See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/5763420

Social anxiety disorder above and below the diagnostic threshold: Prevalence,
comorbidity and impairment in the general population

Article in Social Psychiatry and Psychiatric Epidemiology · May 2008

DOI: 10.1007/s00127-007-0299-4 · Source: PubMed

CITATIONS READS

298 2,783

4 authors, including:

Lydia Fehm Frank Jacobi

Humboldt-Universität zu Berlin Psychologische Hochschule Berlin
111 PUBLICATIONS 4,399 CITATIONS 237 PUBLICATIONS 17,855 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Frank Jacobi on 05 June 2014.

The user has requested enhancement of the downloaded file.

Soc Psychiatry Psychiatr Epidemiol (2008) 43:257–265 DOI 10.1007/s00127-007-0299-4

ORIGINAL PAPER

Lydia Fehm Æ Katja Beesdo Æ Frank Jacobi Æ Agnes Fiedler

Social anxiety disorder above and below the diagnostic

threshold: prevalence, comorbidity and impairment in the
general population

Received: 18 January 2007 / Accepted: 26 November 2007 / Published online: 14 December 2007

j Abstract Background There is a lack of data j Key words epidemiology – prevalence – social
systematically describing subthreshold expressions of anxiety disorder – social anxiety-impairment
social anxiety disorder (SAD) with regard to preva-
lence, comorbidity, and impairment. Methods This
analysis was based on data from the German Health
Survey (GHS) and its Mental Health Supplement Introduction
(GHS-MHS). Social anxiety disorder and its syn-
dromes as well as other mental disorders were as- Social anxiety disorder (SAD) is one of the most fre-
sessed with a standardized diagnostic interview quent anxiety disorders (see [16] for an European
(M-CIDI) in 4,174 adults. Results The 12-month overview; [2] for a recent overview; [23, 24] for results
prevalence rate for threshold SAD was 2.0%, sub- from a representative US sample). The core feature of
threshold and symptomatic social anxiety (one DSM- SAD is a marked and persistent fear of one or more
IV criterion missing/two or more criteria missing) social and performance situations and the fear to act
was found in 3.0 and 7.5% of the participants, in a humiliating or embarrassing way, as described by
respectively. As expected, threshold SAD was char- criterion A in the Diagnostic and Statistical Manual of
acterized by an elevated risk for comorbid disorders Mental Disorders (DSM-IV) [1]. Further criteria in-
and associated with impairment in diverse areas of clude that the feared situations almost invariably in-
life. However, this was also true for the two sub- duce anxiety (criterion B), that the affected person
threshold expressions of social anxiety, which were recognizes the anxiety as being excessive or unrea-
also significantly associated with higher comorbidity sonable (criterion C), that the feared situations are
and greater impairment compared to the control avoided, or endured with intense distress (criterion
group. Conclusions Our results suggest that social D) and that the anxiety or its concomitants interfere
anxiety below the diagnostic threshold is clearly with everyday functioning, or elicit marked distress
associated with adverse outcomes. Prospective de- (criterion E). Additional criteria indicate a minimum
signs should examine the exact temporal and possible duration of six months for individuals under age 18
causal pathways of this burden in order to inform (criterion F) and specify exclusions, such as substance
prevention and early intervention programs. effects and other mental disorders (criterion G) or
general medical conditions that may elicit similar
anxiety symptoms (criterion H).
Women are more frequently affected by SAD than
men [9, 29, 35, 50]. In clinical samples, usually no
Dr. L. Fehm (&) gender differences appear, which has led to the
Institute of Psychology
Humboldt-Unversität zu Berlin
assumption that SAD more strongly interferes with
Rudower Chaussee 18 daily functioning in men than in women [38, 44].
12489 Berlin, Germany It is commonly agreed that the onset of SAD occurs
Tel.: +49-30/2093-9309 during adolescence and early adulthood [11, 12, 15,
Fax: +49-30/2093-9351 35, 50]. The stability of the disorder has yet to be
E-Mail: [email protected]
determined as retrospective designs and clinical
K. Beesdo Æ F. Jacobi Æ A. Fiedler samples usually report a chronic course with low
SPPE 299

Institute of Clinical Psychology and Psychotherapy

Technische Universität Dresden recovery rates, e.g., [8]. In contrast, prospective lon-
Dresden, Germany gitudinal studies in epidemiological samples suggest
258

