© 2000 WILEY-LISS, INC.
DEPRESSION AND ANXIETY 11:99–104 (2000)
THERAPEUTIC ADVANCES: PAROXETINE FOR THE
TREATMENT OF SOCIAL ANXIETY DISORDER
R. Bruce Lydiard, Ph.D., M.D.,1* and Julio Bobes2
Data from early studies of selective serotonin reuptake inhibitors have shown
that these agents are effective in the treatment of social anxiety disorder (social
phobia). This review highlights the outcomes of three large clinical trials of
paroxetine in patients with social anxiety disorder. In two of the studies, pa-
tients received a flexible dose of paroxetine (20–50 mg/day) or placebo; the
third trial was a fixed-dose study, in which patients received paroxetine 20,
40, or 60 mg/day, or placebo. A total of 861 subjects were randomized to
treatment for 12 weeks, in centers across the U.S.A., Canada, Europe, and
South Africa. The primary outcome measures were the Clinical Global Im-
pressions (CGI) Global Improvement item and Liebowitz Social Anxiety Scale
(LSAS) Total Score.
In each of the studies, 45–66% of patients receiving paroxetine were rated
as responders (very much or much improved on the CGI scale). Paroxetine
treatment improved symptoms of social anxiety, as measured by the LSAS,
compared with placebo. Differences between paroxetine and placebo groups
were statistically significant and were clinically relevant within each study. In
general, paroxetine was well tolerated.
Paroxetine is effective for the treatment of social anxiety disorder. Based on
the findings from these studies, a starting dose of 20 mg/day is recommended.
The range of efficacy appears to be 20–50 mg/day for most patients. Depres-
sion and Anxiety 11:99–104, 2000. © 2000 Wiley-Liss, Inc.
Key words: social anxiety disorder; social phobia; clinical trial; randomized;
Social anxiety disorder (social phobia) is a common,
potentially crippling anxiety disorder which affects an
estimated 7–13% of the general population at some
point [Kessler et al., 1994; Wittchen et al., 1999;
Lecrubier et al., 2000]. Most individuals (about 75%)
with social anxiety disorder have the generalized sub-
type (the more severe form) in which most types of
social situations are feared. The remainder forms a
heterogeneous group that includes individuals who
fear a single or several performance situations, but not
most social situations (nongeneralized subtype). The
pathological social fear and avoidance associated with
social anxiety disorder can result in poor academic
achievement, social limitations, and often financial de-
pendence for the sufferer [Davidson, 1994; Schneier
et al., 1992; Weiller et al., 1996; Wittchen and Be-
loch, 1996]. Table 1 shows the main features of social
anxiety disorder. The lifetime prevalence rates of the
disorder found in community studies range from 4%
[Schneier et al., 1992] up to 14% [Magee et al., 1996].
Social anxiety disorder is complicated by a 70–80%
lifetime risk of comorbid psychiatric conditions, such
as depression and alcohol abuse [Merikangas and
Social anxiety disorder is amenable to both cogni-
tive-behavioral and pharmacological treatment, as dis-
cussed in the article by Davidson  and elsewhere
1Department of Psychiatry, Medical University of South
Carolina, Charleston, South Carolina
2Facultad de Medicina, Universidad de Oviedo, Spain
Presented at a satellite symposium to the XXIst CINP Congress,
Glasgow, United Kingdom, 13 July 1998.
Contract grant sponsor: SmithKline Beecham Pharmaceuticals.
*Correspondence to: R. Bruce Lydiard, Ph.D., M.D., Director,
Mood and Anxiety Disorders Program, Department of Psychiatry,
Psychopharmacology Unit, Medical University of South Carolina,
67 President Street, Box 250861, Charleston, SC 294245.
Received for publication 26 November 1999; Accepted 7 Decem-
100Lydiard and Bobes
[Jefferson, 1995; den Boer, 1997; Boerner and Möller,
1998; Davidson, 1998; Lydiard, 1998]. Recently, the
International Consensus Group on Depression and
Anxiety recommended selective serotonin re-uptake
inhibitors (SSRIs) as the therapy of choice for social
anxiety disorder [Ballenger et al., 1998].
