loss anxiety is the fear of, or reaction to, sepa-
ration from a need-gratifying other such as the mother. Castration anxiety is also the fear of being
separated from a highly valued object. However, it is specifi c to concern over bodily injury (typi-

1197 Evidence-Based Psychodynamic Treatments for Anxiety Disorders: A Review
cally the male genitalia) and loss of control and power. Finally, superego anxiety or moral anxiety
is a fear of negative self-evaluation from the conscience or society [ 23 ] .
Following Freud, several dynamic writers have suggested treatments for anxiety that focus on the
role of the unconscious and the use of defense. For example, in discussing panic disorder from a
psychodynamic perspective, it has been suggested that panic symptoms carry a specifi c emotional
signifi cance and are related to intense unconscious confl icts and their defenses [
25, 26 ] . Also, like
Freud, these authors suggest that typical areas of confl ict for panic patients include anger, sexual
desires, and separation [ 25, 26 ] . Similarly, contemporary psychodynamic work on GAD suggests
that worry about current events functions as a defensive avoidance against thinking about more dif-
fi cult issues [ 27 ] . This emphasis on interpersonal and object relationships leads to the next fi gural
contribution of psychodynamic theory.
Object and interpersonal relationships are at the center of several dynamic theories of anxiety. In
fact, although Sullivan, Fromm-Reichmann, Klein, Fairbairn, and Bowlby explain different aspects,
functions, and types of anxiety, internalized object relations and interpersonal relationships are
paramount in each writer’s theory. For example, whereas Sullivan [ 28 ] proposed that anxiety origi-
nates in the anticipation of disapproval from a primary caregiver early in life, Fairbairn [ 29 ] sug-
gested that anxiety centers around separation confl ict. Similarly, Fromm-Reichmann [ 30 ] emphasized
the role of distorted relational views in both the etiology and perpetuation of anxiety whereas Klein
[ 31, 32 ] focused on an infant’s fear of death when he/she is unable to evoke the primary caregiver
on demand. In yet a different vein, Bowlby [ 33 ] focused on the relationship between attachment
style and anxiety.
As highlighted in the discussion of Sullivan, Fromm-Reichmann, Klein, Fairbairn, and Bowlby,
dynamic theorists often link anxiety with object relations and/or interpersonal relationships.
Importantly, this connection has also been supported by empirical research. For example, studies
have shown that worry is often related to interpersonal problems, especially when the worrier is
overly nurturing toward others [ 34 ] and that worry content among GAD patients is more frequently
about relational issues than any other topic [ 35 ] . In addition, insecure attachment patterns have been
clinically reported for many patients with anxiety disorders [ 33, 36 ] . Further, patients suffering from
panic disorder frequently describe controlling or critical parents (e.g. [ 37 ] ). These fi ndings support
psychodynamic theorists’ suggestion that there is a link between anxiety and object relations or
interpersonal relationships. The implications of this connection can be seen in prominent psychody-
namic treatment such as Luborsky’s [ 38 ] Core Confl ictual Relationship Theme (CCRT) which
focuses on patients’ cyclical relational patterns. In sum, psychodynamic theories provide an expla-
nation for the etiology and pathogenesis of anxiety that often includes an emphasis on the underlying
meaning of symptoms, unconscious confl ict, defense, and both object relations and interpersonal
relationships. In order to illustrate how this type of understanding impacts treatment, a clinical
example will now be provided.
Clinical Example
Sara, a 29-year-old homemaker, presented at the emergency room complaining of panic attacks with
prominent symptoms of dizziness, sweating, shortness of breath, and tachycardia. She reported that
the fi rst panic attack occurred the previous week when her husband left for work and that the attacks
had reoccurred every morning since. The panic attacks always began with a fear that something ter-
rible was going to happen to either her husband or son and that one of them would die. In order to
begin to understand the meaning of Sara’s symptoms, her therapist began by exploring the context
in which her symptoms occurred.

120 J. Slavin-Mulford and M.J. Hilsenroth
Therapist: These panic attacks sound very frightening. My hope is that we can begin to understand them
together. Can you tell me about the fi rst morning that it happened?
Patient: I was making breakfast for my 2-year-old son. It was a normal morning.
Therapist: Can you recall what you were thinking about as you were making breakfast?
Patient: When my husband left that morning, he told me to “eat for two” since I am pregnant again. So, I
was thinking about that.
Therapist: When did you learn of your pregnancy?
Patient: Two weeks ago.
Therapist: When people learn they are pregnant, they often experience a wide range of emotions. How are you
experiencing the news?
Patient: It is only something wonderful. Well, I was originally upset about not being able to go back to
work, but now realize that I could not be happier to stay home with our children.
Sara’s response suggests that she may be confl icted about the birth of the new child and her con-
tinued role as a homemaker. On the one hand, she seems to want to be a “good” mother and to take
care of her children. Yet, on the other hand, she appears to be feeling a loss of autonomy about not
returning to work. Thus, her physical symptoms are likely a manifestation of the anger and loss she
feels when her husband leaves her alone with her son during the day and the subsequent guilt that
this anger induces. It seems that Sara is attempting to defend against these feelings by using reaction
formation (when an affect is disguised as its opposite) as seen by her statements that the pregnancy
is “only something wonderful.” When this fails and her resentment surfaces, she uses the defense of
undoing (performing an action which retracts and disavows a previously expressed affect) as evi-
denced by her statement that at fi rst she was upset about the news, but now “could not be happier.”
It seems that when both of these defenses fail, however, panic attacks ensue. In the following inter-
change, the therapist attempts to help Sara gain awareness into this confl ict and become more in
touch with her emotions.
Therapist: I can hear how seriously you take being a good mother and it is very clear to me how much you
love your family. It would be understandable, however, if there was also a part of you that feels
frustrated about not being able to return to work.
Patient: Oh I don’t know, maybe a little frustrated, but not really.
Therapist: Can you tell me about the “little” bit of frustration that is there?
Patient: I’m not sure.
Therapist: What would be the scariest part if you were frustrated?
Patient: My mom never left me alone even for a second as a child. She gave up everything for me. That is
how I knew she loved me. What would my children think if I were to go back to work and what
would my husband and mother think of me as well?
Sara’s object and interpersonal relationships are important in a psychodynamic therapy. Thus,
her therapist spent time exploring Sara’s enmeshed relationship with her mother as well as her
relationships with her husband and son. In addition, they examined Sara’s feelings and fantasies
about her unborn child. Through this work together, Sara and her therapist identifi ed a Core
Confl ictual Relationship Theme (CCRT [ 38 ] ) focusing on what Sara wants from relationships (the
wish; W), the response she expects from others (response from other; RO), and her subsequent
response both affectively and behaviorally (response of self; RS). The specifi c CCRT was that
Sara wants to be loving and giving to others while at the same time she wants to retain a sense of
independence and autonomy. However, she is afraid that others will take advantage of her, aban-
don her, or view her as being selfi sh. As a result, she ignores her own needs and is left feeling
trapped and angry.
Sara’s CCRT was examined in the context of outside relationships as well as in the therapeutic
relationship. The therapeutic relationship was also used to provide Sara with a different relational
experience in which support and open communication were fostered. It was inside of this relationship
that Sara and her therapist were able to make meaning of her physical symptoms, to bring her confl icts
into awareness, and to develop new, healthier ways of relating to herself and others. Although Sara’s

1217 Evidence-Based Psychodynamic Treatments for Anxiety Disorders: A Review
case is encouraging, it is important to examine whether psychodynamic therapy is effi cacious and
effective across many cases. Thus, this review will now turn to the empirical fi ndings of psychody-
namic therapy in the treatment of anxiety disorders.
The Outcome Evidence for Psychodynamic Treatments
of Anxiety Disorders
E f fi cacy of Psychodynamic Therapy Compared with No or Minimal
Treatment for Anxiety
Table 7.1 summarizes the two studies that have compared psychodynamic therapy to a control group
in the treatment of anxiety disorders [ 39, 40 ] . The purpose of this table, and all the other tables in the
empirical sections, is to highlight the magnitude of treatment effects for the primary patient and
clinically rated anxiety measures from each study. When measures of anxiety were not used, general
symptoms or overall improvement will be reported instead. The tables are not meant to constitute
meta-analyses or represent a complete reporting of all outcome effects for every variable examined
in the studies reported. Rather, they are provided to organize the magnitude of effects for the primary
anxiety variables (patient self-report and independent clinical ratings) in a clear manner.
As presented in Table 7.1 , Brom et al. [ 39 ] examined whether brief psychodynamic therapy,
focusing on solving intrapsychic confl ict resulting from trauma, was more effi cacious in the treat-
ment of PTSD than a wait-list control [ 39 ] . They found that on the primary outcome measure of
patient reported total symptoms, psychodynamic therapy was substantially more effective at termi-
nation ( p < 0.05, d = 0.88). 1
The second study was a pilot study in which GAD patients receiving brief supportive-expressive
(SE) psychodynamic therapy were compared to GAD patients receiving brief supportive nondirec-
tive therapy [ 40 ] . 2 The supportive nondirective therapy was conducted using Borkovec and Mathews’
Table 7.1 Psychodynamic treatments versus controls for anxiety disorders
Citation
N , disorder, and type of
dynamic treatment
N , disorder, and type of
comparison treatment Post Tx ES Follow-up Tx Es
Brom
et al. [
39 ]
29 PTSD
Brief (mean = 18.8 sessions)
psychodynamic treatment
focused on solving
intrapsychic confl ict
resulting from trauma
23 PTSD
Wait-list
Pt Rated Gen Symp
d = 0.88
N/A (no follow-up
data for wait-list
group)
Crits-Christoph
et al. [
40 ]
14 GAD
Supportive-expressive
brief psychodynamic
treatment (16 sessions)
13 GAD BAI
14 GAD HAM-A
Supportive, nondirective
therapy (16 sessions)
BAI d = −0.49
HAM-A d = 0.26
N/A
Gen Symp total general symptoms (patient rated), BAI Beck Anxiety Inventory (patient rated), HAM-A Hamilton
Anxiety Inventory (independent rater)
1 Brom et al. [ 39 ] will be discussed in more detail in the Psychodynamic Therapy versus CBT Treatments for Anxiety
section due to its further comparison of psychodynamic treatment to trauma desensitization.
2 The pilot study was published as a subsection of a larger report on the effects of SE in the treatment of GAD and will be
discussed in full in the Pre-post within group effectiveness of psychodynamic therapy in the treatment of anxiety Section.