that the stability on the full diagnostic level is rather Respondents of the German Health Survey (GHS) older than
low [12, 28, 31]. 65 years were excluded because the psychometric properties of the
interview used in the study have not yet been satisfactorily estab-
Social anxiety disorder is associated with a higher lished for use in older populations [25]. The eligible sample size for
risk for a variety of other disorders, primarily other the GHS-MHS was N = 4,773. The conditional response rate of the
anxiety disorders, affective disorders and substance- GHS-MHS was 87.6%, resulting in a total of 4,181 respondents who
related disorders (see [17] for an overview). Especially completed the mental health assessment. For seven individuals, the
information in the SAD section was incomplete, which resulted in a
for depressive disorders, SAD is temporally primary final sample of 4,174 (weighted N = 4,179). The presented results
in the majority of cases [18, 21, 34, 35, 39, 43]. In can be regarded as representative for the German non-institu-
addition, there is first evidence for a potential causal tionalized adult population from 18 to 65 years of age.
relationship between SAD and the development and a
more malignant course of comorbid depressive dis- j Assessment
orders [4, 6, 37].
Psychopathological and diagnostic assessments were based on the
Social anxiety disorder has been shown to be computer-assisted version of the Munich Composite International
associated with marked reduction in quality of life, Diagnostic Interview (M-CIDI) [49], a modified version of the
not only in social, but also in educational and occu- World Health Organization CIDI (version 2.1) [41] for a wider
pational domains [13, 26, 27, 30, 32, 36, 48, 50]. range of mental disorders according to the criteria of DSM-IV than
in previous studies. Psychometric properties of the CIDI were
All current knowledge outlined above has been found to be acceptable to very good [33, 45, 46]. Unlike previous
gathered with regard to threshold SAD, that is, indi- versions of the M-CIDI, the study version focuses strictly on the
viduals fulfilling all diagnostic criteria. Only one assessment of 12-month symptoms and disorders. The standard M-
earlier study, relying on DSM-III criteria, examined CIDI lifetime assessment was only performed when lifetime infor-
subthreshold expressions of SAD [10]. The authors mation was necessary for evaluating current diagnoses (e.g., mood
disorders).
found that subthreshold cases with SAD reported The M-CIDI SAD section begins with a stem question to assess
considerable impairments and disabilities and sug- the presence of strong fears regarding seven social and perfor-
gested to include these cases in the prevalence rate for mance situations (DSM-IV criterion A-1). After at least one fear
SAD. The present study seeks to extend these findings situation was elicited, a subsequent series of nine questions asked
about cognitive elements of anxiety such as fear of humiliating or
in a larger sample with more differentiated levels of embarrassing occurrences, fear of blushing, fear of panic, fear of
subthreshold expressions. showing symptoms of anxiety, etc. (criterion A-2). Criterion B was
Thus, using data from a nationally representative assessed by a list of anxiety symptoms (e.g., sweating, heart racing,
study, the aims of this paper are (1) to describe etc.) of which at least two were required to occur when thinking
about, or when being exposed to such situations. Respondents
prevalence and comorbidity patterns of both DSM-IV further indicated whether they considered either the anxiety or the
SAD and its subthreshold expressions, and (2) to avoidance to be excessive or unreasonable (criterion C), and
examine in detail associated impairments and dis- whether they frequently avoided the situations or, if not, endured
abilities. the situations with distress (criterion D). Criterion E was assessed
by determining whether the respondent reported that the social
fears or avoidance interfered a lot with normal routines, or whether
they sought professional help for the fears.
Besides threshold SAD (all DSM-IV criteria are met) we in-
Methods cluded two conditions characterized by social anxiety, but not
fulfilling all DSM-IV criteria: Individuals in the subthreshold group
j Design and sample met criterion A, but had one other criterion missing. Consistent
with operationalizations of other studies [50], individuals in the
Mental disorders were assessed in the Mental Health Supplement of symptomatic group reported strong social fears, but did not com-
the 1998/99 German National Health Interview and Examination plete two or more of the DSM-IV criteria. Table 1 gives an overview
Survey (GHS-MHS) [5, 19, 20, 47]. Its sample was a stratified of the nature of the missing criteria.
random sample from 113 communities throughout Germany with With regard to onset and course of the social fears, participants
130 sampling units [sampling steps: (1) selection of communities, were asked to retrospectively remember their age at the first epi-
(2) selection of sampling units, and (3) selection of inhabitants sode of the condition. Duration of social anxiety was calculated as
from population registries]. the age at interview minus age of onset.

Table 1 Number of individuals in

the three diagnostic groups failing to Lack of criterion SAD
meet the DSM-IV criteria for social
anxiety disorder (SAD) Symptomatic Subthreshold Threshold
(N = 314) (N = 127) (N = 83)

n % n % n %

A-1 fear of social/performance situation 0 0.00 0 0.00 0 0.00

A-2 fear of humiliation/embarrassment 125 39.80 0 0.00 0 0.00
B anxiety reaction 167 53.23 10 7.61 0 0.00
C excessive or unreasonable 240 76.43 22 18.24 0 0.00
D avoidance or endurance with intense anxiety/distress 232 73.84 53 41.41 0 0.00
E impairment/distress 274 87.36 42 32.74 0 0.00

n and % weighted
 259

Comorbidity with other mental disorders was determined with screening status in order to address different sampling probabilities
regard to the following (threshold) 12-month diagnoses: other and systematic non-response [19]. In the following, we only report
anxiety disorders (panic disorder with and without agoraphobia, weighted Ns and percentages. Logistic regression [odds ratios (OR)
agoraphobia without panic attacks, specific phobias, generalized and 95% CI] were used to quantify the associations between SAD
anxiety disorder, obsessive–compulsive disorder, phobias not conditions and other mental disorders. Mean differences for the
otherwise specified), mood disorders (major depressive disorder), impairment measures over different groups were calculated with
dysthymia, bipolar disorders, dependence from nicotine or alcohol, mean ratios (MR) (from gamma regression, with 95% CI).
eating disorders (anorexia nervosa, bulimia nervosa and eating
disorder not otherwise specified), and an aggregate group of so-
matoform disorders (somatization disorder, hypochondriasis,
undifferentiated somatoform disorder, somatic symptom index 4.6 Results
[14], pain disorder).
Impairment in different life domains was assessed in four ways: j Prevalence of SAD, gender and age group
reduced health related quality of life (physical and mental health
sum scores of the SF-36, [7]; clinical complaints (sum score of a list differences
of psycho-vegetative symptoms mostly related to depression and
anxiety (‘‘Beschwerdeliste’’) [40]; satisfaction in several life domains The total 12-month prevalence rate for DSM-IV SAD
(rated on 7-point-Likert scales), and self reported disability days was 2.0% (Table 2, upper part). Women had an about
within the past 12 months (‘‘How many days within last 12 months twofold risk to develop the disorder as compared to
have you been too sick to carry out usual activities?’’). Health care
utilization is presented in form of the variable ‘‘at least minimal men (OR = 2.1).
intervention’’, a combination of items asking about having ever There seemed to be a decrease in prevalence rates
sought treatment due to psychological, mental, addictive, or psy- in older age groups, however, this decline did not
chosomatic problems, or the recommendation by a doctor to do so. reach statistical significance.
The 12-month prevalence rates for subthreshold
j Statistical analyses and symptomatic SAD were 3.0 and 7.5%, respec-
Prevalence estimates for threshold, subthreshold and symptomatic tively. Women were significantly more frequently af-
SAD [N, %; 95% confidence intervals (CI) available on request] fected by subthreshold SAD (total sample: OR = 1.8;
were calculated with the data weighted for age, gender, region and age 18–29: OR = 4.4; age 50–65: OR = 2.6) compared
Table 2 Twelve-month prevalence
rates of DSM-IV SAD and its Age group Threshold SAD (DSM-IV)
subthreshold expressions (N = 4,179;
men: n = 2,101; women: n = 2,078) Total Men Women Ref. group: men

