European Child & Adolescent Psychiatry

, Volume 15, Issue 3, pp 163–171

Social anxiety disorder in 11–12-year-old children

The efficacy of screening and issues in parent–child agreement

Authors

    • Regional Centre for Child and Adolescent Mental HealthRBUP
  • Svenn Torgersen
    • Dept. of PsychologyUniversity of Oslo
ORIGINAL CONTRIBUTION

DOI: 10.1007/s00787-005-0519-y

Cite this article as:
Kristensen, H. & Torgersen, S. Eur Child Adolesc Psychiatry (2006) 15: 163. doi:10.1007/s00787-005-0519-y

Abstract

Objective

To investigate the level of diagnostic and discriminative accuracy of The Social Anxiety Scale for Children – Revised (SASC-R) for identifying social anxiety disorder (SAD) in a community-based sample of 11–12 year-old children. Parent–child diagnostic agreement was also examined.

Method

A questionnaire including SASC-R and items on impulsive behavior was sent to a population based sample of children, born in 1992. A total of 2568 parents returned their questionnaires (rr: 70%), and 1297 (51%) consented to further participation. An index group [50 high-scoring children on social anxiety (SA-group)] and two contrast groups [(50 high-scoring children on impulsive behavior (Imp-group) and 50 low-scoring children on SA and Imp, (Ls-group)] were selected for participation.

Results

SAD was assigned 35 (23%) of 150 children; 28 (80%) from the SA-group; 7 (20%) from the Imp- group and 0 from the Ls-group. The SASC-R showed relatively high discriminative accuracy for SAD, but was also influenced by other diagnoses. Mother–child agreement was fair (κ=0.46), and mother-only diagnoses were frequent. Both child and parent information are important when diagnosing SAD in this age group.

Key words

social anxiety disorderchildrenscreening

Introduction

Social fears are normal human reactions that in many cases may represent adaptive feelings in social situations [32]. In contrast, social anxiety disorder, SAD, is characterized by intense fear of embarrassment and negative evaluation from others in social situations and by a tendency to avoid such situations [2]. SAD is a common disorder (reported lifetime prevalence of 13% in adults) [20, 21] that may lead to significant stress and impairment [24]. SAD in adults has found to be strongly associated with other anxiety disorders as well as depression and substance abuse in both clinical and community samples [26].

In most cases the onset of SAD is reported to occur in the mid-adolescence and to precede the onset of other psychiatric disorders [28]. However, age of onset is usually based on retrospective data and may be misleading [15]. SAD has been diagnosed in children as young as 7–8-years-old [3], and many adults with SAD report experiencing lifelong shyness [31].

The high prevalence, the significant impairment and comorbidity, and the assumed childhood onset all emphasize the importance of early diagnosis and treatment of SAD. Self-report rating scales may represent the first step in case identification. Studies examining the diagnostic accuracy of dimensional anxiety rating scales usually find that they are most successful in discriminating anxious children from children without psychopathology. Their ability to discriminate between different diagnostic groups have been inconsistent, but indicate a better discrimination between anxiety and externalizing disorders, than between anxiety and depression or between the various anxiety disorders [33].

During the last decade rating scales for measuring social anxiety in children have successfully been developed [29]. Studies including such scales with community samples have found that a large proportion of children (8–12 years) exceed the recommended cut-offs for clinical levels of social anxiety [10, 27], but the relationship between scale scores and diagnostic status has not been explored in this age group. However, a recent study examined this relationship in a sample of socially anxious adolescents using two different scales, the Social Anxiety Scale for Adolescents (the SAS-A) [23] and the Social Phobia and Anxiety Inventory for Children (the SPAI-C) [4, 16]. The sensitivity and specificity for the SAS-A was 44% and 83% and for the SPAI-C 62.5% and 83%.

Parent–child agreement is an important issue when diagnosing childhood SAD. Informant concordance in anxiety disorders in general is reported to be poor with a trend for children to endorse fewer symptoms than their parents [33]. When there is discrepancy between parents and child as informants, a diagnosis is most frequently made on the basis of the parents’ report [30]. However, for most diagnostic categories, including anxiety disorders, both parent-only and child-only diagnoses seem to represent meaningful clinical conditions [18].