122 J. Slavin-Mulford and M.J. Hilsenroth
[ 41 ] treatment manual and focused on directing patients’ attention toward their feelings while creat-
ing an accepting, nonjudgmental, and empathic environment. Supportive therapists were prohibited
from providing direct suggestion or coping techniques, and they were not instructed to focus on
interpersonal relationships. In contrast, SE focused directly on anxiety in the context of problematic
relational patterns.
Both groups showed large within-group improvements, and no signifi cant differences were found
between SE and supportive therapy on the BAI ( d = −0.49), or the HAM-A ( d = 0.26) [ 42 ] . However,
when only patients who achieved symptomatic remission (defi ned as a termination score on the
HAM-A of <7) were examined, a statistically signifi cant difference ( p < 0.05) was found such that
46% of the SE group versus 12.5% of the supportive group achieved remission. Importantly, the
outcome scores were much more variable in the SE group than in the supportive group (SD more
than 50% greater in SE group). The authors state that these fi ndings suggest SE therapy may work
exceptionally well for a subset of GAD patients, but not for other GAD patients. In contrast, sup-
portive treatment may lead to reliable gains for the majority of patients. This suggests that future
research (with more power than a pilot study) should examine whether there are patient characteris-
tics that moderate when psychodynamic therapy is more and less effective. It should also examine
what the specifi c active techniques in psychodynamic therapy are that cause it to lead to greater
remission than supportive therapy. Further, because no follow-up data were provided, it remains
uncertain whether the remitters in the SE group maintained their improvement.
Abbass et al. [
43 ] begin addressing the issues of power and long-term maintenance in their
large meta-analysis evaluating the effi cacy of short-term psychodynamic therapy (STPP) as com-
pared to controls (treatment as usual, medical management, psychotherapeutic support, minimal
psychological interventions, and wait-list). Their meta-analysis is not included in Table 7.1
because their review summarizes data from studies on a variety of mental disorders and does not
focus specifi cally on anxiety disorders. However, their fi ndings are important as they examine
STPP’s effi cacy in reducing anxiety symptoms. Among the 23 studies used in this review, 12 used
anxiety measures. Using these 12 studies in a random effects model, Abbass et al. found that anxi-
ety ratings showed signifi cant ( p < 0.05) and large treatment effects relative to controls at 3- to
9-month follow-up ( d = 0.96), but did not show signifi cant effects when follow-up was less than
3 months ( d = 0.72) or greater than 9 months ( d = 0.85). However, when one of the studies used in
the review [ 44 ] was excluded from analyses, the difference was signifi cant in both the short
( d = 0.96, p < 0.05) and long term ( d = 1.35, p = 0.05). Examining the results with the exclusion of
Sjodin et al. [ 44 ] is important as this study diverged markedly from the results of the other studies
in the review. This is not surprising given that it was a study conducted on peptic ulcer patients
before triple therapy for the eradication of helicobacter pylori was introduced. When this study is
taken out, the fi nding across the other 11 studies is that psychodynamic treatment is more effective
than controls in reducing anxiety symptoms. Although this is important, it is even more important
to determine how psychodynamic treatment compares to other active treatments. To address this
question, the chapter will now review the literature on psychodynamic versus pharmacological
treatments on anxiety.
E f fi cacy of Psychodynamic Treatments Combined or Alone in Relation
to Pharmacological Treatments for Anxiety Disorders
Table 7.2 summarizes the primary patient and clinical anxiety ratings for the randomized control
trial (RCT; [ 46 ] ) and naturalistic study [ 45 ] that have examined the effi cacy of psychodynamic
therapy alone or in combination with medications as compared with pharmacological interventions
only in the treatment of anxiety. These studies vary in the types of patients used, the type of psycho-

1237 Evidence-Based Psychodynamic Treatments for Anxiety Disorders: A Review
dynamic therapy provided, and the types of comparisons made. However, they are similar in that
both provide encouraging results for psychodynamic therapy in the treatment of anxiety.
As presented in Table 7.2 , Wiborg and Dahl [ 46 ] conducted an RCT in which adult outpatients
diagnosed with panic disorder were treated with 9 months of clomipramine or a combination of
clomipramine and 15 sessions of manualized psychodynamic psychotherapy based upon the con-
cepts and techniques of Davanloo (1978), Malan (1976; 1979), and Strupp and Binder (1984) [ 46 ] .
Results showed that, at the end of treatment, all patients who had received therapy and medication
were panic free. In addition, the combined group had signifi cantly less anxiety at the end of treat-
ment than the medication alone group as measured by the patient-rated Panic Attack and Anxiety
Scale (PAAS; p = 0.02, d = 0.55) and the external rated Hamilton Anxiety Inventory (HAM-A;
p = 0.001, d = 1.38). This superiority of combined treatment remained 9 months after treatment ended
on both the PAAS ( p = 0.02, d = 0.76) and HAM-A ( p = 0.03, d = 0.86). These fi ndings suggest that
psychodynamic therapy enhances the effects of pharmacological treatments. However, they do not
attend to the effectiveness of psychodynamic therapy in the absence of medication.
In order to fi ll in this gap, Ferrero et al. [ 45 ] compared the effects of GAD outpatients treated with
medication, brief Adlerian psychodynamic psychotherapy (B-APP), or combined therapy and medi-
cation. B-APP focused on exploring patients’ deep-seated needs, self-esteem, self-image, and rela-
tional patterns. Patients were assigned at intake to their treatment group by a psychiatrist considering
severity and mental ability in treatment placement. Nonetheless, there were no signifi cant differ-
ences between the groups for age, comorbidity of Axis I or II disorders, or ratings on the Hamilton
Rating Scale for Depression (HAM-D), Hamilton Rating Scale for Anxiety (HAM-A), Social and
Occupational Functioning Scale (SOFAS), or Clinical Global Impression (CGI) scores at the
beginning of treatment.
3
No signifi cant differences were found between the groups on level of improvement evaluated
across intake, termination, and 3 months after treatment ended on the CGI ( p = 0.21), HAM-A
( p = 0.31), HAM-D ( p = 0.24), and SOFAS ( p = 0.12) nor on rate of remission for HAM-A ( p < 0.42)
Table 7.2 Psychodynamic treatments versus pharmacological treatments for anxiety disorders
Citation
N , disorder, and type of
dynamic treatment
N , disorder, and type of
comparison treatment Post Tx ES Follow-up Tx Es
Ferrero
et al. [
45 ]
34 GAD post and 3 month 33 GAD post and 3 month HAM-A HAM-A
33 GAD 9 month 25 GAD 9 month d = −0.29 3 month d = 0.21
Brief Adlerian psychotherapy
(10–15 sessions)
Medication
(SSRI or SNRI/NaSSA)
9 month d = 0.13
Ferrero
et al. [
45 ]
34 GAD post and 3 month 20 GAD post and 3 month HAM-A HAM-A
33 GAD 9 month 18 GAD 9 month d = 0.21 3 month d = 0.46
Brief Adlerian psychotherapy
(10–15 sessions)
Brief Adlerian psychotherapy
(10–15 sessions) PLUS
Medication
9 month d = 0.19
Wiborg and
Dahl [ 46 ]
20 Panic manualized
psychodynamic treatment
based on Davanloo, Malan,
Strupp and Binder
(15 sessions) PLUS
clomipramine
20 Panic clomipramine
alone for 9 months
PAAS d = 0.55
HAM-A d = 1.38
9 months after
treatment
ended
PAAS d = 0.76
HAM-A d = 0.86
HAM-A Hamilton Anxiety Inventory (independent rater), PAAS panic attack and anxiety scale (patient rated)
3 It should be noted, however, that there was a nonsignifi cant indication that the therapy alone group had less severity
as indicated by the CGI at the onset of treatment ( p < .07).

