How reliable is the clinical exam in detecting ocular chlamydial infection?

PURPOSE To describe the relationship between the clinical exam for trachoma and the polymerase chain reaction (PCR) for ocular chlamydia. METHODS One hundred children in a trachoma-endemic area of Ethiopia were examined three times and swabbed twice for PCR analysis. The assays were compared, and an analysis of the variance between exam and PCR was performed. RESULTS Inter-examiner agreement was 0.57 (Cohen's K), inter-PCR agreement 0.98, and agreement between examiner and PCR, 0.26-0.34. The positive predictive value of the exam in identifying infection was 66%. Inter-examiner variance accounted for 30% of the total variance between the exam and PCR, with the remainder presumably due to an underlying difference in what the exam and PCR measure. CONCLUSIONS Despite modest inter-grader reliability and correlation with evidence of infection, the clinical exam is widely used due to its convenience and low cost. Efforts to make laboratory tests for ocular Chlamydia trachomatis more affordable would be useful.