Clinical data derived from case reports and open-la-
bel studies of SSRIs have indicated their potential for
use in social anxiety disorder. Acute efficacy has been
demonstrated in small-scale studies of fluvoxamine
[van Vliet et al., 1994], sertraline [Katzelnick et al.,
1995], and paroxetine [Stein et al., 1996] (Table 2). A
relapse prevention study by Stein et al.  showed
that abrupt placebo substitution for paroxetine therapy
following successful treatment of generalized social
anxiety disorder led to relapse of symptoms, while con-
tinued paroxetine therapy was of benefit. Based on the
optimistic preliminary data from these smaller studies,
three large-scale studies of paroxetine in social anxiety
disorder were undertaken. These studies, which repre-
sent the largest sample (n = 861) of SSRI treatment of
patients with social anxiety disorder, were all intention-
ally structured with similar designs to allow easy com-
parison between studies [Stein et al., 1998; Baldwin et
al., 1999] (SmithKline Beecham, unpublished). These
studies are highlighted below.
GENERAL STUDY DESIGN
Flexible-dose studies. Two multicenter, double-
blind, placebo-controlled, flexible-dose clinical trials
of paroxetine were conducted: one in the U.S.A.
and Canada (Study 382), which included 187 sub-
jects, and one in Europe and South Africa (Study
502), which included 290 subjects. Each study em-
ployed a 1-week, single-blind, placebo run-in pe-
riod followed by a 12-week treatment phase. The
dose of paroxetine was 20 mg/day for the first 2
weeks. After the initial 2 weeks, the daily dosage
could be increased by 10 mg daily per week to a
maximum daily dose of 50 mg (Fig. 1).
Fixed-dose study. The 12-week fixed-dose study
(Study 454) was conducted in the U.S.A. and Canada.
A total of 384 patients were randomized to receive
paroxetine 20, 40, or 60 mg/day, or placebo. A 1-
week, single-blind, placebo run-in preceded random
assignment to the treatment groups. Paroxetine was
initiated at 20 mg/day for the first week of the trial
for all patients in the active treatment groups. At
week 2, the dosage was increased to 40 mg/day for
the 40 or 60 mg/day treatment groups and at the
start of the third week, the dosage was further in-
creased to 60 mg/day for the 60 mg/day treatment
group. For the 20 and 40 mg group, additional pla-
cebo tablets were added so that all groups took the
same number of tablets (three per day) of paroxetine
INCLUSION AND EXCLUSION CRITERIA
Inclusion and exclusion criteria were similar across
all studies. Subjects were recruited by referral or from
advertisements in the media. Outpatients with DSM-
IV [American Psychiatric Association, 1994] criteria
for social anxiety disorder (social phobia) were evalu-
ated with the Structured Clinical Interview for the
DSM-IV [First et al., 1995] or the Mini International
Neuropsychiatric Interview [Sheehan et al., 1998].
Patients (men and women) with social anxiety disor-
der, who were at least 18 years old, and had given
written informed consent were eligible for the studies.
The main exclusion criteria were other Axis I dis-
order(s) which might interfere with assessment of
efficacy in the study (e.g., major depression, panic
disorder, schizophrenia, bipolar affective disorder,
body dysmorphic disorder, recent alcohol, or sub-
stance abuse or dependence), or a significant risk of
suicide. In Studies 454 and 502, patients with a
Hamilton Depression Rating Scale (HAM-D) score of
15 or more were also excluded [Hamilton, 1967]. Pa-
tients who required other psychotropic medication
(anxiolytics, antidepressants, or neuroleptics) were ex-
cluded, as were patients with a history of paroxetine
intolerance. Patients with unstable medical disorders,
or conditions that might interfere with study proce-
dures were also excluded.
TABLE 1. Key clinical features of social anxiety
Fear of embarrassment in one or more or social situations
Fear of negative evaluation
Exposure to the situation reliably causes anxiety, often with panic-
The fear is recognized as excessive or unreasonable
The social situation is often avoided, or is endured with dread
Panic attacks are associated with social situations and do not occur
Causes significant distress or interference with in one or more
domains of functioning
Generalized (encompasses most social interactions)
Nongeneralized (a few discrete situations such as public
speaking, eating, writing, etc. in front of others, but not most
TABLE 2. Early studies of SSRIs in social anxiety
n Responders Reference
ParoxetineOpen36 77 Stein et al.
Katzelnick et al.
van Vliet et al.
aSSRI, selective serotonin reuptake inhibitors.