124 J. Slavin-Mulford and M.J. Hilsenroth
or HAM-D ( p < 0.36). In addition, when effect sizes were examined for the primary anxiety measure,
the HAM-A, only small differences were found at termination of therapy. This lack of robust differ-
ences between the groups is related to the fact that all treatment groups showed progressive improve-
ment on all measures at 3 months post-treatment and they all showed long-term conservation of
many of these clinical benefi ts at 9 months post-treatment.
In fact, the only fi nding that signifi cantly differentiated the groups was that patients with a comor-
bid personality disorder made signifi cantly more improvement in the dynamic therapy group than
patients treated with medication alone at 3 months after treatment ended ( p = 0.04). This suggests
that dynamic therapy is as successful as medication on improvement and remission of many symp-
tom variables and that it is even more successful than medication in helping GAD patients with a
comorbid PD improve their social and occupational functioning. Thus, although any conclusions
drawn are tentative due to the small number of studies in this area, these fi ndings combined with
Wiborg and Dahl [
46 ] provide support for the effi cacy and effectiveness of psychodynamic treat-
ment for anxiety disorders when compared with pharmacological approaches. In order to further
explore the effi cacy and effectiveness of psychodynamic therapy, this chapter will now turn to
empirical research that has compared psychodynamic therapy to cognitive and behavioral therapy
for anxiety disorders.
E f fi cacy of Psychodynamic Therapy Compared with Cognitive,
Behavioral, or CBT for Anxiety
Table 7.3 summarizes results from the primary patient and clinical ratings for anxiety or overall
symptoms in studies that have compared psychodynamic therapy to cognitive [ 47, 48, 52 ] , behav-
ioral [ 39, 49, 51, 53 ] , and cognitive-behavioral [ 50 ] therapies in the treatment of anxiety. These
studies vary in the types of patients used, the kind of psychodynamic therapy provided, and the
measures used to assess change. Thus, it is not surprising that the results vary with methodology.
Overall, they provide mixed results for psychodynamic therapy in the treatment of anxiety [ 8, 39, 49,
51 ] , with a few studies raising some questions about the effi cacy or fi delity of the psychodynamic
treatments delivered [ 47, 48 ] .
Durham et al. [ 47, 48 ] is an example of a study that raises concerns about the effi cacy of psycho-
dynamic treatment for anxiety as compared to cognitive therapy. These investigators compared the
effects of cognitive therapy, psychoanalytic therapy, and anxiety management training in an RCT
with GAD adult outpatients. All treatments were delivered at 1 or 2 week intervals during a 6-month
period and were further divided into two groups, one termed the high contact condition (16–20 hours
of treatment) and the other termed the low contact condition (8–10 hours of treatment). They found
that psychoanalytic therapy led to signifi cant improvements on several different measures in both the
high and low contact groups, but that cognitive therapy led to greater improvements. Similarly, low
contact analytic therapy was signifi cantly worse than anxiety management training in lowering
patients’ symptoms on the STAI-T at termination ( d = −0.80) and 6 ( d = −0.90) and 12-month follow-
up ( d = −1.12).
It is important to note, however, that there were several methodological problems in Durham
et al.’s [ 47, 48 ] work. First, there were manuals in the cognitive therapy, but not in the analytic
therapy. Second, there were no checks of adherence to the treatment methods so fi delity to stated
specifi c treatments cannot be evaluated. Third, there was no assessment of therapist competence.
Finally, as Leichsenring et al. [ 54 ] note, there were signifi cant differences in therapist experience
for each group. The psychologists delivering cognitive therapy had between 2 and 10 years of spe-
cifi c training in the cognitive therapy they were providing. In contrast, the psychiatrists conducting

1257 Evidence-Based Psychodynamic Treatments for Anxiety Disorders: A Review
Table 7.3 Psychodynamic treatments versus cognitive, behavioral, or CBT for anxiety disorders
Citation N , disorder, and type of dynamic treatment N , disorder, and type of comparison treatment Post Tx ES Follow-up Tx ES
Brom et al. [
39 ] 29 PTSD Manualized Brief (mean = 18.8 sessions)
psychodynamic therapy focused on solving
intrapsychic conflict resulting from trauma
31 PTSD Trauma desensitization
(mean = 15 sessions)
Gen Symp
d = −0.26
Gen Symp
3 month d = 0.23
Durham et al. [
47, 48 ] 11 GAD post
12 GAD 6 month
10 GAD 12 month
AP (Analytic-based therapy; exploration of
symptoms, relationships, development,
transference, and resistance)
High contact: 16–20 sessions
15 GAD post and 6 months
13 GAD 12 months
CT (cognitive therapy based
on Beck and Emery)
High contact: 16–20 sessions
STAI-T
d = −0.84
STAI-T
6 month d = −0.71
12 month d = −1.33
Durham et al. [
47, 48 ] 13 GAD post
14 GAD 6 months
12 GAD 12 months
AP Low contact: 8–10 sessions
18 GAD post
14 GAD 6 and 12 months
CT Low contact: 8–10 sessions
STAI-T
d = −0.87
STAI-T
6 month d = −1.38
12 month d = −1.38
Durham et al. [
47, 48 ] 13 GAD post
14 GAD 6 months
12 GAD 12 months
AP Low contact: 8–10 sessions
14 GAD post and 6 months
12 GAD 12 months
Anxiety Management (8–10 sessions)
STAI-T
d = −0.80
STAI-T
6 month d = −0.90
12 month d = −1.12
Klein et al. [
49 ] 18 Phobic a
Supportive dynamic (26 weekly sessions) PLUS
imipramine
48 Phobic b
Behavior therapy (26 weekly sessions)
PLUS imipramine
Pt-ODI d = 0.04
IR-ODI d = −0.31
Leichsenring et al. [
50 ] 28 GAD Manualized STDP based on Luborsky
and Crits-Christoph (up to 30 sessions)
29 GAD Manualized CBT focused on
changing and controlling worry
(up to 30 sessions)
STAI-T d = −0.82
HAM-A d = −0.48
6 months
STAI-T d = −0.93
HAM-A d = −0.65
Milrod et al. [
8 ] 26 Panic
Panic-focused psychodynamic psychotherapy
(twice weekly for 12 weeks)
23 Panic
Relaxation Training
(twice weekly for 12 weeks)
PDSS d = 0.95
Pierloot and Vinck [
51 ] Nine Elevated scores on TMAS
Short-term psychodynamic treatment infl uenced
by Malan (20 sessions)
13 Elevated scores on TMAS
Systematic desensitization (20 sessions)
STAI-S d = −0.37
TMAS d = −0.49
STAI-S
3 months
STAI-S d = 0.33
TMAS d = −0.03
Gen Symp total general symptoms (patient rated), STAI-S/ T State-Trait Anxiety Inventory State (S)/Trait (T) (patient rated), HAM-A Hamilton Anxiety Inventory (independent
rater), Pt-ODI patient rating of overall degree of improvement, IR-ODI independent rating of overall degree of improvement, PDSS panic disorder severity scale (independent
rater), TMAS Taylor manifest anxiety scale
a Only 17 of the phobic patients in the ST Group were used in the IR-ODI calculation
b Only 44 of the phobic patients in the BT group were used in the IR-ODI calculation

126 J. Slavin-Mulford and M.J. Hilsenroth
the analytic treatment were training in psychoanalysis and were not reported to have had any special
training in brief therapeutic interventions even though the trial was limited to 6 months of
treatment.
Perhaps even more important than these methodological concerns is Durham, Chambers,
MacDonald, Power, and Major’s [
55 ] later fi nding that cognitive therapy’s initial superiority
over analytic therapy in the 1994 and 1999 studies disappeared by 8-year follow-up. Specifi cally,
they found that 36% of participants in the cognitive group had recovered and 42% of participants
in a combined control group of participants who had received either analytic therapy or anxiety
management had recovered. Unfortunately, because Durham and colleagues [
55 ] combined the
analytic therapy and anxiety management participants in the comparison at 8-year follow-up, it
is impossible to conclude how cognitive therapy directly compared to analytic therapy at
follow-up.
4
A direct comparison between short-term psychodynamic treatment and behavior therapy was
reported by Klein et al. [
49 ] . In this study, phobic patients were randomly assigned to imipramine
plus 26 weekly sessions of behavior therapy (BT), placebo pill plus 26 weekly sessions of BT, or
imipramine plus 26 weekly sessions of supportive dynamic therapy (ST). BT consisted primarily
of systematic desensitization and relaxation training whereas ST focused on expression of feelings
and discussion of interpersonal relationships, anxieties, and confl icts. Results showed that both ST
plus imipramine and BT plus imipramine led to robust improvements. Eighty-fi ve percent of
patients in the BT plus imipramine group and 89% of patients in the ST plus imipramine group
reported having made moderate to marked improvement. Similarly, ratings by an external rater
show that 82% of patients in the BT plus imipramine group and 76% of patients in the ST plus
imipramine group made moderate to marked improvement at termination. Neither of these differ-
ences were signifi cant ( p > 0.10). This suggests that behavioral therapy and short-term psychody-
namic therapy are equally effective in the treatment of phobias when combined with imipramine
and that both are highly effective in combination with imipramine. However, it remains unclear as
to whether either behavioral or psychodynamic therapy would be effective without imipramine and
whether comparing the two treatments without the inclusion of pharmacological treatment would
lead to differential effi cacy.
Pierloot and Vinck [ 51 ] also compared psychodynamic and behavioral therapies in the treatment
of anxiety. Like Klein et al. [ 49 ] , they found both to lead to some positive change, however, their
results were more varied. On two measures, there was a trend, although not signifi cant ( p > 0.05), in
which systematic desensitization led to more positive change at the end of treatment, but not at
3-month follow-up. Specifi cally, at termination, patients in the systematic desensitization group
tended to show greater positive change on the Taylor Manifest Anxiety Scale (TMAS) ( d = −0.49)
and on the State-Trait Anxiety Index, Form State (STAI-S) ( d = −0.37), than patients in the psycho-
dynamic therapy group. Yet, at 3-month follow-up, the difference had disappeared for the TMAS
( d = −0.03) and had trended toward a reversal for the STAI-S ( d = 0.33), with the psychodynamic
therapy group now reporting greater change. Moreover, the positive changes made between termina-
tion and follow-up were signifi cantly greater in the dynamic therapy group than they were in the
systematic desensitization group ( p < 0.05). These fi ndings combined with Durham and colleagues’
[ 55 ] results suggest that cognitive and behavioral treatments may lead to more immediate symptom
reduction. However, psychodynamic treatment may lead to continued improvement after termina-
tion and the difference between the two treatments may diminish.
4 Because there was no direct comparison, Durham et al. [ 55 ] is not included in the table.

