Validity of self-reported sexual behaviors in adolescent women using biomarker outcomes.

Abstract

To determine the criterion-related validity of alternative approaches to the measurement of sexual intercourse using sexually transmitted diseases (STD) as a biomarker (the criterion). Analyses are based on an urban sample of 255 adolescent women, 15 to 19 years of age, treated for genitourinary infections with Neisseria gonorrhoeae, Chlamydia trachomatis or Trichomonas vaginalis who returned 3 months later for reexamination. Subjects completed self-administered questionnaires at enrollment and at 3 months. Genitourinary cultures were obtained at enrollment, 2 to 4 weeks later at a test-of-treatment visit (TOT), and at 3 months. Two hundred fourteen of the 255 adolescents also returned for a TOT culture; 186 of these 214 (73%) were free of infection 2 to 4 weeks after enrollment and 30% (56/186) acquired a subsequent STD by 3 months. The validity of questions about sexual behavior differed. No adolescent who denied interim intercourse by reporting "0" sexual partners or "0" coitions acquired an interval STD. Adolescents who denied regular intercourse (vaginal sex) or failed to indicate the number of interim coitions were at high risk for new STD-23% and 21%, respectively. A new measure of sexual intercourse using both the number of sexual partners and the number of coitions contained no missing data; adolescents classified as not having had interim sexual intercourse were free of infection at 3 months, whereas 32% of those who reported intercourse acquired an interim infection. These data suggest that high-risk urban adolescent women can accurately report whether they have engaged in vaginal intercourse. The validity of the report appears sensitive to the wording and content of the questions.