Theoretical Paper: Taking Control of Social Anxiety Disorder101
EFFICACY AND SAFETY ASSESSMENTS
All patients were assessed weekly for the first 4
weeks of the trial, then at weeks 6, 8, and 12 (end
point). The primary efficacy variables were the pro-
portion of patients with a Clinical Global Impressions
(CGI) Global Improvement score of 1 (very much
improved) or 2 (much improved) [Guy, 1976], and the
mean change from baseline in Liebowitz Social Anxi-
ety Scale (LSAS) Total Score [Liebowitz, 1987].
Changes in anxiety symptoms were also measured us-
ing the Social Anxiety and Distress Scale (SADS)
[Watson and Friend, 1969]. Disability was assessed
with the Sheehan Disability Scale (SDS) [Sheehan,
1983], which assessed patient-rated impairment of
work, social life, and family-life domains. CGI Sever-
ity of Illness scores were used to assess overall disease
severity (except Study 382).
A routine medical examination preceded study par-
ticipation, and monitoring for adverse experiences was
conducted at each visit. Standard laboratory tests were
obtained at baseline and end point. Patients with clini-
cally significant laboratory measures were excluded.
In each study, the proportion of responders (defined
as a CGI Global Improvement score of 1 or 2) was
analyzed by logistic analysis using the categorical
modeling procedure of the Statistical Analysis System
with a model that included an effect for treatment.
The mean change from baseline in LSAS Total Score
and the secondary efficacy scales were analyzed using
parametric analysis of variance. All statistical tests com-
paring paroxetine with placebo were two-tailed and
with an overall significance level of 0.05. In Study 454,
Dunett’s test was used to compare each paroxetine dose
with placebo to maintain an overall significance level of
0.05 and the adjusted level of significance was less than
0.019. Efficacy analyses included the intent-to-treat
population (i.e., patients who took at least one dose of
medication and had at least one postrandomization as-
sessment). The primary time point of interest was end
point for each study. For patients who did not com-
plete the entire study, the last observation during treat-
ment was carried forward to end point.
Patient sample characteristics were similar across
the three studies (Table 3). The mean age of onset of
social anxiety disorder was 15.8 years and the mean
current age of the study population was 36 years (en-
tire patient sample).
Paroxetine-treated patients exhibited clinically and
statistically significant improvement in the primary ef-
ficacy measures (CGI and LSAS scores), compared
with placebo-treated patients.
CGI Global Improvement measures. The pro-
portion of responders (“much improved” or “very
much improved”) in Study 382 was 55% in the par-
oxetine group compared with 24% in the placebo
group, and in Study 502, the response rate was 66%
vs. 32% in the paroxetine and placebo groups, respec-
tively (Fig. 2). The mean paroxetine-placebo differ-
ence in response was similar in each study: 31% in
Study 382 and 34% in Study 502.
In the fixed-dose study, similar percentages of patients
in the paroxetine groups were rated as responders on
the CGI-Improvement assessment (Fig. 2). However,
only the 40 mg/day group was statistically different
from the placebo group. Interestingly, the percentages
of patients rated as responders in the three active fixed-
Figure 1.Design of the flexible dose studies. Adapted from Baldwin et al. .
102Lydiard and Bobes
dose study groups (ranging from 45%, 47%, and 43% at
20, 40, and 60 mg, respectively) was lower than in either
flexible-dose study (55% and 66%), but the placebo re-
sponse rates were comparable (20–30% range) across all
three studies. In the fixed-dose trial, if all three active
drug groups are combined, the mean proportion of re-
sponders in the paroxetine group (45%) was signifi-
cantly greater than with placebo (28%) at week 12 (P <
0.01). Most of the difference between the paroxetine
and placebo groups was due to the percentage of re-
sponders who were “very much improved”: 19%, 21%,
and 22% with increasing doses of paroxetine, compared
with 8% in the placebo group.
LSAS rating. Across all three studies, patients re-
ceiving paroxetine experienced greater improvements
in their social anxiety symptoms than placebo-treated
patients, as measured by the average reduction in LSAS
scores from baseline (Fig. 3). In both flexible-dose stud-
ies, the paroxetine-treated patients exhibited signifi-
cantly more improvement in the LSAS score ratings
than placebo recipients by week 4 and through all of
the subsequent assessments to end point (week 12). At
end point, the differences in LSAS score improvement
between the paroxetine (Study 382, 31 points; Study
502, 29 points) and placebo groups (Study 382, 15
points; Study 502, 16 points) were statistically signifi-
cant. In Study 454, the reduction in LSAS scores in the
20 mg/day dose group (31 points) was statistically
greater than in the placebo group (15 points). The two
higher dose groups (paroxetine 40 and 60 mg/day) both
exhibited the same degree of improvement (25 points)
in LSAS ratings, but this difference was not statistically
significantly different from placebo (Fig. 3).
