Chlamydial disease in adolescent women is a serious public health problem, but secondary preventive efforts through early detection and treatment are encumbered by the cost and complexity of mass screening. This study was undertaken to identify clinical predictors of infection that might narrow the scope of screening adolescent populations. Demographic/clinical data and endocervical smears for the direct-specimen fluorescein-conjugated monoclonal antibody test for Chlamydia trachomatis were collected from 244 consecutive women, 21 years of age or less, attending an adolescent health clinic. Positive direct-specimen fluorescein-conjugated monoclonal antibody test for C trachomatis results were associated with a past history of chlamydial infection, multiple sexual partners, sexual contact with men with urethritis, nonuse of condoms, metrorrhagia, exocervicitis, mucopurulent endocervical discharge, abnormal cervical cytologic features, and isolation of Neisseria gonorrhoeae from the endocervix. These variables were entered into a discriminant analysis to predict direct-specimen fluorescein-conjugated monoclonal antibody test for C trachomatis results. The discriminant function was statistically significant but explained only 17% of between-group variance. Two variables alone, exocervicitis and partners with urethritis, correctly predicted direct-specimen fluorescein-conjugated monoclonal antibody test for C trachomatis results in 79% of all cases (negative predictive value 90%; positive predictive value 35%). When routine screening with reliable laboratory tests is not feasible, selective testing or empirical treatment of adolescent women with either risk factor may be cost-effective alternatives.