n % n % n % OR 95% CI

18–65 83 2.0 28 1.3 55 2.7 2.1 1.3–3.5

18–29 23 2.6 9 2.1 14 3.2 1.6 0.6–4.1
30–39 18 1.7 4 0.8 14 2.6 3.5 1.0–12.5
40–49 19 2.1 5 1.2 13 3.0 2.6 0.9–7.5
50–65 23 1.8 9 1.4 14 2.2 1.7 0.7–4.0

Age group Subthreshold SAD

Total Men Women Ref. group: men

n % n % n % OR 95% CI

18–65 127 3.0 47 2.2 80 3.9 1.8 1.2–2.7

18–29 36 4.1 7 1.6 29 6.6 4.4 1.9–10.4
30–39 36 3.3 17 3.1 19 3.5 1.2 0.6–2.4
40–49 30 3.4 15 3.3 15 3.4 1.0 0.5–2.2
50–65 25 1.9 7 1.1 18 2.7 2.6 1.1–6.2

Age group Symptomatic SAD

Total Men Women Ref. group: men

n % n % n % OR 95% CI

18–65 314 7.5 149 7.1 165 7.9 1.2 0.9–1.5

18–29 115 13.1 60 13.5 55 12.7 1.0 0.6–1.6
30–39 82 7.6 39 6.9 43 8.3 1.2 0.8–1.9
40–49 61 6.8 32 7.1 29 6.4 0.9 0.5–1.6
50–65 56 4.2 18 2.8 38 5.6 2.1 1.2–3.6
Threshold: all DSM-IV criteria are met; subthreshold: criterion A and all but one DSM-IV criteria are met; symptomatic:
individuals report strong social fears, but have at least two DSM-IV criteria missing. n numbers (weighted); % percentages
(weighted), OR odds ratio from multinomial logistic regression
260

to men. For symptomatic social anxiety, this gender SAD as the reference group, but also with symptom-
difference only occurred in the oldest age group (age atic and subthreshold SAD serving as the reference.
50–65: OR = 2.1). Using no SAD as the reference group, not only
In the subthreshold as well as in the symptomatic DSM-IV SAD (OR = 22.2), but also subthreshold
condition, the decline of prevalence with higher age (OR = 10.3) and even symptomatic (OR = 3.4)
reached statistical significance (subthreshold SAD: expressions of this condition were associated with
age 50–65 vs. 18–29: OR = 0.4; 95% CI: 0.2–0.7; age other anxiety disorders, in particular with panic dis-
50–65 vs. 30–39: OR = 0.6; 95% CI: 0.3–0.9; age 50–65 order, agoraphobia, generalized anxiety disorder, and
vs. 40–49: OR = 0.5; 95% CI: 0.3–0.97; symptomatic obsessive–compulsive disorder (see Table 3). A dose–
SAD: age 18–29 vs. age 30–39: OR = 1.9, 95% CI: 1.3– response relationship could be observed for each of
2.6; age 50–65 vs. 30–39: OR = 0.5; 95% CI: 0.4–0.7; the comorbid anxiety conditions under study as
age 50–65 vs. 40–49: OR = 0.6; 95% CI: 0.4–0.9). indicated by increasing ORs from symptomatic SAD
towards the DSM-IV diagnosis. This observation was
supported by analyses using symptomatic and sub-
j Comorbidity with other mental disorders threshold SAD as the reference group, respectively:
DSM-IV SAD was not only associated with other
The vast majority of individuals with DSM-IV SAD anxiety disorders compared to individuals with no
(N = 73/83, 87.8%) had a diagnosis of at least one SAD (OR = 22.2), but also in comparison to indi-
other mental disorder during the past 12-month viduals with symptomatic (OR = 6.6) or subthreshold
period. Among those, 20% had one, 20% two, and (OR = 2.2) expressions of this condition. Similarly,
60% three or more comorbid conditions. Pure cases subthreshold SAD was associated with other anxiety
were younger than comorbid cases (33.5 vs. disorders even when using symptomatic SAD as the
40.5 years), but this difference failed to reach statis- reference group (OR = 3.1).
tical significance (P = 0.065). In the majority of Any mood disorder (including major depression,
comorbid cases, SAD preceded the comorbid disor- dysthymia, bipolar disorders) was strongly associ-
ders (total: 66%; men: 57.5%, women: 70.7%). ated with SAD as well (OR range 3.3–19.7) (see Ta-
To investigate the comorbidity patterns of the ble 4). Likewise, there were significant risk elevations
different diagnostic expressions of SAD, Tables 3 and for somatoform disorders (OR ranging between 1.9
4 show the proportions of other anxiety disorders and and 4.4) in all three levels of SAD that are again
other mental disorders among individuals with most pronounced for threshold SAD. Furthermore, a
symptomatic, subthreshold, threshold SAD and no range of differences in proportions of comorbid
SAD. In addition, associations were not only calcu- conditions was significant between the three SAD
lated for each of these diagnostic categories with no groups, again indicating a dose–response relation-
Table 3 Comorbid anxiety disorders conditional on SAD status (N = 4,179)