Secondary efficacy measures. Improvements in
all secondary efficacy variables were greater with
paroxetine than with placebo. The results from the
bal Improvement scores of 1 or 2 (“responders”). *P < 0.05,
paroxetine vs. placebo.**P < 0.019 (Dunett’s test), paroxetine
vs. placebo. aData from Stein et al. . bData from Baldwin
et al. . cSmithKline Beecham, data on file.
Proportion of patients at end point with CGI Glo-
bowitz Social Anxiety Scale (LSAS) total scores at end point. *P
< 0.05, paroxetine vs. placebo. **P < 0.019 (Dunett’s test),
paroxetine vs. placebo. aData from Stein et al. . bData
from Baldwin et al. . cSmithKline Beecham, data on file.
Social anxiety symptoms–mean reduction in Lie-
TABLE 3. Patient characteristics of the study populations
N = 94
N = 93
n = 139
n = 151
n = 289d
n = 95
Mean age (y)
Age range (y)
aData from Stein et al. .
bData from Baldwin et al. .
cSmithKline Beecham, data on file.
dCombined data from the three active treatment dose groups.
eSome totals greater than 100% due to rounding.
Theoretical Paper: Taking Control of Social Anxiety Disorder103
secondary efficacy measures of the flexible dose stud-
ies are shown in Table 4. In the fixed-dose study, sev-
eral of these variables showed statistical significance
for each dose level, but as might be expected, not all
doses showed statistical superiority on all measures. At
no time was placebo superior to paroxetine.
TREATMENT EFFECTS ON DEPRESSION
In the Europe-South Africa flexible dose study,
change in HAM-D rating was determined for the two
treatment groups. The mean HAM-D scores at baseline
and end point, respectively, were 6.2 and 4.2 for the
paroxetine group, and 6.7 and 6.5 for the placebo group.
Thus, on entry to the study, subjects had low levels of
depression, and little change in depressive symptoms
was observed following treatment. Therefore, the over-
all treatment effect observed for symptoms of social
anxiety disorder does not appear to be dependent on a
change in any symptoms of coexistent depression. Simi-
lar results were observed in the fixed-dose study.
In the flexible-dose studies (20–50 mg/day), the
mean dose of paroxetine at study end point was
36.6 mg/day in the U.S.A. and Canada study (Study
382) and 34.7 mg/day in the European and South
Africa study (Study 502). Not surprisingly, optimal in-
dividual patient dosage varied within both studies. In
the dose-finding trial (Study 454), a good response to
treatment was observed with 20 or 40 mg/day par-
oxetine, with no added efficacy advantage of raising
the dose to 60 mg/day.
Overall, the most common treatment emergent
adverse experiences in the paroxetine group were ab-
normal ejaculation, nausea, somnolence, insomnia,
headache, and asthenia. Headache, insomnia and asthe-
nia, however, were reported at a similar rate in the pla-
cebo groups (Table 5).Most adverse experiences in
both groups were mild or moderate. The rate of ad-
verse experiences classed as “severe” was similar for the
paroxetine (9%) and placebo (6%) groups.
For all studies combined, lack of efficacy was the
most common reason for withdrawal from the placebo
group (12% vs. 2% with paroxetine). However, there
were more withdrawals due to adverse experiences
from the paroxetine group (16% vs. 4% with placebo).
Withdrawal for other reasons included: patients lost to
follow-up (8% paroxetine, 6% placebo), and lack of
efficacy (2% paroxetine, 12% placebo). In the indi-
vidual trials, the reasons for withdrawal followed the
same pattern as for the overall study population.
DISCUSSION AND CONCLUSIONS
This article has summarized three large placebo-
controlled, multicenter studies of the treatment of pa-
tients with social anxiety disorder with paroxetine. As
a class, SSRIs show promise for the treatment of social
anxiety disorder. The three trials described in this re-
view show consistently that paroxetine (20–50 mg/
day) is an effective and well-tolerated treatment for
social anxiety disorder. Paroxetine produced overall
improvement, in addition to providing more specific
benefits with respect to anxiety and avoidance symp-
toms, and social anxiety-related disability.
