Journal of Anxiety Disorders, Vol. 14, No. 4, pp. 345–358, 2000Copyright 2000 Elsevier Science Ltd

Printed in the USA. All rights reserved0887-6185/00 $–see front matter

Pergamon

PII S0887-6185(00)00027-X

A Comparison of the Efficacyof Clonazepam and

Cognitive-BehavioralGroup Therapy for the

Treatment of Social Phobia

Michael W. Otto, Ph.D., Mark H. Pollack, M.D.,Robert A. Gould, Ph.D., John J. Worthington III, M.D.,

Eliza T. McArdle, B.A., and Jerrold F. Rosenbaum, M.D.

Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts, USA

Richard G. Heimberg, Ph.D.

Department of Psychology, Temple University, Philadelphia, Pennsylvania, USA

Abstract—There is a growing body of evidence that social phobia may be treated effec-

tively by either pharmacologic or cognitive-behavioral interventions, but few studies

have examined the relative benefits of these treatments. In this study, we examined the

relative efficacy of pharmacotherapy with clonazepam and cognitive-behavioral group

therapy (CBGT) for treating social phobia. In addition, we examined potential pre-

dictors of differential treatment response. Outpatients meeting Diagnostic and Statisti-

cal Manual of Mental Disorders (3rd ed., revised) criteria for social phobia were ran-

domly assigned to treatment. Clinician-rated and patient-rated symptom severity was

examined at baseline and after 4, 8, and 12 weeks of treatment. All clinician-rated as-

sessments were completed by individuals blind to treatment condition. Patients in both

conditions improved significantly, and differences between treatment conditions were

absent, except for greater improvement on clonazepam on several measures at the 12-

week assessment. Symptom severity was negatively associated with treatment success

for both methods of treatment, and additional predictors—sex, comorbidity with other

The authors thank Bonnie Zucker for assistance in manuscript preparation. Supported in part by

an unrestricted grant from Roche Pharmaceuticals. Portions of this paper were presented at the

Anxiety Disorders Association of America conference in New Orleans, March 21–23, 1997.

Requests for reprints should be sent Michael W. Otto, Ph.D., WACC-815 Massachusetts Gen-

eral Hospital, 15 Parkman Street, Boston, MA 02114.

345

346 M. W. OTTO ET AL.

anxiety or mood disorders, fear of anxiety symptoms, and dysfunctional attitudes—

failed to predict treatment outcome above and beyond severity measures. In summary,

we found that patients randomized to clinical care with clonazepam or CBGT were

equally likely to respond to acute treatment, and pretreatment measures of symptom

severity provided no guidance for the selection of one treatment over another. 2000

Elsevier Science Ltd. All rights reserved.

Keywords: Social phobia; Cognitive-behavioral therapy; Clonazepam; Social anxiety

disorder; Benzodiazepine

The lifetime prevalence of social phobia is approximately 13%, making it

the third most common psychiatric disorder in the community after major de-

pression and alcohol dependence (Kessler et al., 1994). It is characterized by a

persistent fear of one or more social or performance situations that invoke in-

tense anxiety, distress, or avoidance, and which result in significant impair-

ment in social or role functioning (American Psychiatric Association, 1994).

The impairment resulting from social phobia may be extensive and includes

disruption in educational and occupational achievement as well as family and

social relationships (Davidson, Hughes, George, & Blazer, 1993; Schneier et

al., 1994). Comorbid anxiety, mood, and substance use disorders are also com-

mon among individuals with social phobia (Davidson et al., 1993; Schneier,

Hohnson, Hornig, Liebowitz, & Weisman, 1992).

Two modes of treatment, pharmacotherapy and cognitive-behavior ther-

apy (CBT), have demonstrated efficacy in the treatment of social phobia

(Gould, Buckminster, Pollack, Otto, & Yap, 1997). In controlled clinical trials,

monoamine oxidase inhibitors (i.e., phenelzine; Liebowitz et al., 1992; Versi-

ani et al., 1992) and high potency benzodiazepines (Davidson et al., 1993; Mun-

jack, Baltazar, Bohn, Cabe, & Appleton, 1990) have demonstrated efficacy for

social phobia. Initial support also has been provided for selective serotonin re-

uptake inhibitors (SSRIs), including fluvoxamine (van Vliet, den Boer, & Wes-

tenberg, 1994), sertraline (Katzelnick, Kobak, & Greist, 1995), and paroxetine

(Stein et al., 1998). Beta-blockers, although effective for performance anxiety,

appear less effective for social phobia (Pollack & Gould, 1996).