Comorbid anxiety Reference group No SAD Symptomatic SAD Subthreshold SAD Threshold SAD
disorders (n = 3,655) (n = 314) (n = 127) (n = 83)

% % OR 95% CI % OR 95% CI % OR 95% CI

Any other anxiety disorder No SAD 10.1 27.3 3.4 2.6–4.1 53.4 10.3 7.1–15.0 71.2 22.2 13.0–37.9
Symptomatic SAD 3.1 1.9–4.7 6.6 3.7–11.8
Subthreshold SAD 2.2 1.1–4.1
Panic disorder No SAD 1.3 2.9 2.3 1.1–4.7 16.4 15.1 8.6–26.2 25.6 26.4 15.1–46.3
Symptomatic SAD 6.5 2.9–14.7 11.5 5.1–26.0
Subthreshold SAD 1.7 0.9–2.3
Agoraphobia without panic No SAD 0.9 4.8 5.1 2.8–9.2 14.6 17.3 9.6–31.3 17.0 20.7 11.0–38.7
Symptomatic SAD 3.4 1.7–6.9 4.1 1.9–8.6
Subthreshold SAD 1.2 0.6–2.5
Any specific phobia No SAD 5.9 14.3 2.6 1.9–3.7 24.5 5.1 3.3–7.8 29.5 6.6 4.1–10.7
Symptomatic SAD 1.9 1.2–3.2 2.5 1.4–4.4
Subthreshold SAD 1.4 0.7–2.4
GAD No SAD 0.8 2.7 3.5 1.6–7.3 5.9 7.9 3.5–17.5 22.2 35.4 18.5–67.9
Symptomatic SAD 2.3 0.9–5.9 10.3 4.5–23.4
Subthreshold SAD 4.5 1.9–10.9
OCD No SAD 0.3 0.9 3.1 1.0–9.9 4.8 15.9 5.9–42.4 11.5 41.5 17.7–97.2
Symptomatic SAD 5.1 1.4–18.3 13.2 4.0–43.9
Subthreshold SAD 2.6 0.9–7.2
Phobia NOS No SAD 2.6 8.8 3.7 2.4–5.6 9.9 4.2 2.3–7.7 10.7 4.6 2.2–9.1
Symptomatic SAD 1.1 0.6–2.2 1.2 0.6–2.6
Subthreshold SAD 1.1 0.5–2.6

GAD generalized anxiety disorder, OCD obsessive–compulsive disorder, NOS not otherwise specified. n numbers (weighted); % percentages (weighted); OR odds ratio
from multinomial logistic regressions
 261

Table 4 Comorbid other mental disorders conditional on SAD status (N = 4,179)

Comorbid disorders Reference group No SAD Symptomatic SAD Subthreshold SAD Threshold SAD
(n = 3655) (n = 314) (n = 127) (n = 83)

% % OR 95% CI % OR 95% CI % OR 95% CI

Any mood disorder No SAD 8.7 24.1 3.3 2.5–4.5 39.4 6.8 4.6–9.9 65.3 19.7 12.0–32.1
Symptomatic SAD 2.0 1.3–3.2 5.9 3.4–10.2
Subthreshold SAD 2.9 1.6–5.3
MDD No SAD 6.0 18.6 3.6 2.5–5.0 21.3 4.2 2.7–6.5 50.5 15.9 10.0–25.4
Symptomatic SAD 1.2 0.7–1.9 4.5 2.6–7.7
Subthreshold SAD 3.8 2.1–7.1
Dysthymia No SAD 3.0 6.9 2.5 1.5–3.9 20.9 8.7 5.4–13.9 38.1 20.2 12.4–32.9
Symptomatic SAD 3.5 1.9–6.5 8.2 4.4–15.3
Subthreshold SAD 2.3 1.2–4.3
Any bipolar disordera No SAD 0.5 1.2 2.4 0.9–6.1 5.8 11.9 4.9–29.1 5.7 12.6 4.1–33.6
Symptomatic SAD 4.9 1.6–14.8 4.8 1.4–16.6
Subthreshold SAD 0.9 0.3–3.3
Any somatoform disorderb No SAD 9.3 18.7 2.2 1.6–3.1 28.0 3.8 2.5–5.7 31.3 4.4 2.7–7.2
Symptomatic SAD 1.7 1.04–2.7 1.9 1.1–3.4
Subthreshold SAD 1.2 0.6–2.1
Any eating disorderc No SAD 0.3 0.8 3.2 1.1–9.7 2.1 8.4 1.8–39.2 0.0 –
Symptomatic SAD 2.6 0.5–13.9
Subthreshold SAD
Nicotine dependence No SAD 9.1 15.8 1.9 1.3–2.6 21.9 2.8 1.8–4.4 24.0 3.2 1.8–5.4
Symptomatic SAD 1.5 0.9–2.6 1.7 0.9–3.1
Subthreshold SAD 1.1 0.5–2.1
Alcohol dependence No SAD 2.8 4.6 1.7 0.9–3.0 9.8 3.7 1.9–7.1 10.3 3.9 1.8–8.5
Symptomatic SAD 2.2 1.0–5.0 2.4 0.9–5.8
Subthreshold SAD 1.1 0.4–2.7

