OBJECTIVE To compare a rapid, office-based test with standard cell culture for screening of women for Chlamydia trachomatis infections. METHODS An 8-month prospective crossover trial used alternating screening protocols in two Baltimore (Md) sexually transmitted disease clinics from January 2 through August 14, 1991. METHODS Consecutive women attending the two clinics who had no indication for administration of antichlamydial antibiotic therapy (eg, history of recent sexual contact with a partner with a sexually transmitted disease, mucopurulent cervicitis, pelvic inflammatory disease, known gonorrhea, or previously diagnosed Chlamydia infections). METHODS Chlamydia screening was offered according to one of two protocols. Use of the two screening protocols was alternated between clinics each month. In the "rapid test clinic," eligible women were screened with both a 30-minute enzyme immunoassay test and tissue culture. Patients screened with the rapid test were asked to remain in the clinic until their rapid assay results were available so that, if positive, the patients could be treated. In the "routine screening clinic," eligible women were screened for Chlamydia by cell culture. Women identified as being infected with Chlamydia by screening culture were later confidentially notified of their test results by health department disease intervention specialists and referred for therapy. METHODS Performance of screening tests for bringing infected patients to therapy; time intervals between initial clinic visits and therapy; and pelvic inflammatory disease occurring between initial visits and therapy. RESULTS Chlamydia cultures were positive in 100 (6.6%) of 1526 women screened with the solid-phase immunoassay, 47 of which were detected and treated on the basis of rapid test results. In contrast, 93 (74%) of 126 women with positive screening cultures returned to the clinic and received therapy. The median interval between testing and therapy for women with positive screening cultures was 14 days, and three (3.2%) developed pelvic inflammatory disease in the interval between testing and return for therapy. CONCLUSIONS Neither cell culture nor a rapid diagnostic test performed well for ensuring therapy of women with Chlamydia infections. The sensitivity of the rapid diagnostic test was low, and nearly one fourth of the women with positive screening cultures did not return for therapy. Evaluation of screening for Chlamydia should consider the utility of strategies for bringing patients to treatment, as well as the more usual measures of test performance, such as sensitivity, specificity, and predictive values.