Cognitive-behavioral treatments for social phobia have typically empha-

sized cognitive-restructuring and exposure interventions, either alone or in com-

bination, in either individual or group formats (Heimberg & Juster, 1995; Jus-

ter & Heimberg, 1995). Treatment programs emphasizing these interventions

have met with consistent success (for metaanalytic reviews see Feske & Cham-

bless, 1995; Gould et al., 1997; Taylor, 1996). For example, in a comparison of

cognitive-behavioral group treatment (CBGT) with an educational-support-

ive psychotherapy program, significantly greater improvement was found for

the CBGT program at both endpoint and 5-year follow-up assessments (Heim-

berg et al., 1990; Heimberg et al., 1993).

As specific pharmacologic and cognitive-behavioral treatments for social

phobia are established, questions arise about the relative efficacy of these treat-

347CLONAZEPAM VS. CBGT FOR SOCIAL PHOBIA

ments. In one of the few studies in this area, Gelernter et al., (1991) compared

a CBGT with pharmacotherapy (with alprazolam, phenelzine, or pill placebo)

plus instructions for exposure to phobic stimuli. All three treatments were as-

sociated with significant improvement, and no single treatment was consis-

tently superior to the others. More recently, Heimberg, Juster, and Brown

(1994) compared CBGT with phenelzine, pill-placebo, and attention-placebo

conditions. At posttreatment (12 weeks), CBGT and phenelzine produced re-

sponse rates that were equivalent to each other and superior to both the pill-

placebo and the attention-placebo control group. Whereas phenelzine had a

faster onset of therapeutic effect and demonstrated superior outcome on se-

lect measures at endpoint, CBGT tended to be associated with reduced risk of

relapse during maintenance treatment and follow-up.

Findings of approximately equal efficacy of pharmacologic and cognitive-

behavioral treatments for social phobia are reflected in the results of a recent

metaanalysis (Gould et al., 1997). Sixteen studies with 27 treatment conditions

achieved a mean effect size of 0.74 for CBT relative to control conditions

(most frequently a wait-list control), and 11 studies of the efficacy of pharma-

cotherapy relative to pill placebo provided a mean effect size of 0.62. In this

metaanalysis, the very highest, single-study effect size was obtained for the

high potency benzodiazepine clonazepam. In addition, the overall average ef-

fect size for clonazepam was among the highest (ES 5 .72) for pharmacologic

treatment and equal to the overall effect size for CBT (ES 5 .74).

The present study was designed to test the relative efficacy of two accepted

treatments of social phobia: pharmacotherapy with clonazepam treatment and

Heimberg’s CBGT (Heimberg, Juster, & Hope, 1995; Heimberg et al., 1993).

Each treatment has demonstrated impressive effect sizes in earlier trials; the

purpose of the present study was to examine the relative efficacy of these

treatments and potential predictors of differential outcome.

METHOD

Participants

Participants in this study were 27 men and 18 women who met Diagnostic

and Statistical Manual of Mental Disorders (3rd ed., revised [DSM-III-R]) cri-

teria (American Psychiatric Association, 1987) for social phobia as a primary

diagnosis. Mean age (6 standard deviation) of participants was 39.8 6 10.5

years; mean age of onset of the social phobia was 14.6 6 7.9 years. Seventy-six

percent of the sample met criteria for the generalized subtype of social phobia.

Participants were recruited from clinical referrals and local media adver-

tisements, and all provided written informed consent. Exclusion criteria were

the following: (1) a history of failure to respond to or hypersensitivity to 2 mg

of clonazepam treatment, or known allergy to this or other benzodiazepines;

348 M. W. OTTO ET AL.

(2) a history of previous CBT for social phobia that included exposure and

cognitive restructuring interventions; (3) the presence of a serious or unstable

medical condition; (4) current pregnancy or lactation, or failure to take ade-

quate precautions (medically acceptable means of birth control) against preg-

nancy; (5) a history of schizophrenia, other psychosis, or bipolar disorder; (6)

a current diagnosis of alcoholism or other substance abuse or dependence; and

(7) recent or current suicidal tendency. Patients were not taking psychotropic

medications at baseline, and were not engaged in concurrent psychotherapy

for anxiety difficulties, although patients undergoing other treatments (e.g.,

occasional couples therapy) were eligible as long as the focus of this treatment

was not anxiety management.

Comorbid anxiety or affective disorders were permitted as long as social

phobia was the primary diagnosis. Two patients met criteria for comorbid

panic disorder (both received clonazepam), eight met criteria for comorbid

generalized anxiety disorder (three received CBGT and five received clona-

zepam), and one met criteria for comorbid posttraumatic stress disorder

(PTSD; who received clonazepam). Depressive disorders were equally distrib-

uted among patients in each treatment; 16 patients had a lifetime history of

major depression, four had current major depression, and four had current

dysthymia.

Rates of the generalized subtype of social phobia were almost identical in

the two treatment groups (75% for CBGT and 76% for clonazepam). The two

treatment groups did not differ in age (40.8 6 10.9 for CBGT; 39.0 6 10.2 for

clonazepam) or in age of onset of social phobia (14.7 6 8.3 for CBGT; 14.5 6

7.6 for clonazepam). Seventy-five percent (15 of 20) of the patients who were

treated with CBT were men, compared with 48% (12 of 25) of the patients

treated with clonazepam (Fisher’s exact test, p , .08).

