Screening guidelines recommend testing all sexually active female adolescents for Chlamydia trachomatis during a pelvic examination at each clinic visit. Such criteria have been based on cross-sectional studies; new evaluations should take into account multiple clinic visits and assess whether criteria are appropriate when a prior test is negative and risk factors are absent. Because repeated observations on an individual may be correlated, the authors used the generalized estimating equation method. Little information exists on subsequent risk of infection; as control programs develop, approaches targeting high-risk populations for recurrent infections are needed. Using data on females aged 15–19 years who visited family planning clinics more than once from 1988 to 1992 (n = 26, 921) in Region X (Alaska, Idaho, Oregon, and Washington), the authors constructed a retrospective cohort. Teens with chlamydia at their first visit were at high risk for subsequent infection (odds ratio = 1.6, 95% confidence interval 1.4–1.7). Among teens uninfected at the first visit and without risk factors at the second, prevalence at the second visit was 6%. When intervisit correlations using the generalized estimating equation method were taken into account, predictors of chlamydial infection were consistent with those in previous cross-sectional studies: cervicitis, friable cervix, and multiple, new, or symptomatic sex partner(s). These findings support screening sexually active female adolescents at each visit, even if prior tests results are available. Am J Epidemiol 1996; 144: 997-1003.