MDD major depressive disorder

n numbers (weighted), % percentages (weighted); OR odds ratio from multinomial logistic regressions
a
Bipolar-I-disorder, bipolar-II-disorder
b
Somatization disorder, hypochondriasis, undifferentiated somatoform disorder, somatoform pain disorder
c
Anorexia nervosa, bulimia nervosa, eating disorder not otherwise specified

ship for at least affective disorders and somatoform associated with a significantly reduced quality of life
disorders. For eating disorders, nicotine dependence in various domains, e.g., reduced mental health
and alcohol dependence, the associations with sub- (measured by the SF-36 mental health sum scale),
threshold expressions of SAD were weaker or not clinical complaints (‘‘Beschwerdeliste’’), and satisfac-
significant. tion in different life domains, e.g., family, social
relations, work situation and financial situation, but
not with self-perceived physical health (SF-36, phys-
j Age of onset and course ical health sum scale) (see Table 5). The self-reported
As a retrospective self-report measure, a mean dura- number of disability days during the past year was
tion of 16.1 years (SD = 14.5) was found for threshold threefold among individuals with SAD compared to
and 18.6 years (SD = 13.7) for subthreshold SAD (for those without SAD (35 vs. 12 days; MR: 2.9). Inter-
symptomatic SAD, this information was not available, estingly, the differences in impairments and disabili-
as this part of the interview was skipped due to the ties did not only occur in threshold SAD, but also in
interviews’ rules). Men and women did not differ with subthreshold and even symptomatic expressions of
regard to duration (threshold: men M = 16.1, SD = this condition with no SAD as the comparison group.
16.3, women M = 15.8, SD = 13.0; subthreshold: men In a further analysis we explored whether different
M = 21.3, SD = 13.5, women M = 17.1, SD = 13.7). periods of onset of SAD are associated with different
Among individuals with threshold SAD, most degrees of impairment and disability by comparing
participants reported an age of onset during adoles- three groups of individuals with threshold SAD: those
cence and early adulthood, however, 17.8% (n = 15/ with an age of onset of SAD before the 25th year of life
83) reported an onset after their 40th year of life. (‘‘early onset’’; serving as the reference group), those
with an onset between ages 25 and 40, and those with
an onset during or after the fifth decade of life (‘‘late
j Impairments and disabilities onset’’). Among the different indicators of quality of
life (see Table 5) only two significant comparisons
As SAD is known to affect more than only social emerged: the ‘‘late onset’’ group had significantly
domains of life, we examined different facets of lower scores for physical health compared to the
quality of life. Social anxiety disorder was found to be reference group (MR = 0.9; 95% CI 0.7–0.97), and
262

Table 5 Impairment in different life domains, differentiated by diagnostic level of SAD (N = 4,175)

No SAD Symptomatic SAD Subthreshold SAD Threshold SAD

(n = 3655) (n = 314) (n = 127) (n = 83)
(ref. group)

M SD M SD MR 95% CI M SD MR 95% CI M SD MR 95% CI

SF-36: mental healtha 51.5 8.1 46.3 9.3 0.89 0.88–0.92 42.2 11.1 0.82 0.78–0.86 36.9 13.0 0.72 0.66–0.78
SF-36: physical healtha 49.3 8.8 49.5 8.5 1.00 0.98–1.03 47.5 10.2 0.97 0.93–1.00 47.4 9.8 0.96 0.92–1.00
Clinical complaints (Zerssen scoreb)c 15.9 10.5 20.4 10.8 1.28 1.20–1.37 24.5 13.8 1.54 1.38–1.71 29.5 12.6 1.85 1.66–2.06
Satisfaction in several life domainsa
Overall 5.6 1.1 5.3 1.2 0.94 0.91–0.97 4.7 1.4 0.83 0.78–0.87 4.5 1.6 0.79 0.73–0.86
Family 5.8 1.4 5.5 1.6 0.95 0.92–0.98 5.1 1.7 0.87 0.82–0.92 4.8 2.0 0.82 0.75–0.90
Social relations 5.9 1.1 5.7 1.3 0.97 0.94–1.00 5.2 1.5 0.89 0.84–0.93 5.1 1.5 0.87 0.81–0.93
Work situations 5.2 1.6 4.8 1.7 0.93 0.89–0.97 4.6 1.7 0.86 0.80–0.92 4.1 1.8 0.79 0.71–0.87
Financial situation 4.9 1.6 4.4 1.7 0.94 0.89–0.98 4.1 1.9 0.86 0.79–0.94 4.1 2.1 0.77 0.69–0.86
Disability days 12.1 37.0 10.7 22.5 0.88 0.69–1.11 25.9 62.8 2.13 1.39–3.23 34.9 74.6 2.88 1.86–4.44
(last 12 month, self report)