1277 Evidence-Based Psychodynamic Treatments for Anxiety Disorders: A Review
Further support for this argument comes from Brom and colleagues’ [ 39 ] RCT which contrasted
wait-list control, psychodynamic therapy, hypnotherapy, and trauma desensitization in the treatment
of patients with PTSD. Similar to Pierloot and Vinck’s fi ndings, they found that psychodynamic
treatment had slightly weaker effects than trauma desensitization on combined scores of intrusion
and avoidance at post treatment ( d = −0.26). However, the therapeutic gains continued in the psycho-
dynamic treatment group, and at 3-month follow-up, the psychodynamic patients appeared to have
slightly stronger positive effects as compared to trauma desensitization ( d = 0.23).
When within-group effect sizes from pretest to follow-up were examined, psychodynamic treat-
ment had large effects on both the Intrusion ( d = 1.12) and the Avoidance ( d = 0.94) subscales. Trauma
desensitization also fared well when comparing within-group effect sizes from pretest to follow-up.
However, the differences between the two scales for trauma desensitization were greater on the
Intrusion subscale ( d = 1.07) than on the Avoidance subscale ( d = 0.69). Thus, Brom and colleagues
[ 39 ] conclude that both psychodynamic therapy and trauma desensitization are more effective in
treating PTSD than control comparisons, but that the treatments may lead to differential results
depending on the timing of assessments (i.e., at termination or follow-up) as well as the specifi c
measures utilized (i.e., what types of changes are being assessed). Specifi cally, trauma desensitiza-
tion may lead to more improvements in the short term whereas psychodynamic treatment may lead
to greater gains in the long term. Moreover, psychodynamic therapy and trauma desensitization may
be equally effective in reducing intrusion, but psychodynamic therapy may lead to greater gains in
the area of avoidance.
Leichsenring et al. [ 50 ] also found that the relative success of CBT and short-term psychody-
namic therapy (STDP) in the treatment of GAD depended on the measures used for assessment. To
avoid some of the methodological fl aws mentioned earlier in previous studies, Leichsenring and col-
leagues used experienced therapists well trained in their respective approaches, manuals in both
treatments, and adherence and competency checks to ensure that the therapies were being delivered
properly. They found signifi cant positive effects for all outcome measures, indicating that both CBT
and STDP led to improvements (effect sizes ranged from d = 0.41– d = 2.67). This was also true for
all outcome measures at 6-month follow-up, indicating that the improvements were retained. In fact,
when the within-group effect sizes for all measures used in this study are averaged, STDP shows a
large effect at termination ( d = 1.16) and 6-month follow-up ( d = 1.10). Average CBT within-group
effect sizes also show a large effect at termination ( d = 1.73) and 6-month follow-up ( d = 1.71).
When the two groups were compared at termination, no signifi cant differences ( p > 0.05) were
found for the Hamilton Anxiety Rating Scale (HAM-A; d = −0.48), Beck Anxiety Inventory (BAI;
d = −0.33), Hospital Anxiety Scale (HAS; d = −0.57), or the Inventory of Interpersonal Problems
(IIP; d = −0.08) despite the tendency for CBT to have larger effect sizes. Similarly, at 6-month fol-
low-up, CBT continued to have larger effect sizes, but no signifi cant differences ( p > 0.05) were
found for the HAM-A ( d = −0.65), BAI ( d = −0.37), HAS ( d = −0.63), or IIP ( d = −0.07). However,
CBT was signifi cantly ( p < 0.01) superior at termination in reducing symptoms as measured by the
Penn State Worry Questionnaire ( d = −0.98), State-Trait Anxiety Inventory ( d = −0.82), and Beck
Depression Inventory ( d = −0.76). At 6-month follow-up, this signifi cant ( p < 0.05) superiority
remained on all three measures (respective effect sizes: d = −1.04, d = −0.93, d = −0.59). If the
between-group effect sizes of all measures used in this study are averaged, medium effect sizes are
found such that CBT is superior to psychodynamic therapy at both termination ( d = −0.57) and
6-month follow-up ( d = −0.61).
Thus, both treatments were associated with signifi cant improvements in anxiety, depression, and
interpersonal functioning, and no signifi cant differences were found between the two therapies on
the majority of measures. However, CBT was signifi cantly superior in measures of trait anxiety,
worrying, and depression. In discussing these fi ndings, Leichsenring et al. [ 50 ] note that the reduc-
tion of worry is a core element of the CBT treatment used in this trial but not of the applied STDP.
Thus, they suggest that this specifi c difference may account for the superiority of CBT on the Penn

128 J. Slavin-Mulford and M.J. Hilsenroth
State Worry Questionnaire and in part, on the State-Trait Anxiety Inventory. (This measure also
contains several items about worry). One way to look at this would be to accept that the two treat-
ments focus differentially on symptoms. Thus, some types of outcome measures will favor one treat-
ment while others will present the inverse pattern.
Although no outcome measures were associated with more positive change for psychodynamic
therapy in this study, other studies have suggested that psychodynamic therapy may be more effec-
tive in some domains [
39, 56 ] . For example, Gibbons and colleagues used a pooled study database
of fi ve trials utilizing similar methods with varied study populations in order to examine the unique
and common mechanisms of change across psychodynamic and cognitive psychotherapies. They
found that psychodynamic therapy led to signifi cantly ( p < 0.05) greater gains in self-understanding
of interpersonal patterns than cognitive therapy between intake and termination ( d = 0.47).
Importantly, improvements made during treatment in self-understanding were signifi cantly predic-
tive ( p < 0.05) of improvement in symptoms of anxiety on the BAI from termination to follow-up,
controlling for change on the BAI from intake to termination. This suggests that changes in self-
understanding preceded symptom change in anxiety. Thus, it seems that psychodynamic therapy’s
focus on self-understanding of interpersonal patterns may be an important mechanism of change in
the treatment of anxiety symptoms.
Other promising results for psychodynamic therapy in the reduction of anxiety come from
Milrod, Leon, Busch and colleagues’ [
53 ] RCT comparing panic focused psychodynamic psycho-
therapy (PFPP) and relaxation training in the treatment of 49 patients with panic disorder. Results
found that, at termination, PFPP had been signifi cantly more effective than relaxation training in
reducing the severity of a broad range of symptoms on the Panic Disorder Severity Scale ( p < 0.01,
d = 0.95). At termination, PFPP had also been more successful in lowering functional impairment as
measured by the Sheehan Disability Scale ( p = 0.01, d = 0.74), and it trended toward being more
effective in the reduction of depressive symptoms on the Hamilton Depression Scale ( p = 0.07,
d = 0.53). Finally, relaxation training had a signifi cantly higher attrition rate than PFPP: 7% of
patients in the PFPP group versus 34% of patients in the relaxation training group dropped out of
treatment ( p = 0.03).
Follow-up analyses for this RCT were reported by Milrod, Leon, Barber, Markowitz, and Graf [ 8 ]
in order to determine whether Axis II comorbidity moderated the treatment effects found for PFPP
and relaxation training. Results showed that on the Panic Disorder Severity Scale (PDSS), PFPP was
superior to relaxation therapy both for patients with ( d = 1.19) and without ( d = 0.55) an Axis II dis-
order and that the effect size was even larger for the personality disorder patients. In addition, when
only Cluster C was considered, PFPP outperformed relaxation therapy for patients with ( d = 1.35)
and without ( d = 0.69) a Cluster C personality disorder and again the effect size was even larger for
the Cluster C group. Because only fi ve subjects with an Axis II diagnosis did not have a Cluster C
personality disorder, the effects of other Axis II disorders could not be separated from the effects of
Cluster C. Thus, the conclusion which can be drawn is that Cluster C comorbidity increased the
effi cacy of PFPP while it decreased the effi cacy of relaxation therapy. These results are consistent
with the APA Practice Guidelines for Panic Disorder [ 16 ] , which recommend psychodynamic psy-
chotherapy for PD patients with comorbid personality disorders.
The research presented thus far has compared psychodynamic treatments to cognitive and behav-
ioral treatments as well as to controls and pharmacological interventions. Results have generally
suggested that psychodynamic therapy demonstrates effi cacy in relation to controls and that it is
generally as effi cacious as medication. However, the conclusions with regards to pharmacological
interventions are preliminary due to the small number of studies in this area. Finally, the effi cacy of
psychodynamic therapy in relation to cognitive, behavioral, and CBT is more mixed. While psycho-
dynamic therapy produces large treatment effects, there is a general trend for CBT to demonstrate
small to moderate effects over psychodynamic treatments for anxiety disorders.