Type of disorder, age and gender of child and parental psychopathology represent modulating factors with regard to parent–child agreement [13]. Social concerns and fear of negative evaluation are core symptoms of SAD, and therefore the children may tend to underreport their social anxiety symptoms. However, studies on parent–child agreement that have included a measure of social desirability have demonstrated conflicting results [7, 13]. With regard to age, some studies report a higher parent–child agreement on SAD in older children (>9 years) [30] whereas other studies have failed to replicate significant differences between age groups [5, 13]. The data on gender have shown a higher parent–child agreement on SAD in boys compared with girls [14] and vice versa [5]. Finally, in general, parental psychopathology is suggested to make them more accurate reporters, but also to result in overreporting of symptoms [13].

The major aim in the present study was to examine the level of diagnostic and discriminative accuracy of The Social Anxiety Scale for Children – Revised (SASC-R) [23] in identifying children with SAD in a population based sample of 11–12 year old children. The relationship between a SAD diagnosis and self- and observer rating of state anxiety was also explored. Furthermore, we wanted to examine parent–child agreement concerning the presence of a SAD diagnosis in this age group and the influence of maternal psychopathology on parent–child agreement. Diagnostic interviews were conducted in three screened groups, an index group of socially anxious children (SA-group) and two contrast groups, one with impulsive children (Imp-group) and one with children with low scores on both social anxiety and impulsivity (Ls-group).

Method

Participants

The screened sample in this study consists of 150 children aged 11–12 years; 75 girls and 75 boys. The children had participated in a cohort study on asthma bronchiale launched in 2001. This 2001 cohort (n = 7266) consisted of all children who were born in 1992 and lived in Oslo in 2001 (n = 5285) plus children born in Oslo in 1992 who had participated in a previous asthma birth cohort, but who were living outside Oslo in 2001 (n = 1981). A total of 4831 (response rate 66%) agreed to participate and of these 3653 (75%) also agreed to be sent invitations for later research projects. (Fig. 1)
https://static-content.springer.com/image/art%3A10.1007%2Fs00787-005-0519-y/MediaObjects/787_2005_519_f1.gif
Fig. 1

Recruitment of the participants in the study

In 2003, 3642 of 3653 families received a postal questionnaire on their children’s physical and mental health. Included in this questionnaire were SASC-R [23], selected items on impulsive behavior [6, 12] and an invitation to join an additional study on social anxiety in childhood. The reason for the inclusion of impulsivity items was twofold: To ensure a comparison group with problem behavior in addition to a presumably more well-functioning group and to contrast anxiety disorders with disruptive disorders. A total of 2568 families (rr = 70%) returned the questionnaires, and 1297 families (50.5%) wanted to participate in the social anxiety study. The difference in mean SASC-R total score between those who consented (n = 1297) and those who declined (n = 1271) were: 35.9 (10.6) vs. 35.0 (9.6) (F = 9.49, p = 0.002), respectively. The study was approved by the Regional Ethics Committee for Medical Research.

The second author selected participants in the three screening groups on the basis of the parental scale scores on the SASC-R and the selected impulsivity items. The aim was to have 25 children of each gender in the respective groups. The social anxiety score and the impulsivity score were transformed to z-scores. Those with the highest scores were selected for the SA-group and the Imp-group. As some of the candidates were excluded or declined to participate (n = 42), those with the next highest score were chosen. Children who had high scores both on social anxiety and impulsivity were not excluded, and in such cases, the highest z-score determined which group they were assigned to. The selection of 25 girls and 25 boys to the Ls-group was based on the lowest z-scores on the SASC-R scale and the impulsivity scale combined. Of the 42 selected families who had to be replaced, a total of 23 families were excluded. Exclusion criteria included language problems (both parents from outside Europe) (n = 16); mild mental retardation (n = 3); twins (n = 2); adopted child (n = 2). Another 19 families did not want to participate when contacted due to various reasons such as illness in family, house move etc.

The mean age of the children at assessment was 11.6 years (0.37). The mean total SASC-R scores for the SA-group were: 61.0 (6.8); the Imp-group: 41.9 (9.9) and the Ls-group: 20.7 (2.8), respectively. A total of 61% of the children lived with both their biological parents. The ethnic composition included both parents Norwegian or Nordic (82%); one parent Norwegian/one European (10%); one parent Norwegian /one non European (8%). Educational attainment categories were: mothers (n = 150): without H.S. diploma: 2 %; H.S. graduate: 39.3%; some college education: 38.0 %; degree from 4-year college or more: 20.7%; fathers (n = 148): without H.S. diploma: 7.3%; H.S. graduate: 34.7%; some college education: 31.3%, degree from 4 years of college or more: 25.3%. Income mothers (n = 150) included: Below 200,000 NK (19.3%); between 200,000 and 500,000 NK (72%) and higher than 500,000 NK (8.7%). Income fathers (n = 150) in the corresponding categories were 16.7%; 56.0% and 27.3%, respectively.