In the flexible-dose studies, a more robust response
was seen than in the fixed-dose study. There were
similar baseline sample characteristics and sample sizes
in the treatment groups. This suggests that factors in-
herent in the differences between study designs may
have reduced the rates of response in the fixed-dose
study. It may be that some patients needed higher or
lower doses than allowed by the fixed-dose design. Be-
cause all subjects were committed to a single dose, in-
dividual optimization was not possible. Additionally,
no further improvement in efficacy in the high-dose
TABLE 4. Efficacy of paroxetine in the treatment of
social anxiety disorder, data represents mean change
from baseline at end point.
Paroxetine Placebo Paroxetine Placebo
CGId severity of illness
SADS total score
aData from Stein et al. .
bData from Baldwin et al. .
cND = not described.
dCGI, Clinical Global Impressions.
eLSAS, Liebowitz Social Anxiety Scale.
fSDS, Sheehan Disability Scale.
*Statistically significant improvement with paroxetine compared to with
placebo (P < 0.05).
TABLE 5. Most common adverse experiences reported
over 12 weeks of treatment, combined data for the
three large-scale studies
(n = 522)
(n = 339)
aCorrected for gender.
104Lydiard and Bobes
(60 mg/day) group suggests that for most patients
there is no advantage in prescribing doses higher than
50 mg/day. Based on the accrued evidence, a starting
dose of 20 mg/day is recommended. The pattern of
adverse experiences attributed to paroxetine in these
studies is similar to those reported for paroxetine in
previous studies [reviewed by Gunasekara et al., 1998].
Paroxetine is the first SSRI to receive approval for use
in social anxiety disorder in the U.K., other European
countries, and more recently in the U.S.A.
The process of development of paroxetine as a
treatment for social anxiety disorder, as described in
this article, is the first important step in an ongoing
effort to develop effective and safe treatment for this
condition. Since approximately 60–70% of patients
with social anxiety disorder respond to treatment with
SSRIs, there is a clear need for further expansion of
the treatment armamentarium. Whether patients who
are responsive or incompletely responsive to one SSRI
may respond to another SSRI or a monoamine oxidase
inhibitor remains an open and important clinical ques-
tion. It appears that the other SSRIs, some of the
newer non-SSRI antidepressant agents, high potency
benzodiazepines, and the anticonvulsant gabapentin
may also be effective as treatments for this prevalent
and serious disorder [Lydiard, 1998; Ballenger et al.,
1998; Pande et al., 1999]. Studies comparing these
newer agents both with each other as well as with the
other efficacious treatments are needed to provide im-
portant clinical information on their relative efficacy
Acknowledgments. “Face to Face with Social
Anxiety Disorder” supported by an educational grant
from SmithKline Beecham Pharmaceuticals.
American Psychiatric Association. 1994. Diagnostic and statistical
manual of mental disorders. Fourth edition. Washington, DC:
American Psychiatric Association Inc.
Baldwin D, Bobes J, Stein DJ, Ingebor S, Faure M. 1999. A random-
ized, double-blind, placebo-controlled study of paroxetine in the
treatment of social phobia/social anxiety disorder. Br J Psychiatr
Ballenger JC, Davidson JRT, Lecrubier Y, Nutt D. 1998. Consensus
statement on social anxiety disorder. J Clin Psychiatr 59 (suppl
Boerner RJ, Möller HJ. 1998. Therapeutic strategies for patients
with social phobia. Int J Psychiatr Clin Prac 2:107–113.
Davidson JR. 1994. Social phobia: outlook for the ’90s. J Clin Psy-
Davidson JR.1998. Pharmacotherapy of social anxiety disorder / so-
cial phobia. J Clin Psychiatr 59 (suppl 17):47–51.
Davidson JRT. 2000. Social anxiety disorder under scrutiny. Depres-
sion and Anxiety 11:93–98.
den Boer JA. 1997. Social phobia: epidemiology, recognition and
treatment. BMJ 315:796–800.
First M, Spitzer RL, Williams, JBW, Gibbon M. 1995. Structures
clinical interview for DSM-IV—patient edition. Washington, DC:
American Psychiatric Press.