Procedures

Patients referred to the study were initially screened by telephone to deter-

mine whether they met basic diagnostic and inclusion criteria. Patients were

subsequently scheduled for a diagnostic and medical assessment. Diagnostic

evaluation was performed with the Structured Clinical Interview for DSM-III-

R (Spitzer et al., 1990). To assess treatment outcome, symptom severity was

evaluated at baseline, week 4, week 8, and week 12 (posttreatment). All clini-

cian-rated assessments were completed by individuals blind to treatment con-

dition and included the Clinical Global Impression Scale for severity (CGI;

Guy, 1976) with scores ranging from 1 (“not ill”) to 7 (“among most severely

ill”), the Liebowitz Social Anxiety Scale (LSAS; Liebowitz, 1987), and the

Hamilton Rating Scale for Anxiety (HAM-A; Hamilton, 1959). Patient-rated

measures included the Social Interaction Anxiety Scale (SIAS; Mattick &

Clarke, 1989; Heimberg et al., 1992), the Social Phobia Scale (Mattick & Clarke,

349CLONAZEPAM VS. CBGT FOR SOCIAL PHOBIA

1989), the brief version of the Fear of Negative Evaluation Scale (FNE; Leary,

1983), and the Rathus Assertiveness Schedule (RAS; Rathus, 1973).

For purposes of data analysis, the CGI severity score and the SIAS were

chosen on an a priori basis to serve as the primary outcome measures. In addi-

tion to these continuous variables, treatment response was assessed according

to a responder criteria of a CGI severity score of 1 or 2.

In addition to social phobia severity, five additional patient characteristics

were examined as predictors of treatment outcome: patient gender, presence

of a current comorbid depressive disorder (major depression or dysthymia),

presence of a comorbid anxiety disorders (other than simple phobia), severity

of fears of anxiety sensations as assessed by the Anxiety Sensitivity Index (Pe-

terson & Reiss, 1992), and severity of depression-related dysfunctional cogni-

tions as assessed by the Dysfunctional Attitudes Scale (Weissman, 1979).

Treatment Conditions

Clonazepam treatment. Patients in this treatment condition were treated with

clonazepam and monitored on a weekly basis by a psychiatrist experienced in

the pharmacologic treatment of anxiety disorders. Patients were prescribed

0.25 mg twice daily for the first 2 days, increasing to 0.5 mg twice daily by the

end of the first week. Dose was then increased by 1.0 mg per week up to a max-

imum daily dose of 2.0 mg twice daily. Dose titration was flexible and based on

the clinician’s assessment of the relative balance of efficacy and side effects.

Weekly interactions between psychiatrists and patients included review of

anxiety symptoms and potential medication side effects, adjustment of dosage

schedule, supportive interactions encouraging attempts to complete new so-

cial interactions, and monitoring of the success of these interactions as they

were attempted. Prohibited interventions included review of the cognitive-be-

havioral model of social phobia, identification or modification of catastrophic

cognitions associated with social phobia, relaxation training, diaphragmatic

breathing training, or structuring of specific exposure assignments.

CBGT treatment. The CBGT intervention consisted of 12 group treatment ses-

sions of 2.5 hours duration each. This treatment was based on the successful

social phobia treatment protocol developed by Heimberg and associates (Heim-

berg et al., 1995; Hope & Heimberg, 1993) and was guided by a detailed treat-

ment manual (Heimberg, 1991). Core elements of CBGT include: (1) a review

of the cognitive-behavioral model of social phobia; (2) training in the identifi-

cation, analysis, and restructuring of dysfunctional cognitions; (3) exposure to

anxiety-provoking events in the group in a true-to-life fashion; (4) rehearsal of

cognitive restructuring procedures in the context of exposure exercises; and

(5) assigned homework to complete cognitive-restructuring and exposure ex-

ercise situations rehearsed in the context of the group. Treatment groups

350 M. W. OTTO ET AL.

ranged in size between three and eight members, and each was led by two

therapists. Therapists included interns, fellows, and staff psychologists experi-

enced in behavior therapy and supervised by either the first or third author.