Screening measures

SASC-R

The Social Anxiety Scales for Children Revised, SASC-R, [22, 23] is a self-report inventory (child and parent version) designed to assess children’s subjective feelings of social anxiety and the behavioral consequences of the social anxiety. The instrument consists of 22 items (four of the items are filler items) that are rated on a 5-point scale from 1 (not at all) to 5 (all the time).

The instrument contains three subscales; Fear of Negative Evaluation (FNE; eight items), Social Avoidance and Distress specific to New Situations (SAD-New; six items), and General Social Avoidance and Distress (SAD-G; four items). Subscale scores are calculated by summing-up ratings for the items included in the particular subscale. The SASC-R total score is the sum of the three subscale scores. Scores may range from 18 to 90. Recommended clinical cut-off scores are 52 for boys and 54 for girls [23].

Acceptable internal consistency for the subscales has been reported both for the child version (alphas: 0.86–0.90; 0.74–0.79 and 0.60–0.70) [10, 11, 22] and the parent version (0.91; 0.87; 0.78) [23]. Substantial test–retest reliability and concurrent validity have also been demonstrated [22, 27]. In the current study the Cronbach’s alpha coefficients for the subscales were: 0.96, 0.95 and 0.90, respectively.

ITSC and SDQ

The Infant–Toddler Symptom Checklist (ITSC) [6] had been used in two different Norwegian studies of children with restlessness and ADHD (unpublished). The ITSC was developed for children who experience problems of self-regulation. Each item on the checklist is rated on a 5-point scale from “never” to “always”. Five items that appeared to discriminate successfully between restless children/children with ADHD and controls, were selected. In addition, the Strength and Difficulties Questionnaire (SDQ) was applied [12]. The SDQ measures prosocial behavior and psychopathology. The items are rated as “not true”, “somewhat true”, and “certainly true”. Among the SDQ items, the six items that correlated most strongly with the five ITSC items were selected. Together the 11 items cover attention problems, restlessness, demanding behavior, stimuli sensitiveness, disobedience and aggression. The Cronbach’s Alpha was 0.78.

Assessment measures

Kiddie-SADS P/L

The Kiddie- SADS P/L is a semistructured interview providing DSM-IV Axis I child psychiatric (present and lifetime) diagnoses [19]. The instrument does not assess pervasive developmental disorders (PDD) or Axis II disorders. The diagnoses are scored as definite, probable (equal or greater to ¾ of symptom criteria met), or not present. The K-SADS P/L has a child (6–18 years) and a parent version. Usually the interviewer starts with the parent as the informant, then administers the K-SADS to the child and finally makes consensus scores on the basis on all available information. Studies have reported sufficient interrater reliability and test–retest reliability [1, 19]. To test interrater reliability in the present study, 20 audiotaped interviews (mother and child combined) were chosen at random and scored independently by an experienced child psychiatrist. Interrater reliability for the most prevalent lifetime consensus diagnoses were: SAD : 0.81; specific phobia: 0.81; any depressive disorder: 0.81; ADHD: 0.91 and any disruptive disorder: 0.91.

ASSQ

The Autism Spectrum Screening Questionnaire (ASSQ) is a 27-item checklist for the assessment of Asperger’s disorder and other high-functioning autism spectrum disorders. Items are scored as does not apply, applies to some extent and definitely applies. Total scores range from 0 to 54. The cut-off score for a diagnosis in the present study was 22. The instrument is found to be a valid and reliable screening measure with good to excellent sensitivity and specificity for autism spectrum disorders [9].

Mini-International-Neuropsychiatric-Interview (M.I.N.I.)

The M.I.N.I. [34] is a brief structured diagnostic interview for adults screening 17 axis-1 disorders for current and lifetime diagnoses according to both DSM-IV [2] and ICD-10 [35]. The validity and reliability of the M.I.N.I are reported to be satisfactory [25, 34].