Gunasekara NS, Nobel S, Benfield P. 1998. Paroxetine. An update of
its pharmacology and therapeutic use in depression and a review
of its use in other disorders. Drugs 55:85–120.
Guy W. 1976. ECDEU assessment manual for psychopharmacology.
Washington DC: Department of Health, Education and Welfare
Publications (ADM) 76-338. p 217–222.
Hamilton M. 1967. Development of a rating scale for primary de-
pressive illness. Br J Soc Clin Psychol 6:278–296.
Jefferson JW. 1995. Social phobia: a pharmacologic treatment over-
view. J Clin Psychiatry 56 (suppl 5):18–24.
Katzelnick DJ, Kobak KA, Greist JH, Jefferson JW, Mantle JM,
Serlin RC. 1995. Sertraline for social phobia: a double-blind, pla-
cebo-controlled crossover study. Am J Psychiatr 152:1368–1371.
Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M,
Eshleman S, Wittchen HU, Kendler KS. 1994. Lifetime and 12-
month prevalence of DSM-III-R psychiatric disorders in the
United States: results from the National Comorbidity Survey.
Arch Gen Psychiatry 51:8–19.
Lecrubier Y, Wittchen HU, Faravelli C, Bobes J, Patel A, Knapp M.
2000. A European perspective on social anxiety disorder. Eur Psy-
chiatry (in press).
Liebowitz MR. 1987. Social phobia. Modern Problems in Pharma-
Lydiard RB. 1998. The role of drug therapy in social phobia. J Aff
Dis 50 (suppl 1):S35–39.
Magee WJ, Eaton WW, Wittchen HU, McGonagle KA, Kessler
RC. 1996. Agoraphobia, simple phobia, and social phobia in the
National Comorbidity Survey. Arch Gen Psychiatry 53:159–168.
Merikangas KR, Angst J. 1995. Comorbidity and social phobia: evi-
dence from clinical, epidemiologic and genetic studies. Eur Arch
Psychiatry Clin Neurosci 244:297–303.
Pande AC, Davidson JR, Jefferson JW, Janney CA, Katzelnick DJ,
Weisler RH, Greist JH, Sutherland SM. 1999. Treatment of social
phobia with gabapentin: a placebo-controlled study. J Clin
Schneier, Johnson J, Hornig C D, Liebowitz MR, Weissman MM.
1992. Social phobia. Comorbidity and morbidity in an epidemio-
logic sample. Arch of Gen Psychiatry 49:282–288.
Sheehan DV. 1983. The anxiety disease. New York: Charles Scribner
& Sons. p.148–149.
Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller
E, Hegueta T, Baker R, Dunbar GC. 1998. The Mini-Interna-
tional Neuropsychiatric Interview (MINI): the development and
validation of a structured diagnostic psychiatric interview. J Clin
Psychiatry 59 (suppl 20):22–57.
Stein MB, Chartier MJ, Hazen AL, Kroft CD, Chale RA, Cote D,
Walker JR. 1996. Paroxetine in the treatment of generalized social
phobia: open-label treatment and double-blind placebo-controlled
discontinuation. J Clin Psychopharmacol 16: 218–222.
Stein MB, Liebowitz MR, Lydiard RB Pitts CD, Bushnell W, Gergel I.
1998. Paroxetine treatment of generalised social phobia (social anxi-
ety disorder). A randomised controlled trial. JAMA 280:708–713.
van Vliet IM, den Boer JA, Westenberg HG. 1994. Psychopharma-
cological treatment of social phobia; a double blind placebo con-
trolled study with fluvoxamine. Psychopharmacology (Berl).
Watson P, Friend R. 1969. Measurement of social-evaluative anxiety.
J Consulting Clin Psychol 33:448–457.
Weiller E, Bisserbe JC, Boyer P, Lepine JP, Lecrubier Y. 1996. Social
phobia in general health care: an unrecognised undertreated dis-
abling disorder. Br J Psychiatry 168:169–174.
Wittchen HU, Beloch E. 1996. The impact of social phobia on qual-
ity of life. Int Clin Psychopharmacol 11 (suppl 3):15–23.
Wittchen HU, Nelson CB, Kessler RC. 1999. Social fears and DSM-IV
social phobia in a community sample of adolescents and young adults:
prevalence, risk factors and comorbidity. Psychol Med 29:309–329.