MR mean ratios from negative binominal regression and 95% confidence intervals (CI); reference group: no social phobia; bold: P < 0.05
a
Higher scores indicate better health/higher satisfaction
b
‘‘Beschwerdeliste’’
c
Higher scores indicate more complaints

they were more satisfied with their social relations assess social anxiety of a significant degree only (‘‘Did
(MR = 1.2; 95% CI 1.02–1.4). you feel strong anxiety in the following situations, or
did you avoid these situations…?’’, ‘‘Do you think you
j Help-seeking behavior experience this fear or avoid these situations much
stronger than other people?’’).
Sixty-five percent of all individuals with threshold While the majority of SAD cases expectedly re-
DSM-IV SAD (N = 54/83) indicated to have accessed ported an onset of social anxiety in adolescence or
professional mental health services at least once. For early adulthood, there seems to be a small portion of
other mental disorders apart from SAD this rate was cases with threshold SAD with a later onset after the
36.3% (N = 435/1,301). Individuals with comorbid fourth decade of life. As the results have not been
SAD had a more than threefold risk of seeking ser- adjusted for individuals having not yet reached the
vices compared to those cases without other mental age of 40, our findings might underestimate the ac-
disorders (OR = 3.26, 95% CI 1.99–5.36). However, as tual number of late-onset cases. On the other hand,
health care use was not assessed with regard to spe- due to the restriction on 12-month prevalence it may
cific symptoms or disorders, no statements about be that individuals experienced previous episodes of
disorder-related treatment rates or the adequacy of SAD which were inaccurately recalled and therefore
the treatment were available. possibly resulted in shifts of dating the first age of
onset towards later ages. Among the questions
determining age of onset, participants were asked to
Discussion indicate their very first episode in life, but we cannot
rule out completely whether there were episodes that
Our study addressed SAD in a large unselected sample were not reported—either voluntarily or involun-
not only on the full DSM-IV diagnostic level, but also on tarily. Another reason for a possible misdating of
a syndrome level including symptomatic and sub- onsets might be the broad age range of the sample
threshold expressions of this condition with regard to that should be kept in mind when interpreting ret-
patterns of 12-month prevalence, comorbidity and rospective age-of-onset reports. However, due to lack
quality of life. of basic epidemiological data on SAD according to
The prevalence of threshold SAD in our sample DSM-IV criteria we still felt that it would be
was 2.0%, which corresponds to the median preva- important to report age-of-onset findings among our
lence in a recently published review for the European 12-month SAD cases despite the methodological
countries [16]. Subthreshold and symptomatic limitations and the need for replication of findings
expressions of SAD were found in 3.0 and 7.5% of the in more appropriate samples (studies beginning in
participants, respectively. One would have expected childhood or adolescence) and study designs (pro-
social anxiety to be more prevalent on a broad syn- spective-longitudinal studies).
drome level, as social anxiety is usually regarded as Regarding comorbidity, our results are consistent
being normally distributed in the general population. with findings of other studies [23, 30] in that SAD is
Our results might be explained by the wording of the a highly comorbid disorder. Even within the 12-
respective questions in the interview that intended to month time frame that we used in our study, we
 263

found a very low portion of only 12.2% pure cases. j Strengths and limitations
However, our findings expand former knowledge on
comorbidity patterns in SAD by showing that A major strength of the study lies in the unselected
comorbidity is lower, but still significant among representative population sample. This is important,
subjects with subthreshold and symptomatic because clinical samples are subject to diverse selection
expressions of this condition. biases (e.g., due to symptom severity, varying inclu-
Moreover, SAD seems to precede comorbid dis- sion/exclusion of patients in treatment facilities) and
orders in the majority of cases. It thus may function represent only a small portion of all individuals affected
as a potential causal risk factor for the development of by the condition. In contrast, unselected, representa-
comorbid disorders and/or their more malignant tive samples allow broader inferences on a general
course as suggested by recent analyses of a prospec- population level and are of special interest for the
tive-longitudinal community study among adoles- determination of the prevalence and the actual burden
cence and young adults on the example of depression of the disorder. It could be speculated that only more
[4, 37]. The temporal relation between comorbid impaired cases seek treatment and that clinical samples
conditions and SAD has to be more closely investi- represent the upper end of a continuum of impairment
gated in future research in order to further investigate by SAD. A direct comparison of indicators of impair-
whether an increased risk for the secondary onset of ment and disability in non-clinical and clinical samples
comorbid conditions can be found irrespective of age would allow for testing this assumption.
or rather in specific age or time frames, whether SAD The use of a well-established diagnostic interview
can be considered a causal risk factor for other con- is a further strength of the study but since the SAD
ditions, and if so, what the causal determinants of assessment focused on the previous 12 months,
such risk associations might be. The answer to these analyses covering longer periods of time are limited
questions would have considerable clinical impact (e.g., with regard to the temporal patterns of comor-
with regard to targeting and timing early screening bid disorders, or the course of social anxiety over
and early intervention as well as on prevention. This time). It is possible that some of the symptomatic and
is especially of relevance as we could show that social subthreshold cases have met criteria for threshold
anxiety is associated with reduced health related diagnosis in the past (>1 year), but due to the re-
quality of life, reduced satisfaction in several life do- stricted diagnostic time frame this cannot be further
mains, more clinical complaints and more disability investigated (as well as comparisons with completely
days. Also, help seeking (lifetime contact with mental remitted cases).
health professionals) is increased compared with
other mental disorders. However, other studies sug-
gest that individuals with SAD usually consult mental Conclusion
health services not until many years after the onset of
their disorder. For example Wang et al. [42] reported Taken together our results point to an enormous
that only 3.4% of individuals of SAD make treatment burden posed on the many individuals suffering from
contact during the first year after onset of this con- SAD as well as its subthreshold expressions. As SAD is
dition. The discrepancies between the findings might likely to precede other disorders, the early detection
be explained by the fact that our help-seeking and treatment of SAD is of utmost importance.
assessment was not specifically targeted on SAD but
rather on mental health in general. Therefore, it is
likely that the help seeking is influenced by the Note
presence of comorbid conditions revealing overall
higher rates. Data of this study are available as Public Use File
Lastly, special attention should be given to the from Dr. Frank Jacobi (manual and variable
strong impairments reported by subthreshold and description in German language): Dr. Frank Jacobi,
symptomatic SAD cases. At first glance this seems to Institute of Clinical Psychology and Psychotherapy,
contradict the fact that in many subthreshold cases Chemnitzer Str. 46, 01187 Dresden, Germany;
the DSM-criterion of subjective impairment was e-mail: [email protected]
missing. It has to be differentiated in future analyses, For further information about the Core Survey
whether the impairment among these cases can be (GHS-CS) and its Public Use File contact the Robert
traced back to comorbid disorders or if indeed the Koch-Institute, Dr. Heribert Stolzenberg, Nordufer 20,
impairment criterion in the DSM classification has to D-13353 Berlin, Germany; e-mail: [email protected]
be challenged critically (for a discussion see [3, 22]).
Either way, syndromes of SAD below the diagnostic
threshold seem to be sensitive indicators of psycho- j Acknowledgments The authors thank Hans-Ulrich Wittchen as
well as two anonymous reviewers for their helpful comments on
pathology, impairment and disability—a fact with earlier versions of the manuscript. Reported data on mental dis-
important clinical implications in regard to diagnos- orders were assessed in the Mental Health Supplement of the
tics, intervention and prevention. German Health Survey (GHS-MHS), conducted by the Max-Planck-
264