1297 Evidence-Based Psychodynamic Treatments for Anxiety Disorders: A Review
Another way to examine the effects of psychodynamic treatment of anxiety disorders is to evaluate
pre-post within-group treatment changes. This allows an evaluation of the amount of improvements
patients make in psychodynamic therapy between intake and termination and between intake and
follow-up. This allows a direct evaluation of the amount of change patients make during psychody-
namic therapy. Thus, this chapter will now move to discussing the effectiveness of psychodynamic
therapy in leading to change.
Pre-Post Within-Group Effectiveness of Psychodynamic
Therapy in the Treatment of Anxiety
Table 7.4 summarizes the within-group (pre-post change) effect sizes for the primary patient- and
clinician-rated outcome measures of anxiety (overall symptoms or improvement when a measure of
anxiety was not provided). This table only includes published studies that use psychodynamic ther-
apy in the treatment of anxiety disorders. When the pre-post treatment effect sizes of these studies
are averaged (random effect, weighted for sample size), psychodynamic psychotherapy has a large
mean effect for patient ( d = 1.05; range d = 0.26–3.23) and clinician ( d = 1.62; range d = 0.57–2.29)
ratings on the primary outcome scales for the studies ( N = 269). These summary scores are not meant
to constitute a meta-analysis or represent a complete reporting of all outcome effects for every vari-
able examined in the studies reported. Rather, they are provided to organize the magnitude of effects
for the primary anxiety variables in a clear manner. However, whereas this summary score is not a
meta-analysis, it is important to note that the summary effect sizes are comparable to Stewart and
Chambless’s [ 58 ] meta-analytic fi nding that effect sizes for CBT range from 0.83 to 2.59 depending
on the specifi c anxiety disorder being targeted. For example, they found that across 11 effectiveness
studies of CBT for GAD, the mean pre-post effect size was 0.92 for generalized anxiety. Across 17
effectiveness studies of CBT for panic disorder, the mean pre-post effect size for panic attacks was
1.01. And, across 11 effectiveness studies of social anxiety disorder, the mean pre-post effect size for
social anxiety symptoms was 1.04.
5
In addition to examining the summary of within-group effect sizes across numerous studies, it is
also useful to look directly at effectiveness studies of psychodynamic psychotherapy for anxiety
disorders in naturalistic settings. Reviewing both RCT’s and effectiveness research in naturalistic
settings is important as they provide different types of information. Effi cacy studies tend to have
high internal validity, enabling them to provide information about which treatment leads to the best
results under controlled conditions. In contrast, effectiveness research has higher external validity
and helps to identify if a treatment is feasible and effective in real-world settings [ 61 ] . In order to
examine whether psychodynamic treatments of anxiety are likely to be effective as well as effi ca-
cious, this chapter will now turn to research examining the effectiveness of psychodynamic treat-
ments for anxiety in naturalistic settings.
Three studies have examined the effectiveness of psychodynamic psychotherapy for anxiety dis-
orders in naturalistic settings [
21, 27, 57 ] . Crits-Christoph et al.’s [ 27 ] study was an open trial of
brief supportive-expressive (SE) psychotherapy in the treatment of GAD; Milrod et al.’s [ 21 ] study
was an open trial of panic-focused psychodynamic psychotherapy (PFPP) in the treatment of panic
disorder; and Slavin-Mulford et al.’s [ 57 ] study was an open trial of short-term psychodynamic psy-
chotherapy in the treatment of anxiety spectrum disorders. All three studies suggest that psychody-
namic therapy is effective in treating anxiety.
5 The measure of effect size in this study was Hedge’s g [ 59 ] rather than Cohen’s d [ 60 ] which is more commonly
reported in meta-analyses. The two measures are based on slightly different computational formulas, but with large
samples, the choice of formula leads to limited or no change in transforming g to d .

130 J. Slavin-Mulford and M.J. Hilsenroth
Crits-Christoph et al.’s [ 27 ] study included 26 adult outpatients with GAD. Exclusionary criteria
included acute medical disorders, any current or past history of schizophrenic disorders, bipolar
disorder, Cluster A Axis II disorders, and anyone who met criteria within the past year for substance
Table 7.4 Psychodynamic treatments for anxiety disorders pre–post follow-up
Citation N , disorder, and type of dynamic treatment Pre–Post ES Pre-follow-up ES
Ferrero et al. [
45 ] 34 GAD post and 3 month
33 GAD 9 month
Brief Adlerian psychotherapy (10–15 sessions)
HAM-A
d = 0.57
HAM-A
3 month d = 1.58
9 month d = 1.53
Wiborg and
Dahl [ 46 ]
20 panic
Manualized psychodynamic treatment
based on Davanloo, Malan, Strupp and Binder
(15 sessions) PLUS clomipramine
PAAS d = 3.23
HAM-A d = 2.29
9 month
PAAS d = 2.54
HAM-A d = 1.88
Brom et al. [
39 ] 29 PTSD
Manualized brief (mean = 18.8 sessions)
psychodynamic focused on solving intrapsychic
confl ict resulting from trauma
Gen Symp
d = 0.92
Gen Symp
3 month
d = 1.28
Durham
et al. [
48 ]
11 GAD post
12 GAD 6 month
10 GAD 12 month
AP (analytic-based therapy w/exploration of
symptoms, current relationships, development,
transference, and resistance-no support used)
High contact condition: 16–20 sessions
STAI-T
d = 0.26
STAI-T
6 month d = 0.39
12 month d = 0.23
Durham
et al. [
48 ]
13 GAD post
14 GAD 6 month
12 GAD 12 month
AP: Low contact: 8–10 sessions
STAI-T
d = 0.49
STAI-T
6 month d = 0.12
12 month d = 0.13
Klein
et al. [
49 ] a
18 Phobic PTODI
17 Phobic ERODI
Supportive dynamic (26 weekly sessions) PLUS
imipramine
89% Pt-ODI
76% IR-ODI
Leichsenring
et al. [
50 ]
28 GAD
Manualized STDP based on Luborsky
and Crits-Christoph (up to 30 sessions)
STAI-T d = 1.02
HAM-A d = 2.14
6 month
STAI-T d = 0.94
HAM-A d = 2.02
Milrod et al. [
8 ] 26 Panic
Panic-focused psychodynamic psychotherapy
(twice weekly for 12 weeks)
PDSS d = 2.07
Pierloot and
Vinck [ 51 ]
9 Elevated scores on Taylor Manifest Anxiety Scale
Short-term psychodynamic tx infl uenced by
Malan (20 sessions)
STAI-T d = 0.52 STAI-T 3 months
d = 0.54
Crits-Christoph
et al. [
40 ]
61 GAD
Brief supportive-expressive psychotherapy
(16 weekly sessions + 3 monthly booster sessions)
BAI d = 1.25
HAM-A d = 1.18
Milrod
et al. [
21 ]
21 Panic 16 weeks
17 panic 6 month
PFPP (2× weekly 24 sessions)
ASI d = 1.19
PDSS d = 2.08
6 month
ASI d = 1.66
PDSS d = 1.81
Slavin-Mulford
et al. [
57 ]
12 GAD, 1 Panic, 4 PTSD, and 4 Anxiety Disorder NOS
STPP (duration varied; mean sessions = 29)
BSI Anx d = 0.89
GAF d = 1.44
HAM-A Hamilton Anxiety Inventory (independent rater), PAAS Panic Attack and Anxiety Scale (patient rated), Gen
Symp total general symptoms (patient rated), STAI-T State-Trait Anxiety Inventory (patient rated), Pt-ODI patient
rating of overall degree of improvement, IR-ODI independent rating of overall degree of improvement, PDSS Panic
Disorder Severity Scale (independent rater), BAI Beck Anxiety Inventory (patient rated), ASI Anxiety Sensitivity
Inventory (patient rated), STPP short-term psychodynamic psychotherapy, BSI Anx Brief Symptom Inventory anxiety
subscale (patient rated), GAF Global Assessment of Functioning Scale (independent rater)
a Klein et al. [ 49 ] did not provide information for calculating within-group effect sizes. However, the percentage of
patient improvement as reported by the patients and independent raters was provided