State-anxiety measure

An observational and a self-report measure of state anxiety were included as a validation for a SAD diagnosis. The overall present level of anxiousness was scored on a five-point Likert-type scale: 1: not anxious at all; 2: a little anxious; 3. somewhat anxious; 4: rather anxious; 5: very anxious. The rating was administrated upon arrival to evaluate the presumably most stressful moment of the assessment for children with social anxiety. The children were asked to score how anxious they felt at the moment. The observer’s score was rated from a global impression of the child’s social behavior (i.e. hiding behind the mother, gaze aversion, sweaty palms, muteness).

Assessment procedure

The selected families were contacted by phone and the direct assessment was carried out at an outpatient clinic (n = 63) or in the child’s home (n = 87) 4–12 months after the screening. The first author interviewed the children while maternal interviews were performed simultaneously by research assistants (two experienced educational therapists and one master level student in psychology). All raters were blind to screening group status. Kiddie-SADS was administered to the mother (by the first author) 1–3 weeks after the assessment of the child. If the interview with the child gave any indication that the child might qualify for a diagnosis of Asperger’s disorder, an ASSQ interview was added to the parental version of Kiddie-SADS. Diagnostic questions on selective mutism according to DSM-IV were added to the social anxiety questions in the Kiddie SADS in all cases.

Statistics

Chi-square analysis or the Fisher exact test was used to examine differences between groups for categorical data. For continuous data, independent t-tests or ANOVAs were used. Multiple regression analyses were performed to examine the independent contribution of selected diagnoses to the SASC-R total score. Agreement between pairs of raters is described using the κ statistic for multiple raters. All p values are calculated as two-tailed.

Results

Demographics

The three screening groups did not differ significantly with regard to ethnicity or family income. However, children in the Ls-group more often lived with both their parents compared to both the SA-group and the Imp-group (Ls: n = 38; SA: n = 28; Imp: n = 25, χ2 = 7.77, df = 2, p < 0.05). Furthermore the fathers in the Imp-group reported less education than the other two groups (Degree from 4 years of college or more: Ls: n = 35; SA: n = 30; Imp: n = 20; χ2 = 9.05, df = 2, p < 0.05).

Rates of diagnoses

The rates of present and lifetime diagnoses (Kiddie-SADS) by selected group and gender are presented in Table 1 (consensus) and Table 2 (child as informant). While the prevalence of the categories “any disorder” and “any anxiety disorder” was high in both the SA-group and the Imp-group, SAD was the only single diagnostic category that discriminated the SA-group from both of the other contrast groups. This was also the case when the child was the informant. Furthermore, with the child as the informant, there were significantly more boys than girls fulfilling the criteria for SAD (boys: n = 14 (78%), girls. n = 4 (22%), χ2 = 6.3, p < 0.01).
Table 1

Rates (%) of present (P) and lifetime (L) consensus diagnoses (DSM-IV) by selection group (SA = social anxious; Imp = impulsive) and gender (F = female; M = male)