Data Analysis

Differences between treatment groups were examined using analysis of co-

variance (ANCOVA), treating the baseline value of each outcome variable as

the covariate. ANCOVAs were used to achieve two goals: (1) statistical ad-

justment of outcome scores to equate participants on baseline severity, and (2)

evaluation of baselines scores (the covariates) as predictors of treatment out-

come. To examine whether baseline severity was differentially predictive of

outcome for the two treatment groups (and thereby assess the homogeneity of

regression assumption required by ANCOVA), preliminary analyses exam-

ined the significance of the interaction between the covariate and treatment

group. Alternate procedures (allowing for heterogeneity of regression; Max-

well & Delaney, 1990) were applied if heterogeneity of regression was evident,

as evaluated by the significance of the interaction term. All outcome measures

were assessed at the end of the trial in both completer (examining data only

from patients completing the trial) and endpoint (data from all patients, car-

rying data from the last available visit forward) analyses. To examine the time-

course of significant differences at outcome further, follow-up tests examined

differences between treatment groups at weeks 4 and 8. In the absence of sig-

nificant differences between groups, improvement over the course of treat-

ment was evaluated using t-tests for repeated measures. Attrition in each

group, and all other contingency table analyses, were examined using Fisher’s

exact tests. Alpha was set at p , .05 for all statistical tests.

RESULTS

Attrition

During the course of the study, 10 of 25 patients (40%) starting clonazepam

discontinued treatment before reaching study endpoint (week 12). Five of

these 10 patients dropped out because of excessive side effects; the remaining

six patients reported lack of efficacy or were lost to follow-up. Five of 20 pa-

tients (25%) who started CBGT discontinued treatment before endpoint.

Two patients attributed their attrition to lack of efficacy and three to inconve-

nience or time constraints associated with the group treatment. Patients who

reached week 12 were considered “completers” (n 5 30). Differences between

groups in attrition during treatment did not reach significance (p , .35).

351CLONAZEPAM VS. CBGT FOR SOCIAL PHOBIA

Comparisons Between Treatments

Table 1 presents means and standard deviations for outcome measures at

each assessment point. Endpoint ANCOVAs, carrying forward the last obser-

vation for each outcome variable, revealed no significant differences between

treatment groups for any clinician- or patient-rated variable. The two treat-

ments were also similar in rates of remission (defined as CGI-severity , 2).

Twenty percent (5 of 25) of the patients treated with clonazepam and 25% (5

of 20) of the patients treated with CBGT met criterion for remission at end-

point. CGI data at each assessment point are provided in Figure 1.

Completer analyses, examining only those patients completing the week-12

assessments, indicated no significant differences on clinician-rated variables.

Patient-rated outcome variables were marked by missing scores for a large

proportion of patients in each treatment group; nonetheless, available scores

revealed significant differences between groups for three outcome variables:

the SIAS (F 5 6.28; df 5 1, 16; p , .05), FNE (F 5 4.15; df 5 1, 20; p , .05),

and RAS (F 5 5.67; df 5 1, 20; p , .05). In all three cases, outcome was better

at visit 12 for patients treated with clonazepam relative to patients treated

with CBGT. The time course of these significant effects was further evaluated

by examining differences between groups at weeks 4 and 8; no significant dif-

ferences between treatments were evident at these evaluation periods. Exami-

nation of patients who completed week 8 but not week 12 assessments indi-

cated that on average, patients dropping out of clonazepam treatment were

more severe on the SIAS and FNE at week 8 than their counterparts; in con-

trast, this trend was not evident for these measures for CBGT patients who did

not complete week 12, but was evident for the RAS.

Improvement Across Treatment

Examination of change in outcome between baseline and endpoint re-

vealed significant improvements for all measures (all p values , .001). Like-

wise, significant changes were evident in all completer analysis (all p values ,

.005). For examination of changes in the SIAS, FNE, and RAS (which demon-

strated differences between groups at week 12), within-group changes were

considered for each group individually; each group demonstrated significant

improvement across treatment (p values , .05). To allow comparison with

published estimates of the pre- to posttreatment effect sizes of CBT for social

phobia (Feske & Chambless, 1995), we used the following equation to com-

pute the CBGT effect size for the present study: (Mpretest 2 Mposttest)/SDpretest.

The average effect size for measures specific to social anxiety or avoidance

(the CGI, SIAS, LSAS, Social Phobia Scale, and FNE) was 0.92; the effect size

352M

.W

.O

TT

OE

TA

L.

TABLE 1

Means and Standard Deviations for Each Outcome Variable at Baseline and Each Week of Treatment

Baseline Week 4 Week 8 Week 12 Endpoint

Outcome

Variable CLZP CBGT CLZP CBGT CLZP CBGT CLZP CBGT CLZP CBGT

CGI–severity 4.8 4.9 3.8 4.0 3.5 3.7 2.7 3.4 3.5 3.6

(0.8) (0.8) (0.7) (1.1) (1.1) (1.2) (1.1) (1.0) (1.3) (1.3)

SIAS 46.9 43.1 37.0 42.1 33.1 36.0 24.6** 39.6 30.9 35.4

(17.9) (12.9) (18.3) (16.3) (17.3) (9.8) (13.9) (14.8) (17.6) (12.4)

LSAS–fear 37.1 34.7 25.3 26.7 21.5 23.4 16.6 21.2 24.8 23.5

(12.9) (11.1) (11.5) (13.3) (13.0) (11.9) (10.3) (10.0) (15.1) (12.3)

