AIM--To determine when a fluorescence assay for Chlamydia trachomatis elementary bodies in the specimen buffer is of most value as a verification test for genital specimens reactive on screening enzyme immunoassay (EIA). METHOD--Genital swabs from high and medium prevalence populations were tested using EIA. Samples with absorbance values greater than the positive threshold and those within the range of 30% below this value were verified by the MicroTrak direct fluorescence assay (DFA) test. Quotients derived from the threshold value and specimen absorbances were used to establish confidence limits for the EIA. RESULTS--Of 13,283 swabs tested, 474 from the high risk group and 236 from the medium risk group were reactive on EIA and confirmed by DFA. Thirty six (5.9%) patients with confirmed reactive samples would have been missed if the kit criteria alone were followed. When confidence limits were applied to the calculated quotients, only those samples with an EIA quotient of > or = 4.0 were universally confirmed by the DFA. CONCLUSION--A scheme of testing which uses the DFA to verify EIA reactive specimens over a specified range was found to improve the sensitivity and specificity of the EIA screening test.