1317 Evidence-Based Psychodynamic Treatments for Anxiety Disorders: A Review
dependence or abuse, OCD, eating disorder, major depression, or borderline personality disorder.
These patients received manualized SE from therapists who had more than 10 years post doctoral
experience and special training in SE for GAD. Treatment consisted of 16 weekly sessions and 3
monthly booster sessions. It was conducted using Luborsky’s [
38 ] general SE manual in conjunction
with Crits-Christoph et al.’s [ 62 ] GAD specifi c SE manual. This approach focuses on understanding
anxiety in the context of interpersonal/intrapsychic confl icts using the Core Confl ictual Relational
Theme which focuses on patients’ cyclical relational patterns.
Results showed that, at the end of 16 weeks, 79% of patients no longer qualifi ed for GAD. This
considerable reduction in diagnosis occurred alongside a signifi cant and large decrease in anxiety
symptoms as measured by the HAM-A ( p < 0.01, d = 1.41) and BAI ( p < 0.01, d = 1.99). Moreover,
signifi cant improvements were found for worry as measured by the Penn State Worry Questionnaire
( p < 0.01, d = 0.95), interpersonal functioning as measured by the Inventory of Interpersonal Problems
( p < 0.05, d = 0.25), and depression as measured by the Hamilton Depression Rating Scale ( p < 0.01,
d = 1.15) and Beck Depression Inventory ( p < 0.01, d = 1.09).
In a later study [ 62 ] , additional SE patients were added to the 1996 sample, and it is this larger
sample that is reported in Table 7.4 . These additional SE patients were obtained from the following:
13 patients originally excluded due to comorbid MDD, seven patients originally excluded because
they met all the criteria of DSM-IV GAD except that their worry was only in one sphere, and 15
patients who had been randomized to SE in a pilot study comparing SE and supportive, nondirect
therapy. With this combined sample, Crits-Christoph et al. [ 62 ] found that SE therapy led to statisti-
cally and clinically meaningful change on the HAM-A ( p < 0.01, d = 1.18), BAI ( p < 0.01, d = 1.25)
and Inventory of Interpersonal Problems (IIP; p < 0.01, d = 1.07). These numbers are somewhat lower
than were reported in the 1996 paper. Because the 2005 paper did not examine these differences, it
is unclear as to what caused the drop. One possibility is that the comorbid MDD patients responded
less favorably than the noncomorbid patients to SE. However, this possibility remains the work of
future research.
At this time, the most important point is that with this larger sample, SE remained highly effective.
In fact, these effect sizes are similar in magnitude to those found for CBT treatments. Specifi cally,
this investigation demonstrated effect sizes of 1.18 and 1.25 for reducing anxiety symptoms which
is comparable to the Stewart and Chambless [ 58 ] fi nding that across 11 effectiveness studies of CBT
for GAD, the mean pre-post effect size was 0.92 for generalized anxiety.
6 These fi ndings suggest that
psychodynamic therapy is effective in treating anxiety disorders in the short term. However, because
no follow-up data was collected, it is unclear as to whether these gains are maintained.
Milrod et al.’s [ 21, 26 ] study helps to begin addressing the issues of how well changes are main-
tained after treatment. Because their 2001 work is an expanded report of their 2000 study and pro-
vides additional assessment measures and increased sample, only the results presented in the 2001
report will be discussed. In this study, 21 patients with panic disorder were recruited, 81% of whom
had at least one comorbid Axis I diagnosis. Unlike Crits-Christoph et al.’s [ 27 ] study, major depres-
sion was not used as an exclusionary category and 24% of patients had comorbid major depression.
Other Axis I comorbidities included: dysthymia (24%), GAD (24%), specifi c phobia (24%), and
social phobia (14%).
Patients were treated with 24 sessions at twice-weekly intervals of Panic-Focused Psychodynamic
Psychotherapy (PFPP). PFPP is a manualized treatment that focuses on the underlying emotional
meanings of panic symptoms and on current social and emotional functioning through utilization of
free association, exploration of fantasies, interpretation of defenses, and the therapeutic relationship.
Results showed signifi cant and substantial improvements at termination on the Panic Disorder
Severity Scale ( d = 2.08, p < 0.001), Hamilton Anxiety Scale ( d = 1.72, p < 0.001), Hamilton
6 The measure of effect size in this study was Hedge’s g [ 59 ] rather than Cohen’s d [ 60 ] which is more commonly
reported in meta-analyses. The two measures are based on slightly different computational formulas, but with large
samples, the choice of formula leads to limited or no change in transforming g to d .

132 J. Slavin-Mulford and M.J. Hilsenroth
Depression Scale ( d = 0.89, p < 0.002), and Sheehan Disability Scale ( d = 1.55, p < 0.001). In addition,
signifi cant and substantial improvements were found at termination on the Body Sensations
Questionnaire ( d = 1.30, p < 0.001), Agoraphobic Cognitions Questionnaire ( d = 1.27, p = 0.001),
Marks and Mathews Fear Questionnaire 1 ( d = 1.12, p = 0.002), Marks and Mathews Fear Questionnaire
2 ( d = 1.39, p < 0.001), and Anxiety Severity Index ( d = 1.19, p < 0.001). Moreover, these gains were
maintained on all measures at 6-month-follow-up. In fact, when termination scores were compared
to follow-up scores, there was a trend on the Anxiety Sensitivity Index ( d = 0.31, p = 0.08) and on the
Marks and Mathews Fear Questionnaire 1 ( d = 0.42, p = 0.10) for the gains to increase. Thus, it seems
that at follow-up, additional small effect size gains were made even over the treatment termination
scores, although these levels of chance did not reach traditional levels of signifi cance.
When Milrod et al.’s [
21 ] within-group effect sizes are compared to the within-group effect sizes
obtained in CBT for panic disorder, PFPP seems to do as well if not better than CBT. Specifi cally,
the effect sizes found on the Panic Disorder Severity Scale at termination in Milrod et al.’s [ 21 ] study
was 2.08. Comparatively, Stewart and Chambless [ 58 ] report a mean pre-post effect size of 1.02 for
general anxiety symptoms, 1.01 for panic attacks, and 0.83 for avoidance in their meta-analytic
review of 17 effectiveness studies that used cognitive, behavioral, or a combination of cognitive and
behavioral interventions in the treatment of panic disorder.
7 Thus, like Crits-Christoph et al.’s [ 27 ]
results, this suggests that psychodynamic therapy is highly effective in the treatment of anxiety dis-
orders and that these large within-group changes are consistent with the within-group changes found
in CBT.
Although Crits-Christoph et al. [ 27 ] and Milrod et al. [ 21 ] provide important information about
the outcome of psychodynamic therapy, they do not explore how the process of therapy (e.g., tech-
niques) is related to change. Slavin-Mulford et al.’s [ 57 ] study helps to begin addressing this gap by
examining which therapeutic techniques were most related to outcome in an open trial of short-term
psychodynamic therapy for anxiety disorders. In this naturalistic study, 21 patients with a diagnosed
anxiety disorder (12 GAD, one Panic, four PTSD, and four Anxiety Disorder NOS) were accepted
into treatment regardless of comorbidity. The majority of patients had at least one comorbid Axis
I diagnosis (57% mood disorder; 10% eating disorder; 5% adjustment disorder) as well as a comor-
bid personality disorder (67%).
These patients received once or twice weekly 50–60-minute sessions of short-term psychody-
namic psychotherapy (STPP). The decision about frequency of sessions was decided collaboratively
between patient and therapist through a consideration of the patient’s needs.
8 Treatment was orga-
nized, aided, and informed (but not prescribed) by the technical guidelines delineated in four treat-
ment manuals [ 38, 63– 65 ] . Additional technical material specifi c to the STPP treatment of anxiety
[ 62 ] was actively integrated into the treatment of these patients. Key features of the STPP model
include [ 66 ] : [ 1 ] Focus on affect and the expression of emotion; [ 2 ] The identifi cation of patterns in
actions, thoughts, feelings, experiences, and relationships with these patterns being explored/formu-
lated using the Core Confl ictual Relationship Theme (CCRT) format [ 67 ] ; [ 3 ] Emphasis on past
experiences; [ 4 ] Focus on interpersonal experiences; [ 5 ] Emphasis on the therapeutic relationship/
alliance; [ 6 ] Exploration of wishes, dreams, or fantasies; and [ 7 ] Exploration of attempts to avoid
topics or engage in activities that may hinder the progress of therapy.
Results showed statistically signifi cant changes at the end of treatment on anxiety symptoms,
global distress, interpersonal distress, and social/occupational functioning.
7 The measure of effect size in this study was Hedge’s g [ 59 ] rather than Cohen’s d [ 60 ] which is more commonly
reported in meta-analyses. The two measures are based on slightly different computational formulas, but with large
samples, the choice of formula leads to limited or no change in transforming g to d .
8 The mean number of sessions attended by these twenty-one patients was 29 (SD = 15), and the median was 24. The
maximum number of sessions attended by a patient was 64.

1337 Evidence-Based Psychodynamic Treatments for Anxiety Disorders: A Review
Specifi cally, there was a signifi cant and large decrease in anxiety symptoms as assessed by patient
self-report on the Brief Symptom Inventory Anxiety Scale ( p < 0.01, d = 0.89). Likewise, global
symptomatic distress as measured by the clinician-rated Global Assessment of Functioning scale
(GAF, d = 1.44) and patient-rated Global Severity Index (GSI, d = 0.92) were both shown to signifi -
cantly decrease over the course of treatment ( p < 0.01) with large effects. In addition, signifi cant
( p < 0.05) improvement in interpersonal functioning was also found with changes ranging from small
(patient-rated Interpersonal Sensitivity subscale of the Brief Symptom Inventory, d = 0.33) to large
(externally rated Global Assessment of Relational Functioning, d = 1.23). Moreover, the patient-
rated Social Adjustment Scale ( d = 0.53) and clinician-rated Social and Occupational Functioning
Assessment Scale ( d = 0.84) also showed signifi cant changes ( p < 0.05) with medium to large effects.
Finally, at termination, most patients (76%) reported that their anxiety symptoms were within two
standard deviations of the normative mean. This is similar to or even more encouraging than over
psychodynamic (e.g. [
40, 45, 48 ] ) and CBT [ 18, 68, 69 ] studies on anxiety disorder patients.
Importantly, Slavin-Mulford and colleagues’ [ 57 ] study also examined the relationship between
the use of specifi c therapeutic techniques and subsequent change. A signifi cant direct process-out-
come link between psychodynamic-interpersonal therapist techniques and changes in anxiety symp-
toms was observed ( r = 0.46, p = 0.04). Further, results showed that several individual PI techniques
were meaningfully related to outcome. These included: [ 1 ] focusing on wishes, fantasies, dreams, and
early memories; [ 2 ] linking current feelings or perceptions to the past; [ 3 ] highlighting patients’ typi-
cal patterns; and [ 4 ] helping patients to understand their experiences in new ways. This compilation
of techniques fi ts well with psychodynamic theories for anxiety as well as the empirically supported
treatments derived from them [ 21, 27 ] . Thus, Slavin-Mulford and colleagues’ [ 57 ] work along with
the other research presented previously suggests that psychodynamic treatments and techniques are
likely helpful in the treatment of anxiety disorders. This being said, many questions remain such as
which patient (e.g., Axis I and II comorbidity, alliance) and therapist (e.g., experience, adherence,
competence, alliance) variables moderate the outcome of psychodynamic therapy in the treatment of
anxiety disorders. Thus, it will be the job of future research to begin fi lling in these gaps. This chapter
will now summarize the research presented so far and offer a road map for the research still needed.
Summary and Future Directions
Psychodynamic therapy is widely used to treat anxiety [ 12 ] , and many studies have begun examining
its effi cacy and effectiveness with promising results. The few effectiveness studies which have been
conducted in naturalistic settings indicate that psychodynamic treatments for anxiety demonstrate
large effects [ 21, 27, 57 ] . Patients in these studies treated with psychodynamic therapy evidenced
considerable reduction in diagnosis, anxiety symptoms, depression, and global distress (i.e., large
effects). In addition, randomized control trials suggest that psychodynamic treatment for anxiety
symptoms tends to be more effi cacious than controls [ 43 ] . Moreover, the few studies which have
compared psychodynamic therapy to medication, suggest that psychodynamic therapy is as effi ca-
cious as pharmacological interventions (e.g. [ 45, 46 ] ). Finally, the effi cacy of psychodynamic ther-
apy in relation to cognitive, behavioral, and CBT is more mixed, although there is a general trend for
CBT to demonstrate small to moderate effects over psychodynamic treatments for anxiety disorders
(e.g. [ 49, 50 ] ). Importantly, however, many of the studies comparing CBT to psychodynamic therapy
found large effects for both treatments (e.g. [ 50 ] ).
Despite these encouraging fi ndings, much work remains. For example, there are no diagnosis-
specifi c controlled trials of psychodynamic therapy for social phobias or specifi c phobias [ 1 ] and
there has only been one randomized control trial which has included a psychodynamic treatment
for PTSD [ 39 ] . The need for more psychodynamic research on GAD [ 62 ] and panic disorder [ 21 ]