LSAS–avoid 32.4 30.5 21.2 22.5 17.6 19.4 12.5 19.0 20.0 20.5

(14.0) (12.9) (12.5) (13.8) (12.1) (13.0) (11.0) (12.4) (17.4) (13.9)

HAM-A 15.6 11.8 9.0 9.2 8.9 10.3 6.8 7.2 9.9 8.2

(6.8) (5.8) (4.8) (5.2) (6.3) (6.6) (6.4) (4.9) (7.1) (6.1)

SPS 34.7 25.5 26.1 24.3 18.8 18.1 12.6 17.5 19.2 18.0

(18.0) (9.4) (17.3) (12.0) (13.0) (9.9) (6.6) (9.2) (16.6) (9.8)

FNE 44.8 44.2 39.6 40.1 35.1 42.1 32.6* 40.9 37.3 40.2

(9.6) (10.1) (9.2) (13.9) (8.7) (6.7) (5.2) (9.2) (10.2) (9.1)

RAS 227.6 223.5 214.9 215.4 4.2 210.8 5.1* 216.2 21.0 211.9

(28.6) (27.7) (25.3) (30.2) (29.5) (23.0) (28.4) (27.1) (30.3) (27.3)

Note. CLZP 5 clonazepam; CGBT 5 cognitive-behavior group therapy; CGI 5 clinician global impression of severity; SIAS 5 Social Interaction Anxiety

Scale; LSAS 5 Liebowitz Social Anxiety Scale (fear and avoidance scales); HAM-A 5 Hamilton Rating Scale for Anxiety; SPS 5 Social Phobia Scale,

FNE 5 Fear of Negative Evaluation scale; RAS 5 Rathus Assertiveness Schedule.

Table values are for study completers, except for the endpoint column, which provides data for all participants, with the data from last available visit

carried forward.

Significant differences between groups at the designated assessment point are: * p , .05; ** p , .005.

353CLONAZEPAM VS. CBGT FOR SOCIAL PHOBIA

Fig. 1. Mean Clinical Global Impression (CGI) severity scores for the treatment groups

at each study visit (completers) and at endpoint (last observation carried

forward). CGBT 5 cognitive-behavioral group therapy.

for severity of general anxiety symptoms was 0.62 as assessed by the HAM-A.

Comparable metaanalytic results have not been published for high-potency

benzodiazepines.

Prediction of Treatment Response

For each of the endpoint analyses, excluding CGI, baseline severity was a

significant and positive predictor of endpoint outcome (R values ranged from

.48 to .60; all p values , .005). For these analyses, there was no evidence of

heterogeneity of regression, indicating that baseline severity was not differen-

tially predictive of outcome for CBGT versus clonazepam.

For the completer analyses, baseline values were predictive of outcome

scores for three measures: the SIAS, RAS, and HAM-A (R values from .45 to

.51; all p values , .05). The analysis of the LSAS avoidance score revealed sig-

nificant heterogeneity of regression. Examination of each treatment groupseparately indicated a significant positive association between the baselineand week 12 scores for the CBGT group, but no such association for patientstreated with clonazepam. This results suggest that for clonazepam treatment,improvement in avoidance was not similar across levels of severity (as tendedto be true for CBGT and for the analyses of other outcome measures).

In addition to the baseline values of the outcome measures, gender, depres-sion comorbidity, anxiety comorbidity, and Anxiety Sensitivity Index andDysfunctional Attitudes Scale scores were examined individually and in inter-action with treatment group as predictors of the primary outcome measures,the CGI and the SIAS. None of these variables increased prediction over thatprovided by baseline severity scores.

354 M. W. OTTO ET AL.

DISCUSSION

Overall, our comparative trial of clonazepam and CBGT for social phobiaindicated that both were effective treatments and revealed few significant dif-ferences in efficacy. Our sample size was not large, but adequate power, in ex-cess of .95, was available for detection of large treatment or predictor effectsizes in the endpoint ANCOVAs. In the endpoint analyses, there was no evi-dence of significant differences between treatment groups, and both treatmentgroups achieved significant pre- to posttreatment changes. In addition, therewere no differences in early treatment effects, as evaluated by week 4 andweek 8 results, and no differences in clinician-rated measures at week 12.However, there was evidence of significantly better outcome on several pa-tient-rated measures of social anxiety and avoidance, anxiety-related cogni-tions, and assertiveness at week 12 for patients treated with clonazepam. Thistreatment advantage was obtained for a subsample of patients completing theSIAS, FNE, and RAS, and may be partially explained by attrition of the moresevere patients in the clonazepam treatment group.