Diagnoses

High screen SA

High screen Imp

Low screen SA/Imp

Fn = 25

Mn = 25

Totaln = 50

Fn = 25

Mn = 25

Totaln = 50

Fn = 25

Mn = 25

Totaln = 50

(P) Any depressive dis

4

2

12

6

(L) Any depressive disb, f

28

40

34

16

40

28

12

6

(P) Separation anxiety disa, e

16

12

14

8

20

14

(L) Separation anxiety disb, f

36

32

34

20

28

24

4

2

(P) Panic dis

4

2

(L) Panic dis

4

2

4

2

(P) Specific phobiab, f

36

48

42

44

36

40

4

2

(L) Specific phobiab, f

36

52

44

44

36

40

8

8

8

(P) Social anxiety disb, d

52

56

54

4

24

14

(L) Social anxiety disb, d

52

60

56

4

24

14

(P) Obsessive-compulsive dis

8

8

8

4

2

4

2

(L) Obsessive-compulsive dis

12

8

10

4

4

4

4

8

6

(P) Posttraumatic stress dis

(L) Posttraumatic stress dis

4

2

(P) Generalized anxiety dis

16

12

14

8

12

10

4

2

(L) Generalized anxiety disb

20

20

20

8

12

10

4

2

(P) Adjustment dis w/anxiety

(L) Adjustment dis w/anxiety

4

2

(P) Any anxiety disb, c, f

72

80

76

56

56

56

4

8

6

(L) Any anxiety disb, d, f

80

88

84

56

56

56

12

20

16

(P) Any enuresis/encopresis

4

8

6

4

12

8

(L) Any enuresis/encopresisb, f

24

32

28

12

32

22

(P) ADHDd, f

8

52

30

(L) ADHDd, f

16

52

34

(P) ADD

4

12

8

4

12

8

(L) ADDe

4

16

10

12

16

14

(P) Oppositional defiant disd, f

4

2

12

28

20

(L) Oppositional defiant disd, f

4

12

8

24

32

28

(P) Conduct dis

4

2

(L) Conduct dis

4

2

(P) Adjustment dis w/conduct

(L) Adjustment dis w/conduct

8

4

8

4

(P) Any ADHD/disruptive disd, f

4

16

10

20

72

46

(L) Any ADHD/disruptive disd, f

8

24

16

48

76

62

8

4

(P) Any Tourette/tic disa

16

24

20

8

20

14

8

4

(L) Any Tourette/tic disa

24

32

28

16

20

18

24

12

(P) Asperger’s dis

4

2

4

2

(L) Asperger’s dis

4

2

4

2

(P) Selective mutism

4

2

(L) Selective mutism

4

2

(P) Any disd, f

76

84

80

68

88

78

4

16

10

(L) Any disd, f

88

96

92

84

92

88

12

36

24

Note:p-values (two-tailed) based on χ2 for differences between total group rates (1df): aSA>Ls, p < 0.05; bSA>Ls, p < 0.01; cSA>Imp, p < 0.05; dSA>Imp, p < 0.01; eImp>Ls, p < 0.05; fImp>Ls, p < 0.01

Dis = Disorder; ADHD = attention deficit hyperactivity disorder; ADD = attention deficit disorder

Table 2

Rates (%) of present (P) and lifetime (L) diagnoses (DSM-IV, with child as informant) by selection group (SA = social anxious; Imp = impulsive) and gender (F = female; M = male)

Diagnoses

High screen SA

High screen Imp

Low screen SA/Imp

Fn = 25

Mn = 25

Totaln = 50

Fn = 25

Mn = 25

Totaln = 50

Fn = 25

Mn = 25

Totaln = 50

(P) Any depressive dis

(L) Any depressive dis

12

4

8

8

12

10

(P) Separation anxiety dis

8

4

8

4

6

(L) Separation anxiety dis

12

4

8

16

4

10

(P) Panic dis

4

2

––

(L) Panic dis

4

2

(P) Specific phobiab, f

28

20

24

20

16

18

(L) Specific phobiab, f

28

28

28

20

16

18

8

4

(P) Social anxiety disb, c

12

40

26

4

12

8

(L) Social anxiety disb, d

12

44

28

4

12

8

(P) Obsessive-compulsive dis

8

4

4

2

4

2

(L) Obsessive-compulsive dis

8

4

4

4

4

4

4

4

(P) Generalized anxiety dis

4

2

(L) Generalized anxiety dis

4

2

(P) Any anxiety disb, f

36

48

42

32

32

32

4

2

(L) Any anxiety disb, f

40

60

50

32

32

32

12

4

8

(P) Any enuresis/encopresis

4

2

(L) Any enuresis/encopresis

8

8

8

12

12

12

4

2

(P) ADHD

16

8

(L) ADHD

16

8

(P) Oppositional defiant dis

4

2

(L) Oppositional defiant dis

4

2

4

2

(P) Any ADHD/disruptive dis

16

8

(L) Any ADHD/disruptive dis

4

2

16

8

(P) Any Tourette/tic dis

4

2

8

4

(L) Any Tourette/tic dis

8

4

6

4

8

6

8

4

(P) Any disb, f

40

48

44

32

52

42

4

2

(L) Any disb, f

60

68

64

44

60

52

12

16

14

Note:p-values (two-tailed) based on χ2 for differences between total group rates (1df): bSA>Ls, p < 0.01; cSA>Imp, p < 0.05; dSA>Imp, p < 0.01; fImp>Ls, p < 0.01

Dis = Disorder; ADHD = attention deficit hyperactivity disorder; ADD = attention deficit disorder. Diagnostic groups with 0 prevalence are omitted

The mean state anxiety score (five-point Likert-type scale ranging from not anxious at all to very anxious) in the total sample was 2.2 (0.92) (observer’s rating) and 1.8 (0.73) (child’s rating). Both the observer’s and the child’s rating were significantly higher for the children who were assigned a current consensus diagnosis of SAD (n = 34) compared to the children without SAD (n = 116) (observer: SAD: 3.12 (0.84) vs. non-SAD: 1.88 (0.74), t = −7.73, p < 0.001; self-report: SAD: 2.18 (0.80), vs. non-SAD: 1.70 (0.68), t = −3.18, p < 0.01).