134 J. Slavin-Mulford and M.J. Hilsenroth
has also been called for in recent research. Moreover, given that anxiety disorders are frequently
comorbid with one another [ 3 ] , more studies examining the anxiety spectrum as a whole would
help to represent the types of patients who actually present for treatment in clinical practice. Finally,
it will be important for future research to examine the moderating effects of Axis I and Axis II
comorbidity, as well as therapeutic alliance and aspects of technique.
Concerns with regards to the moderating effects of Axis I comorbidity were raised by Crits-
Christoph et al.’s [
27, 40 ] fi nding that the pre-post effect sizes declined when patients with comorbid
MDD were included. However, other types of patients were added at this time making it unclear as
to whether the comorbid MDD was involved in this decrease in effect size. Unfortunately, other
research has generally failed to clarify the effects of comorbidity as many studies have utilized strin-
gent exclusionary criteria such as eliminating patients with comorbid major depression (e.g. [ 50 ] ).
The potential moderating effects of Axis II comorbidity have also been raised by a few studies
which suggest that psychodynamic therapy for anxiety disorders may be even more effective than
CBT or pharmacological approaches when patients have a comorbid personality disorder [
8, 45 ] .
Specifi cally, Ferrero et al. [ 45 ] found that psychodynamic therapy was as successful as medication
among GAD patients when Axis II comorbidity was controlled, but that it was even more successful
than medication in helping GAD patients with a comorbid PD improve their social and occupational
functioning at 6-month follow-up. In addition, Milrod et al. [ 8 ] found that Cluster C comorbidity
increased the effi cacy of PFPP whereas it decreased the effi cacy of relaxation therapy. These fi nd-
ings suggest that future research should examine which Axis I and II disorders moderate different
anxiety disorders and the direction of these effects.
Future research should also continue to examine how the process of psychodynamic therapy for
anxiety disorders is related to outcome. Slavin-Mulford et al.’s [ 57 ] work has begun this process by
suggesting that specifi c psychodynamic/interpersonal techniques are related to reduction in anxiety
in short-term psychodynamic psychotherapy. However, much work in this area remains. For exam-
ple, it will be important to examine which intervention strategies work best with which types of anxi-
ety disorder patients (i.e., comparison between patients with and without a personality disorder as
well as comparisons between different disorders such as GAD versus panic). If similar process
research was also conducted for CBT and other treatments of anxiety disorders, it may be possible
to determine which process variables are critical for sustained change in patients with different anxi-
ety disorders. This type of process and outcome research could also help to explain which treatment
factors in psychodynamic therapy lead it to have more gradual but also longer-term progress than
other forms of treatment with some patients (e.g. [ 39, 51 ] ). In addition, process and outcome research
could provide important information about which specifi c factors in psychodynamic therapy and
CBT lead them to help patients differentially, depending on the symptoms being targeted [ 39, 50 ] .
Thus, while research generally suggests that psychodynamic therapy is useful in the treatment of
anxiety disorders, there is still much work to be done.
References
1. Fonagy P, Roth A, Higgitt A. Psychodynamic psychotherapies: evidence-based practice and clinical wisdom. Bull
Menninger Clin. 2005;69(1):1–58.
2. Somers JM, Goldner EM, Waraich P, Hsu L. Prevalence and incidence studies of anxiety disorders: a systematic
review of the literature. Can J Psychiatry. 2006;51(2):100–13.
3. Andrews G, Slade T, Issakidis C. Deconstructing current comorbidity: data from the Australian national survey of
mental health and well-being. Br J Psychiatry. 2002;181:306–14.
4. Moffi tt TE, Harrington HL, Caspi A, Kim-Cohen J, Goldberg D, Gregory AM, et al. Depression and generalized
anxiety disorder: cumulative and sequential comorbidity in a birth cohort followed prospectively to age 32 years.
Arch Gen Psychiatry. 2007;64:651–60.

1357 Evidence-Based Psychodynamic Treatments for Anxiety Disorders: A Review
5. Bruce SE, Yonkers KA, Otto MW, Eisen JL, Weisberg RB, Pagano M, et al. Infl uence of psychiatric comorbidity
on recovery and recurrence in generalized anxiety disorder, social phobia, and panic disorder: a 12-year prospec-
tive study. Am J Psychiatry. 2005;162:1179–87.
6. Joormann J, Kosfelder J, Schulte D. The impact of comorbidity on the course of anxiety treatments. Cognitive
Ther Res. 2005;29(5):569–91.
7. Kessler RC, Mickelson KD, Barber C, Wang P. The association between chronic medical conditions and work
impairment. In: Rossi AS, editor. Caring and doing for others: social responsibility in the domain of the family,
work, and community. Chicago: University of Chicago Press; 2001. p. 403–26.
8. Milrod B, Leon AC, Barber JP, Markowitz JC, Graf E. Do comorbid personality disorders moderate panic-focused
psychotherapy? An exploratory examination of the American Psychiatric Association practice guideline. J Clin
Psychiatry. 2007;68:885–91.
9. Siev J, Chambless DL. Specifi city of treatment effects: cognitive therapy and relaxation for generalized anxiety
and panic disorders. J Consult Clin Psychol. 2007;75(4):513–22.
10. Klerman GL, Weissman MM, Ouellette R, Johnson J, Greenwald S. Panic attacks in the community social mor-
bidity and health care utilization. J Am Med Assoc. 1991;265:742–6.
11. Swinson RP, Cox B, Woszczyna CB. Use of medical services and treatment for panic disorder with agoraphobia
for social phobia. Can Med Assoc J. 1991;147:878–83.
12. Goisman RM, Warshaw MG, Keller MB. Psychosocial treatment prescriptions for generalized anxiety disorder,
panic disorder, and social phobia, 1991–1996. Am J Psychiatry. 1999;156(11):1819–21.
13. Borkovec TD, Ruscio AM. Psychotherapy for generalized anxiety disorder. J Clin Psychiatry. 2001;62:37–42.
14. Starcevic V. Issues in the pharmacological treatment of anxiety disorders. Australas Psychiatry.
2005;13(4):371–4.
15. Woods JH, Katz JL, Winger G. Benzodiazepines: use, abuse, and consequences. Pharmacol Rev. 1992;44:151–347.
16. American Psychiatric Association (APA). Practice guideline for the treatment of patients with panic disorder. 2nd
ed. Washington, DC: American Psychiatric Association; 2009.
17. Mitte K. Meta-analysis of cognitive-behavioral treatments for generalized anxiety disorder: a comparison with
pharmacotherapy. Psychol Bull. 2005;131(5):785–95.
18. Barlow DH, Gorman JM, Shear MK, Woods SW. Cognitive-behavioral therapy, imipramine, or their combination
for panic disorder. J Am Med Assoc. 2000;283:2529–36.
19. Knijnik DZ, Kapczinski F, Chachamovich E, Margis R, Eizirik CL. Psychodynamic group treatment for general-
ized social phobia. Rev Bras Psiquiatr. 2004;26(2):77–81.
20. Schottenbauer MA, Glass CR, Arnkoff DB, Tendick V, Gray SH. Nonresponse and dropout rates in outcome stud-
ies on PTSD: review and methodological considerations. Psychiatry: Interpers Biol Process. 2008;71(2):134–68.
21. Milrod B, Busch F, Leon AC, Aronson A, Roiph J, Rudden M, et al. A pilot trial of brief psychodynamic psycho-
therapy for panic disorder. J Psychother Pract Res. 2001;10:239–45.
22. Zerbe K. Through the storm: psychoanalytic theory in the psychotherapy of the anxiety disorders. Bull Menninger
Clin. 1990;54(2):171–83.
23. Freud S. Inhibitions, symptoms, and anxiety. The Standard Edition of the Complete Psychological Works of
Sigmund Freud. 1926;75–176.
24. Compton A. A study of the psychoanalytic theory of anxiety: I. The development of Freud’s theory of anxiety.
J Am Psychoanal Assoc. 1972;20:3–44.
25. Busch F, Milrod BL, Singer MB. Theory and technique in psychodynamic treatment of panic disorder. J Psychother
Pract Res. 1999;8(3):234–42.
26. Milrod B, Busch F, Leon AC, Shapiro T, Aronson A, Roiphe J, et al. Open trial of psychodynamic psychotherapy
for panic disorder: a pilot study. Am J Psychiatry. 2000;157(11):1878–80.
27. Crits-Christoph P, Connolly MB, Azarian K, Crits-Christoph K, Shappell S. An open trial of brief supportive-
expressive psychotherapy in the treatment of generalized anxiety disorder. Psychother: Theor Res Pract Training.
1996;33(3):418–30.
28. Sullivan HS. The interpersonal theory of psychiatry. New York: Norton; 1953.
29. Fairbairn WD. An object-relations theory of personality. New York: Basic Books; 1952.
30. Fromm-Reichmann F. Psychiatric aspects of anxiety. In: Thompson CM, Mazer M, Witenberg E, editors. An
outline of psychoanalysis. New York: Modern Library; 1955. p. 113–33.
31. Klein M. Envy and gratitude and other works 1946–1963. New York: Delacorte P/Seymour Lawrence; 1975.
32. Klein M. Love, guilt, and reparation and other works 1921–1945. New York: Delacorte P/Seymore Lawrence;
1975.
33. Bowlby J. A secure base: clinical applications of attachment theory. London: Routledge; 1988.
34. Borkovec TD, Ray WJ, Stobger J. Worry: a cognitive phenomenon intimately linked to affective, physiological,
and interpersonal behavioral processes. Cognitive Ther Res. 1998;22:561–78.
35. Breitholtz E, Johansson B, Ost LG. Cognitions in generalized anxiety disorder and panic disorder patients: a
prospective approach. Behavior Res Ther. 1995;37:533–44.