Examination of predictors of treatment outcome in the ANCOVAs pro-vided no consistent evidence of differential prediction of outcome for the twotreatments. Endpoint analyses provided consistent evidence for a positive re-lationship between baseline and outcome severity for both treatment groups;higher pretreatment severity was associated with higher posttreatment sever-ity. Similar, but less consistent, evidence for this relationship was evident incompleter analyses. In multiple regression equations, no other predictor (in-cluding patient gender, anxiety comorbidity, depression comorbidity, anxietysensitivity, and dysfunctional attitudes), examined alone or in interaction withtreatment group, offered a significant increase in prediction over that pro-vided by baseline severity for the main outcome measures. There was evi-dence for differential prediction for clonazepam and CBGT treatment foronly one measure; for LSAS avoidance scores, pretreatment severity was as-sociated with posttreatment outcome only for the CBGT group.

Our results at endpoint are consistent with efficacy estimates provided ineffect-size analysis of controlled studies; clonazepam and CBGT have almostidentical between-group effect size estimates in the Gould et al. (1997) reviewof controlled treatments. Evaluation of within-group effect sizes also revealedconsistency with the available treatment-outcome literature. Feske andChambless (1995) derived effect sizes based on within-group change from pre-to posttreatment in 12 studies of CBT for social phobia. For measures of socialanxiety, the mean pre- to posttreatment effect size was 0.90. In the presentstudy, we found a near identical pre- to posttreatment effect size of 0.92 forCBGT treatment for measures of social anxiety and avoidance. Likewise,Feske and Chambless (1995) reported a mean effect size of 0.58 for measuresof anxiety and depression severity; we obtained a CBGT effect size of 0.62 forchanges in general anxiety as assessed by the HAM-A.

355CLONAZEPAM VS. CBGT FOR SOCIAL PHOBIA

The results of our study suggest that, on average, patients randomized to

clonazepam and CBGT treatment are equally likely to respond during the

acute treatment phase, although treatment completers may enjoy a stronger ad-

vantage with clonazepam. It is important to note that clonazepam treatment,

similar to the pharmacotherapy conditions examined by Gelertner et al., (1991),

was conducted in the context of clinicians’ encouragement of patients to enter

avoided social situations. This encouragement to conduct self-guided expo-

sure may well have aided treatment gains for the clonazepam condition, pro-

viding a clinically meaningful index of improvement when this element of psy-

chosocial care is used as part of standard psychopharmacology. To disallow

these supportive interventions is to construct a treatment that diverges from

recommendations for standard pharmacologic treatment of social phobia

(Sutherland & Davidson, 1995). Nonetheless, it should be understood that

these procedures may have increased efficacy estimates for the clonazepam

condition and may have provided a more stringent test of the relative efficacy

of CBGT.

Examination of measures of social phobia severity revealed no consistent

evidence for the selection of one treatment over another; predictors of poorer

response to one treatment were also predictors of poorer response to the

other treatment. As such, there is little empirical evidence to guide treatment

choice firmly in the short-term, and choice of treatment may depend more on

treatment availability and patient preference. The issue of treatment prefer-

ence is clouded by the exceptionally low rates at which individuals with social

phobia seek treatment (Schneier et al., 1992), and in our sample we observed

an average 24-year interval between the onset of the social phobia and entry

into our treatment trial. Treatment seeking may be limited by the tendency of

those with social phobia to avoid the social anxiety associated with the inter-

personal interactions inherent in treatment, but our study provided no evi-

dence that once initiated, a psychosocial treatment is less tolerable than a

pharmacologic treatment. Overall, we observed a 40% drop-out rate for clo-

nazepam treatment and a 25% percent drop-out rate for CBGT.

Treatment choice should also be informed by long-term outcome, but few

studies have addressed this issue in social phobia. Available studies indicate

the return of symptoms after medication discontinuation (Davidson, Tupler, &

Potts, 1994; Versiani et al., 1992) and that maintenance or additional treat-

ment is common among individuals who initiated pharmacotherapy (Suther-

land, Tupler, Colket, & Davidson, 1996). There is also tentative evidence for

slippage of treatment effects even if pharmacotherapy is maintained (Liebo-

witz et al., 1999). In contrast, a number of studies of CBT for social phobia in-

dicate that additional treatment gains are achieved over follow-up intervals

(Gould et al., 1997; Heimberg et al., 1993; Mersch, Emmelkamp, & Lips, 1991;

Turner, Beidel, & Cooley-Quille, 1995), suggesting that patients continue to

apply principles and skills learned during the short-term treatment phase.

356 M. W. OTTO ET AL.

In summary, our study provides evidence of generally equal efficacy for

clonazepam and CBGT, with some evidence of more optimal acute outcome

for patients maintaining clonazepam treatment at the end of the trial. Our

study is limited by a relatively low sample size and a relatively high attrition

rate. However, effect size comparisons indicate our acute outcome findings

are consistent with previous studies and are consistent with other recent com-

parisons of CBGT and pharmacotherapy (Heimberg et al., 1994). The best

strategy for long-term treatment of patients with social phobia remains an

open question, with some evidence suggesting consolidation and extension of

treatment gains at follow-up in patients treated with CBT.