SASC-R’s specificity in diagnostic identification

To further explore the SASC-R’s ability to specifically identify SAD, a multiple regression analysis was performed. The SASC-R total score was included as the dependent variable and dichotomic diagnostic categories of lifetime diagnoses of ADHD, “any depression disorder”, separation anxiety disorder, specific phobia, SAD and general anxiety disorder were entered as independent variables. “Any present depressive disorder” (girls) and present separation anxiety disorders (both genders) were excluded from the analyses due to low prevalence. The results are presented in Table 3.
Table 3

Multiple regression analysis (enter procedure) (standardized Beta)

Diagnoses

SASC-R total score

 

Female

Male

 

Presenta

Lifetimeb

Presentc

Lifetimed

ADHD

0.01

0.05

0.19*

0.18

Any depressive disorder

0.25**

0.01

0.09

Separation anxiety disorder

0.19*

0.16

Specific phobia

0.31**

0.19

0.23*

0.20*

Social anxiety disorder

0.48***

0.43***

0.52***

0.49***

General anxiety disorder

0.04

0.02

0.16

0.11

aF = 11.37; p = 0.000; adj. R2 = 0.36

bF = 9.69; p = 0.000; adj. R2 = 0.41

cF = 12.18, p = 0.000, adj. R2 = .43

dF = 11.80; p = 0.000; adj. R2 = 0.47

*p < 0.05; **p < 0.01; ***p < 0.001; SASC-R = social anxiety scale for children – revised; ADHD = attention deficit hyperactivity disorder. Diagnoses excluded because of too low prevalence are marked with –

Parent–child agreement

Table 4 presents the parent–child agreement for SAD and composite diagnoses of anxiety disorders, depressive disorders, elimination disorders, AD(H)D/disruptive disorders and any disorders. Agreement for SAD was fair for the whole group. However, when analyzing by gender, there was a non-significant tendency for higher parent–child agreement for boys than girls. Overall the children reported markedly less social anxiety compared to the mothers’ report. Of the 35 children with SAD, a total of 17 (48.6%) was mother-only diagnoses and 5 child-only (14.3%).
Table 4

Parent–child agreement on Kiddie-SADS for social anxiety disorder and composite diagnoses, present (P) and lifetime (L)

Diagnoses

 

κ-values

  

Females

Males

Total

Social anxiety disorder

P

0.30*

0.53**

0.44**

L

0.30*

0.55**

0.46**

Any anxiety disorder

P

0.47**

0.48**

0.48**

 

L

0.54**

0.38**

0.46**

Any depresssive disorder

L

0.38**

0.23*

0.27**

Any enuresis/encopresis

L

0.38**

0.38**

0.38**

Any AD(H)D/disruptive disorder

P

 

0.24**

0.21

L

−0.03

0.18*

0.13*

Any tic disorder

P

0.32**

0.28**

0.29**

L

0.29*

0.26*

0.28**

Any disorder

P

0.42**

0.41**

0.42**

L

0.54**

0.42**

0.49**

*p < 0.01; **p < 0.001; ADHD = attention deficit hyperactivity disorder

The observer’s rating of state anxiety did not differ significantly between those who reached diagnostic level for SAD with the child as informant (n = 18) and those who were assigned a diagnosis by mothers’ report alone (n = 17). However, the state anxiety self-report was significantly higher in the group with a SAD diagnosis with the child as informant versus mother alone [2.4 (0.9) vs. 1.9 (0.5), t = 2.28, df = 26, p = 0.03] There was no difference in state anxiety scores between children with SAD assessed at home (secure setting) compared to children with SAD assessed at the out-patient clinic (novel setting).

A mother-only SAD diagnosis was significantly more frequent among mothers with a lifetime diagnosis of depression compared with mothers without depression (n = 13 (76.5%) vs. n = 4 (23.5%), χ2 = 5.04, df = 1, p < 0.03). Corresponding analyses with maternal social phobia or “any anxiety disorder” showed no significant differences. In the SAD group with depressive mothers (n = 20), the state anxiety self-report was significantly lower in the children with a mother-alone SAD diagnosis (n = 13) compared to children with a SAD diagnosis as a result of the child interview (n = 7) [1.9 (0.5) vs. 2.9 (1.1), t = 2.70; df = 18; p = 0.02]. The observer’ rating did not differ between the two groups.

