The interrater agreement was calculated with respect to the total score. Cohen’s kappa was 0.916 with a cut-off of 8.

Data Analysis

In the interviewed sample, a stepwise regression analysis was conducted using the DIB-R scores as the criterion and the items from the STCPD as predictors.

Results

Of the 107 subjects, 22 (21%) were diagnosed with BPD (4 boys, 18 girls; mean age 16.7 years, SD 1.7; range 15 to 20). The regression equation obtained with the stepwise regression analysis was applied to the STCPD items, and a cut-off value of 8.315 was identified in the interviewed sample. It divided this sample into 2 categories containing 24 (score higher than the cut-off) and 83 (score lower than the cut-off) adolescents, respectively.

The frequency of BPD in the entire sample was obtained by applying the identified cut-off value of 8.315. It divided the entire sample into 2 categories containing 173 and 1029 adolescents, respectively. Table 1 shows the distribution of these 173 adolescents as a function of age and sex.

Discussion

To our knowledge, this study is the first to use a validated structured diagnostic interview to assess BPD prevalence in a community sample of adolescents.

This study proposes a novel method for assessing the prevalence of mental disorders. Its advantage is that a time-consuming structured interview can be used in a limited subsample, and a questionnaire in a larger sample. The appropriate cut-off for determining the pathology most relevant to the studied population was identified based on regression analysis.

We estimated overall BPD frequency to be 14%, which is close to that found by Bernstein and others (10.8%). This comparison is legitimate because a reasonably high correspondence has been reported in adolescents between the DIB and the DSM-III and its revised form (12).

The fact that the observed frequency in the interviewed sample was greater than the inferred frequency in the whole sample may indicate that adolescents with psychological problems used the interview to obtain attention and support from a psychologist. The interviewed sample could thus be seen as intermediate between a clinical and a community sample. A consequence is that distributions of STCPD and DIB-R scores are closer to a normal distribution and more appropriate for determining a precise cut-off.

The overall BPD frequency was 10% for boys and 18% for girls. The 95% CIs were 3.1% to 23.3% for boys and 5.7% to 33% for girls. These findings are in keeping with the preponderance of girls in clinical samples (1,2). This female preponderance was lesser in the study by Bernstein and others, which found a prevalence of 10.1% for boys and 11.5% for girls (7). Another difference is their observation that BPD prevalence peaked in early adolescence: they found nearly twice as many cases between ages 11 and 14 years as they found between ages 15 and 17 years.

Our study also observed a peak of frequency at age 14 years, but we found that, after a decline, the frequency increased in late adolescence. It should be noted, however, that the number of subjects in each age group is insufficient to determine a precise frequency; these results have to be confirmed in a larger sample.

The high frequency of BPD found in this study, contrasted with a much lower prevalence in adults, further challenges the validity of diagnosing this disorder in adolescents and the validity of diagnostic criteria for this age group. Given the prevalence of “borderline” symptoms in adolescents who may not have a BPD, further studies are necessary to evaluate the possibility of narrowing the diagnostic criteria for this age group.

References

1. Ludolph PS, Westen D, Misle B, Jackson A, Wixom J, Wiss FC. The borderline diagnosis in adolescents: symptoms and developmental history. Am J Psychiatry 1990;147:470–6.
2. Grilo CM, Becker DF, Fehon DC, Walker ML, Edell WS, McGlashan TH. Gender differences in personality disorders in psychiatrically hospitalized adolescents. Am J Psychiatry 1996;153:1089–91.
3. Garnet KE, Levy KN, Jonathan JF, Mattanah JJF, Edell WS, McGlashan TH. Borderline personality disorder in adolescents: ubiquitous or specific? Am J Psychiatry 1994;151:1380–2.
4. Mattanah JJF, Becker DF, Levy KN, Edell WS, MacGlashan TH. Diagnosis stability in adolescents followed up 2 years after hospitalization. Am J Psychiatry 1995;152:889–94.
5. Becker DF, Grilo CM, Edell WS, McGlashan TH. Comorbidity of borderline personality disorder with other personality disorders in hospitalized adolescents and adults. Am J Psychiatry 2000;157:2011–6.
6. Johnson JG, Cohen P, Skodol AE, Oldham JM, Kasen S, Brook JS. Personality disorders in adolescence and risk of mental disorders and suicidality during adulthood. Arch Gen Psychiatry 1995;56:805–11.
7. Bernstein DP, Cohen P, Velez CN, Schawb-Stone M, Siever LJ, Shinsato L. Prevalence and stability of the DSM-III-R personality disorders in a community-based survey of adolescents. Am J Psychiatry 1993;150:1237–43.
8. Dowson JH. Assessment of DSM-III-R personality disorders by self-report questionnaire: the role of informants and a screening test for co-morbid personality disorders (STCPD). Br J Psychiatry 1992;161:344–52.
9. Zanarini MC, Gunderson JG, Franckenburg FR, Chauncey DL. The revised diagnostic interview for borderlines: discriminating BPD from others axis II disorders. J Pers Disord 1989;3:10–8.
10. Yannchyshyn G, Kutcher S, Cohen C. The diagnostic interview for borderlines : reliability and validity in adolescents. J Am Acad Child Adolesc Psychiatry 1986;25:427–9.
11. Atlas JA, Postelnek J. Applicability of the revised diagnostic interview for borderlines to hospitalized adolescents. Psychological Reports 1994;74:297–8. 12. Meijer M, Goedhart AW, Treffers, PDA. The persistence of borderline personality disorder in adolescence. J Pers Disord 1998;12:13–22.