REFERENCES

American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders

(3rd ed., revised). Washington, DC: Author.

American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders

(4th ed.). Washington, DC: Author.

Davidson, J. R. T., Hughes, D. L., George, L. K., & Blazer, D. G. (1993). The epidemiology of so-

cial phobia: Findings from the Duke Epidemiological Catchment Area Study. Psychological

Medicine, 23, 709–718.

Davidson, J. R. T., Potts, N., Richichi, E., Krishnan, R., Ford, S. M., Smith, R., Wilson W. H. (1993).

Treatment of social phobia with clonazepam and placebo. Journal of Clinical Psychopharma-

cology, 13, 423–428.

Davidson, J. R. T., Tupler, L. A., & Potts, N. L. S. (1994). Treatment of social phobia with benzodi-

azepines. Journal of Clinical Psychiatry, 55(Suppl 6), 28–32.

Feske, U., & Chambless, D. L. (1995). Cognitive behavioral verses exposure only treatment for so-

cial phobia: A meta-analysis. Behavior Therapy, 26, 695–720.

Gelernter, C. S., Uhde, T. W., Cimbolic, P., Arnkoff, D. B., Vittone, B. J., Tancer, M. E., & Bartko,

J. J. (1991). Cognitive-behavioral and pharmacological treatments of social phobia: A con-

trolled study. Archives of General Psychiatry, 48, 938–945.

Gould, R. A., Buckminster, S., Pollack, M. H., Otto, M. W., & Yap, L. (1997). Cognitive-behavioral

and pharmacological treatment for social phobia: A meta-analysis. Clinical Psychology: Sci-

ence & Practice, 4, 291–306.

Guy, W. (1976). Assessment manual for psychopharmacology. Washington, DC: US Government

Printing Office.

Hamilton, M. (1959). The assessment of anxiety states by rating. British Journal of Medical Psy-

chology, 32, 50–55.

Heimberg, R. G., Dodge, C. S., Hope, D. A., Kennedy, C. R., Zollo, L., & Becker, R. E. (1990).

Cognitive behavioral group treatment for social phobia: Comparison with a credible placebo

control. Cognitive Therapy and Research, 14, 1–23.

Heimberg, R. G. (1991). Cognitive behavioral treatment of social phobia in a group setting: A treat-

ment manual (2nd ed.). Unpublished manuscript.

Heimberg, R. G., Mueller, G. P., Holt, C. S., Hope, D. A., & Liebowitz, M. R. (1992). Assessment

of anxiety in social interaction and being observed by others: The social interaction anxiety

scale and the social phobia scale. Behavior Therapy, 23, 53–73.

Heimberg, R. G., Salzman, D. G., Holt, C. S., & Blendell, K. A. (1993). Cognitive-behavioral group

treatment for social phobia: Effectiveness at five-year follow-up. Cognitive Therapy and Re-

search, 17, 325–339.

357CLONAZEPAM VS. CBGT FOR SOCIAL PHOBIA

Heimberg, R. G., & Juster, H. R. (1995). Cognitive-behavioral treatments: Literature review. In

R. G. Heimberg, M. R. Liebowitz, & D. A. Hope (Eds.), Social phobia: Diagnosis, assessment

and treatment (pp. 261–309). New York: Guilford Press.

Heimberg, R. G., Juster, H. R., & Hope, D. A., & Mattia, J. I. (1995). Cognitive behavioral group

treatment for social phobia: Description, case presentation and empirical support. In M. B.

Stein. (Ed.), Social phobia: Clinical and research perspectives (pp. 293–321). Washington, DC:

American Psychiatric Press.

Hope, D. A., & Heimberg, R. G. (1993). Social phobia and social anxiety. In D. H. Barlow (Ed.),

Clinical handbook of psychological disorders: A step-by-step treatment manual (2nd ed., pp. 99–

136). New York: Guilford Press.

Juster, H. R., & Heimberg, R. G. (1995). Social phobia: Longitudinal course and long-term out-

come of cognitive-behavioral treatment. Psychiatric Clinics of North America, 18, 821–842.

Katzelnick, D. J., Kobak, K. A., & Greist, J. H. (1995). Sertraline in social phobia: A double-blind,

placebo-controlled crossover study. American Journal of Psychiatry, 152, 1368–1371.

Kessler, R. C., McGonagle, K. A., Zhao, S., Nelson, C. B., Hughes, M., Eshleman, S., Wittchen,

H., & Kendler, K. S. (1994). Lifetime and twelve month DSM-III-R psychiatric disorders in the

United States. Archives of General Psychiatry, 51, 8–19.

Leary, M. R. (1983). A brief version of the Fear of Negative Evaluation scale. Personality and So-

cial Psychology Bulletin, 9, 371–375.