Discussion

Diagnostic and discriminative accuracy

A primary goal of the present study was to investigate the diagnostic and discriminative accuracy of a social anxiety rating scale for detecting SAD in a community based sample of 11–12-year-old children. Due to the minimal time required to complete the inventory applied (10 min), the SASC-R has been recommended as a screening tool for large community samples [27]. To our knowledge this is the first study that has screened for SAD using the SASC-R followed by a diagnostic interview in a community sample in this age group. The results show that the SASC-R, was fully successful in identifying children with SAD when comparing groups with high and low social anxiety scores as rated by their parents. The lifetime prevalence of SAD in these groups was 56% and 0%, respectively. Moreover, the instrument also showed a specific effect in that the other contrast group of children (the Imp-group), screened with impulsivity items, also had significantly lower SAD prevalence than the SA-group (56% vs. 14%). This difference was found despite the inclusion of children with both high scores in impulsivity and social anxiety. In this study, the SASC-R’s efficacy in selecting children with SAD is generally in line with previous findings showing that dimensional rating scales tend to discriminate individuals with anxiety from controls and from individuals with disruptive disorders [8].

Further analysis of the SASC-R’s discriminative ability shows that in addition to being most sensitive to SAD, the instrument is also influenced by other conditions such as specific phobia in both genders, depression and separation anxiety disorder in girls and ADHD in boys. GAD was the only anxiety disorder that showed no independent predictive value on the SASC-R total score in both genders. Since this is the first study to explore the discriminative accuracy of the SASC-R in a population sample, the results must be replicated. However, considering the fact that rating scales tend to distinguish poorly between anxiety and depression and between different anxiety disorders, the result seems promising [17].

Parent and child agreement

Nearly half of the SAD diagnoses were assigned based on mothers’ report alone. This is not surprising because the children were selected for participation from the scores of the parental version of the SASC-R. Both the fact that the children with a mother-only diagnosis of SAD rated themselves less anxious, and the higher proportion of mother-only diagnoses in depressive mothers may suggest an overreport of SAD in the sample. However, the observational anxiety ratings oppose such a conclusion.

On the other hand, the underreporting of social anxiety symptoms by the children compared to both mother report and observed state anxiety may support the notion that children with SAD often want to present themselves in a socially desirable way [7]. The ability of children in this age group to communicate about inner states may also explain this finding.

Furthermore, the parent–child agreement on SAD in the current study was somewhat higher than has been reported in some previous studies [13]. The age of the children in the present study may have contributed to this result. This is in line with Rapee et al.’s [30] finding of a markedly greater parent–child diagnostic agreement for children older than 9 years. The use of the same interviewer across informants may also have increased the level of agreement.

The gender differences in our sample showed that males both reported more symptoms of social anxiety and tended to show a greater diagnostic agreement with their mothers compared to females. This result is consistent with a previous study with an outpatient sample of children and adolescents [13], but contrary to other studies reporting higher parent–child agreement in girls [5, 30]. There is no apparent explanation for these discrepancies across studies and more research to clarify this issue is needed.

Limitations

A major limitation in the current study is that the screened sample has gone through several stages, thus questioning the representativity of the final sample. The procedure may have selected well functioning families who endured participation in several studies over time. On the other hand, the screening instrument seemed successful in detecting highly socially anxious children from the relatively large sample of families who agreed to participate. The exclusion of children of immigrant families also restricts the generalizibility of the findings. Further limitations include the time span between administration of instruments and also the fact that father reports were not obtained.

Clinical implications

Overall, the high rate of SAD in this community-based sample of 11–12-year-olds confirms this condition as a childhood disorder in line with previous research suggesting an earlier onset of SAD than usually reported [3, 15]. The finding emphasizes the necessity of early detection and treatment of SAD. The high prevalence of psychopathology in the SA-group in general (also among those who did not reach diagnostic level for SAD) further supports this suggestion. Hopefully this may prevent a more serious development of the disorder. When assessing social anxiety symptoms in 11–12-year-olds, clinicians should always obtain information from both the parent and the child.

Copyright information

© Steinkopff Verlag Darmstadt 2006