Institute of Psychiatry, Munich. Principal investigator was Hans- 21. Kessler RC et al (1999) Lifetime co-morbidities between social
Ulrich Wittchen. Reported sociodemographic and somatic health phobia and mood disorders in the US National Comorbidity
status variables come from the Core Survey (GHS-CS), conducted Survey. Psychol Med 29:555–567
by the Robert Koch-Institute, Berlin. Principal investigators of the 22. Kessler RC et al (2003) Mild disorders should not be eliminated
GHS-CS were Dr. Bärbel-Maria Kurth and Dr Wolfgang Thefeld. from the DSM-V. Arch Gen Psychiatry 60(11):1117–1122
23. Kessler RC et al (2005) Prevalence, severity, and comorbidity of
12-month DSM-IV disorders in the National Comorbidity
Survey Replication. Arch Gen Psychiatry 62(6):617–627
24. Kessler RC et al (2005) Lifetime prevalence and age-of-onset
References distributions’ of DSM-IV disorders in the national comorbidity
survey replication. Arch Gen Psychiatry 62(6):593–602
1. APA (1994) Diagnostic and statistical manual of mental dis- 25. Knauper B, Wittchen HU (1994) Diagnosing major depression
orders, 4th edn. American Psychiatric Association, Washing- in the elderly—evidence for response bias in standardized
ton, DC diagnostic interviews. J Psychiatr Res 28(2):147–164
2. Baumeister H, Harter M (2007) Prevalence of mental disorders 26. Lochner C et al (2003) Quality of life in anxiety disorders: a
based on general population surveys. Soc Psychiatry Psychiatr comparison of obsessive–compulsive disorder, social anxiety
Epidemiol 42(7):537–546 disorder, and panic disorder. Psychopathology 36(5):255–262
3. Beals J et al (2004) Challenges in operationalizing the DSM-IV 27. Mendlowicz MV, Stein MB (2000) Quality of life in individuals
clinical significance criterion. Arch Gen Psychiatry with anxiety disorders. Am J Psychiatry 157(5):669–682
61(12):1197–1207 28. Merikangas KR et al (2002) The spectrum of social phobia in
4. Beesdo K et al (2007) Incidence of social anxiety disorder and the Zurich Cohort Study of young adults. Biol Psychiatry
the consistent risk for secondary depression in the first three 51(1):81–91
decades of life. Arch Gen Psychiatry 64(8):1–10 29. Meyer C et al (2000) Lebenszeitprävalenz psychischer Störun-
5. Bellach B-M, Knopf H, Thefeld W (1998) Der Bundes-Ges- gen in der erwachsenen Allgemeinbevölkerung - Ergebnisse der
undheitssurvey 1997/1998. Das Gesundheitswesen 60:59–68 TACOS-Studie [Lifetime prevalence of mental disorders in the
6. Bittner A et al (2004) What characteristics of primary anxiety adult general population: Findings from the TACOS study].
disorders predict subsequent major depressive disorder? J Clin Nervenarzt 71:535–542
Psychiatry 65(5):618–626 30. Meyer C et al (2004) Impact of psychiatric disorders in the
7. Brazier JE et al (1992) Validating the SF-36 health survey general population: satisfaction with life and the influence of
questionnaire: new outcome measure for primary care. Br Med comorbidity and disorder duration. Soc Psychiatry Psychiatr
J 305:160–164 Epidemiol 39:435–441
8. Bruce SE et al (2005) Influence of psychiatric comorbidity on 31. Müller N (2002) Die soziale Angststörung bei Jugendlichen und
recovery and recurrence in generalized anxiety disorder, social jungen Erwachsenen: Erscheinungsformen, Verlauf und Kon-
phobia, and panic disorder: a 12-year prospective study. Am J sequenzen, Münster, Waxmann
Psychiatry 162(6):1179–1187 32. Quilty LC et al (2003) Quality of life and the anxiety disorders.
9. Davidson JRT et al (1993) The epidemiology of social phobia: Anxiety Disord 17:405–426
findings from the Duke Epidemiological Catchment Area Study. 33. Reed V et al (1998) To what degree the Composite International
Psychol Med 23:709–718 Diagnostic Interview (CIDI) correctly identifies DSM-IV dis-
10. Davidson JRT et al (1994) The boundary of social pho- orders? Testing validity issues in a clinical sample. Int J
bia—exploring the threshold. Arch Gen Psychiatry 51(12):975– Methods Psychiatr Res 7:142–155
983 34. Schatzberg AF et al (1998) McLean Hospital Depression Re-
11. DeWit DJ et al (1999) Antecedents of the risk of recovery from search Facility: early-onset phobic disorders and adult-onset
DSM-III-R social phobia. Psychol Med 29:569–582 major depression. Br J Psychiatry 173(suppl 34):29–34
12. Degonda M, Angst J (1993) The Zurich Study XX. Social phobia 35. Schneier FR et al (1992) Social phobia: comorbidity and mor-
and agoraphobia. Eur Arch Psychiatry Clin Neurosci 243:95–102 bidity in an epidemiologic sample. Arch Gen Psychiatry
13. Eng W et al (2005) Domains of life satisfaction in social anxiety 49:282–288
disorder: relation to symptoms and response to cognitive- 36. Schneier FR (2003) Social anxiety disorder—is common, un-
behavioral therapy. J Anxiety Disord 19:143–156 derdiagnosed, impairing, and treatable. Br Med J
14. Escobar JI et al (1989) Somatic symptom index (SSI)—a new 327(7414):515–516
and abridged somatization construct—prevalence and epide- 37. Stein MB et al (2001) Social anxiety disorder and the risk of
miological correlates in 2 large community samples. J Nerv depression: a prospective community study of adolescents and
Ment Dis 177(3):140–146 young adults. Arch Gen Psychiatry 58:251–256
15. Faravelli C et al (2000) Epidemiology of social phobia: a clinical 38. Turk CL et al (1998) An investigation of gender differences in
approach. Eur Psychiatry 15:17–24 social phobia. J Anxiety Disord 12:209–223
16. Fehm L et al (2005) Size and burden of social phobia in Europe. 39. Van Ameringen M et al (1991) Relationships of social phobia
Eur Neuropsychopharmacol 15:453–462 with other psychiatric illness. J Affect Disord 21:93–99
17. Fehm L, Wittchen H-U (2004) Comorbidity in social anxiety 40. Von Zerssen D, Koeller DM (1976) B-L Beschwerden-Liste.
disorder. In: Bandelow B, Stein D (eds) Social anxiety disorder. Klinische Selbstbeurteilungs-Skalen (KSB-S) aus dem Münch-
Dekker, New York, pp 49–63 ener Psychiatrischen Informations-System (PSYCHIS Mün-
18. de Graaf R et al (2003) Temporal sequencing of lifetime mood chen). Beltz, Weinheim
disorders in relation to comorbid anxiety and substance use 41. WHO (1997) Composite International Diagnostic Interview
disorders. Findings from the Netherlands Mental Health Survey (CIDI, Version 2.1). World Health Organization, Geneva
and Incidence Study. Soc Psychiatry Psychiatr Epidemiol 42. Wang PS et al (2005) Failure and delay in initial treatment
38:p1–p11 contact after first onset of mental disorders in the national
19. Jacobi F et al (2002) Estimating the prevalence of mental and comorbidity survey replication. Arch Gen Psychiatry
somatic disorders in the community: aims and methods of the 62(6):603–613
German National Health Interview and Examination Survey. Int 43. Weiller E et al (1996) Social phobia in general health care: an
J Methods Psychiatr Res 11(1):1–18 unrecognised undertreated disabling disorder. Br J Psychiatry
20. Jacobi F et al (2004) Prevalence, comorbidity and correlates of 168:169–174
mental disorders in the general population: results from the 44. Weinstock LS (1999) Gender differences in the presentation
German Health Interview and Examination Survey (GHS). and management of social anxiety disorder. J Clin Psychiatry
Psychol Med 34(4):597–611 60(suppl 9):9–13
 265