Liebowitz, M. R. (1987). Social phobia. Modern Problems in Pharmacopsychiatry, 22, 141–173.

Liebowitz, M. R., Heimberg, R. G., Schneier, F. R., Hope, D. A., Davies, S., Holt, C. S., Goetz, D.,

Juster, H. R., Lin, S. H., Bruch, M. A., Marshall, R. D., & Klein, D. F. (1999). Cognitive-behav-

ioral group theory versus phenelzine in social phobia: Long-term outcome. Depression and

Anxiety, 10, 89–98.

Liebowitz, M. R., Schneier, F. R., Campeas, R., Hollander, E., Hatterer, J., Fyer, A., Gorman, J.,

Papp, L., Davies, S., & Gully, R. (1992). Phenelzine vs. atenolol in social phobia: A placebo-

controlled comparison. Archives of General Psychiatry, 49, 290–300.

Mattick, R. P., & Clarke, J. C. (1989). Development and validation of measures of social phobia

scrutiny fear and social interaction anxiety. Unpublished manuscript.

Maxwell, S. E., & Delaney, H. D. (1990). Designing experiments and analyzing data. A model com-

parison perspective. Belmont, CA: Wadsworth Publishing Company.

Mersch, P. P. A., Emmelkamp, P. M. G., & Lips C. (1991). Social phobia: Individual response pat-

terns and the long-term effects of behavioral and cognitive interventions: A follow-up study.

Behaviour Research and Therapy, 29, 357–362.

Munjack, D. J., Baltazar, P. L., Bohn, P. B., Cabe, D. D., & Appleton, A. A. (1990). Clonazepam in

the treatment of social phobia: A pilot study. Journal of Clinical Psychiatry, 51, 35–40.

Peterson, R. A., & Reiss S. (1992). Anxiety sensitivity index revised test manual. Worthington,

Ohio: International Diagnostic Services, Inc.

Pollack, M. H., & Gould, R. A. (1996). The pharmacotherapy of social phobia. International Clini-

cal Psychopharmacology, 11, 71–75.

Rathus, S. A. (1973). A 30-item schedule for assessing assertive behavior. Behavior Therapy, 4,

398–40.

Schneier, F. R., Hohnson, H., Hornig, C. D., Liebowitz, M. R., & Weissman, M. M. (1992). Social

phobia: Comorbidity and morbidity in an epidemiologic sample. Archives of General Psychia-

try, 49, 282–288.

Schneier, F. R., Heckelman, L. R., Garfinkel, R., Campeas, R., Fallon, B. A., Gitow, A., Street, L.,

Del Bene, D., & Liebowitz, M. R. (1994). Functional impairment in social phobia. Journal of

Clinical Psychiatry, 55, 322–331.

Spitzer, R. L., Williams, J. B. W., & Gibbon, M., & First, M. B. (1990). Structured clinical interview

for DSM-III-R—Patient edition (with psychotic screen). Washington, DC: American Psychiatric

Press.

358 M. W. OTTO ET AL.

Stein, M. B., Liebowitz, M. R., Lydiard, R. B., Pitts, C. B., Bushnell, W., & Gergel, I. (1998). Paro-

xetine treatment of generalized social phobia (social anxiety disorder): A randomized, con-

trolled trial. Journal of the American Medical Association, 280, 708–713.

Sutherland, S. M., & Davidson, J. R. T. (1995). b-Blockers and benzodiazepines in pharmacother-

apy. In M. B. Stein (Ed.), Social phobia: Clinical and research perspectives (pp. 323–326). Wash-

ington, DC: American Psychiatric Press.

Sutherland, S. M., Tupler, L. A., Colket, J. T., & Davidson, J. R. T. (1996). A 2-year follow-up of

social phobia: Status after a brief medication trial. Journal of Nervous and Mental Disorders,

184, 731–738.

Taylor, S. (1996). Meta-analysis of cognitive-behavioral treatments for social phobia. Journal of

Behavior Therapy and Experimental Psychiatry, 27, 1–9.

Turner, S., Beidel, D., & Cooley-Quille, M. R. (1995). Two-year follow-up of social phobics treated

with social effectiveness therapy. Behaviour Research and Therapy, 33, 553–555.

van Vliet, I. M., den Boer, J. A., & Westenberg, H. G. M. (1994). Psychopharmacological treatment

of social phobia: A double-blind placebo controlled study with fluvoxamine. Psychopharmacol-

ogy, 115, 128–134.

Versiani, M., Nardi, A. E., Mundim, F. D., Alves, A. B., Liebowitz, M. R., & Amrein, R. (1992)

Pharmacotherapy of social phobia: A controlled study with moclobemide and phenelzine. Brit-

ish Journal of Psychiatry, 161, 353–360.

Weissman, A. N. (1979). The Dysfunctional Attitude Scale: A validation study. Dissertation Ab-

stracts International, 40, 1389–1390.