136 J. Slavin-Mulford and M.J. Hilsenroth
36. Vertue FM. From adaptive emotion to dysfunction: pathways to social anxiety disorder. Rev Pers Soc Psychol.
2003;7(2):170–91.
37. Silove D. Perceived parental characteristics and reports of early parental deprivation in agoraphobic patients. Aust
N Z J Psychiatry. 1986;20:365–9.
38. Luborsky L. Principles of psychoanalytic psychotherapy: a manual for supportive–expressive treatment. New
York: Basic Books; 1984.
39. Brom D, Kleber RJ, Defares PB. Brief psychotherapy for posttraumatic stress disorders. J Consult Clin Psychol.
1989;57(5):607–12.
40. Crits-Christoph P, Connolly Gibbons MB, Narducci J, Schamberger M, Gallop R. Interpersonal problems and the
outcome of interpersonally oriented psychodynamic treatment of GAD. Psychother: Theor Res Pract Training.
2005;42:211–24.
41. Borkovec TD, Mathews AM. Treatment of non-phobic anxiety disorders: a comparison of nondirective, cognitive,
and coping desensitization therapy. J Consult Clin Psychol. 1998;56:877–84.
42. Crits-Christoph P. Personal Communication. August 11, 2009.
43. Abbass AA, Hancock JT, Henderson J, Kisely S. Short-term psychodynamic psychotherapies for common mental
disorders. Cochrane Database of Systematic Reviews(4). doi:10.1002/14651858.CD004687.pub3 Art. No.:
CD004687; 2006.
44. Sjodin I, Svedlund J, Ottosson JO, Dotevall G. Controlled study of psychotherapy in chronic peptic ulcer disease.
Psychosomatics. 1986;27(3):187–200.
45. Ferrero A, Piero A, Fassina S, Massola T, Lanteri A, Abbate Daga G, et al. A 12-month comparison of brief psy-
chodynamic psychotherapy and pharmacotherapy treatment in subjects with generalized anxiety disorders in a
community setting. Eur Psychiatry. 2007;22:530–9.
46. Wiborg IM, Dahl AA. Does brief dynamic psychotherapy reduce the relapse rate of panic disorder. Arch Gen
Psychiatry. 1996;53:689–94.
47. Durham RC, Murphy T, Allan T, Richard K, Treliving LR, Fenton GW. Cognitive therapy, analytic psychotherapy,
and anxiety management training for generalized anxiety disorder. Br J Psychiatry. 1994;165:315–23.
48. Durham RC, Fisher PL, Treliving LR, Hau CM, Richard K, Steward JB. Psychotherapy and anxiety management
training for generalized anxiety disorder: symptom change, medication usage and attitudes to treatment. Behav
Cogn Psychother. 1999;27(1):19–35.
49. Klein DF, Zitrin CM, Woerner MG, Ross DC. Treatment of phobias. II. Behavior therapy and supportive psycho-
therapy: are there any specifi c ingredients. Arch Gen Psychiatry. 1983;40:139–45.
50. Leichsenring F, Salzer S, Jaeger U, Kachele H, Kreische R, Leweke F, et al. Short-term psychodynamic psycho-
therapy and cognitive-behavioral therapy in generalized anxiety disorder: a randomized, controlled trial. Am J
Psychiatry. 2009;166:875–81.
51. Pierloot R, Vinck J. Differential outcome of short-term dynamic psychotherapy and systematic desensitization in
the treatment of anxious out-patients a preliminary report. Psychol Belg. 1978;18(1):87–98.
52. Svartberg M, Stiles TC, Seltzer MH. Randomized, control trial of the effectiveness of short-term dynamic psycho-
therapy and cognitive therapy for cluster c personality disorders. Am J Psychiatry. 2004;161(5):810–7.
53. Milrod B, Leon AC, Busch F, Rudden M, Schwalberg M, Clarkin J, et al. A randomized controlled clinical trial of
psychoanalytic psychotherapy for panic disorder. Am J Psychiatry. 2007;164(2):265–72.
54. Leichsenring F, Rabung S, Leibing E. The effi cacy of short-term psychodynamic psychotherapy in specifi c psy-
chiatric disorders. Arch Gen Psychiatry. 2004;61:1208–16.
55. Durham RC, Chambers JA, MacDonald RR, Power KG, Major K. Does cognitive-behavioral therapy infl uence
the long-term outcome of generalized anxiety disorder? An 8–14 year follow-up of two clinical trials. Psychol
Med. 2003;33:499–509.
56. Gibbons MBC, Crits-Christoph P, Barber J, Wiltsey Stirman S, Gallop R, Goldstein L, et al. Unique and common
mechanisms of change across cognitive and dynamic psychotherapies. J Consult Clin Psychol. 2009;77:801–13.
57. Slavin-Mulford J, Hilsenroth M, Weinberger J, Gold J. Therapeutic interventions related to outcome in psychody-
namic psychotherapy for anxiety disorder patients. Journal of Nervous and Mental Disease. 2011;199:214–21.
58. Stewart RE, Chambless DL. Cognitive-behavioral therapy for adult anxiety disorders in clinical practice: a meta-
analysis of effectiveness studies. J Consult Clin Psychol. 2009;77:595–606.
59. Hedges LV. Estimation of effect size from a series of independent experiments. Psychol Bull. 1982;92:490–9.
60. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale: Erlbaum; 1988.
61. Westen D, Novotny CM, Thompson-Brenner H. The empirical status of empirically supported psychotherapies:
assumptions, fi ndings, and reporting in controlled clinical trails. Psychol Bull. 2004;130:631–63.
62. Crits-Christoph P, Crits-Christoph K, Wolf-Palacio D, Fichter M, Rudick D. Brief supportive-expressive psycho-
dynamic therapy for generalized anxiety disorder. In: Barber JP, Crits-Christoph P, editors. Dynamic therapies for
psychiatric disorders. New York: Basic Books; 1995. p. 43–83.
63. Book H. How to practice brief psychodynamic psychotherapy: the core confl ictual relationship theme method.
Washington, DC: American Psychological Association; 1988.

1377 Evidence-Based Psychodynamic Treatments for Anxiety Disorders: A Review
64. Strupp H, Binder J. Psychotherapy in a new key. New York: Basic Books; 1984.
65. Wachtel P. Therapeutic communication: principles and effective practice. New York: Guilford; 1983.
66. Blagys MD, Hilsenroth MJ. Distinctive features of short-term psychodynamic-interpersonal psychotherapy:
a review of the comparative psychotherapy process literature. Clin Psychol Sci Prac. 2000;7:167–88.
67. Luborsky L, Crits-Christoph P. Understanding transference: the core confl ictual relational theme method. 2nd ed.
Washington, DC: APA; 1997.
68. Borkovec TD, Costello E. Effi cacy of applied relaxation and cognitive-behavioral therapy in the treatment of
generalized anxiety disorder. J Consult Clin Psychol. 1993;61:611–9.
69. Borkovec TD, Newman MG, Pincus AL, Lytle R. A component analysis of cognitive–behavioral therapy for
generalized anxiety disorder and the role of interpersonal problems. J Consult Clin Psychol. 2002;70:288–98.