45. Wittchen HU (1994) Reliability and validity studies of the 49. Wittchen H-U, Pfister H (eds) (1997) DIA-X-Interviews: Man-
WHO-Composite International Diagnostic Interview (CIDI): a ual für Screening-Verfahren und Interview; Interviewheft für
critical review. Int J Psychiatr Res 28:57–84 Längsschnittuntersuchung (DIA-X-Lifetime); Ergänzungsheft
46. Wittchen HU et al (1998) Test–retest reliability of the com- (DIA-X-Lifetime); Interviewheft Querschnittsuntersuchung
puterized DSM-IV version of the Munich-Composite Interna- (DIA-X-Monate); Ergänzungsheft (DIA-X-Monate); PC-Pro-
tional Diagnostic Interview (M-CIDI). Soc Psychiatry Psychiatr gramm zur Durchführung des Interviews (Längs- und Quer-
Epidemiol 33(11):568–578 schnittsuntersuchung); Auswertungsprogramm. Frankfurt,
47. Wittchen H-U et al (1999) Affektive, somatoforme und Angs- Swets and Zeitlinger
tstörungen in Deutschland - Erste Ergebnisse des bundesweiten 50. Wittchen H-U, Stein MB, Kessler RC (1999) Social fears and
Zusatzsurveys ‘‘Psychische Störungen’’. Gesundheitswesen social phobia in a community sample of adolescents and young
Sonderheft 2:S216–S222 adults: prevalence, risk factors and co-morbidity. Psychol Med
48. Wittchen H-U et al (2000) Disability and quality of life in pure 29:309–323
and comorbid social phobia. Findings from a controlled study.
Eur Psychiatry 15(1):46–58

View publication stats