Abstract

OBJECTIVE: To test the diagnostic validity of clinical algorithms for the detection of Chlamydia trachomatis in an urban population of married women in Turkey. DESIGN: Cross-sectional population-based survey. SUBJECTS: A systematic sample of 867 women who reported the use of contraceptive methods. MAIN OUTCOME MEASURES: Sensitivity, specificity and positive predictive value of clinical algorithms for the diagnosis of C trachomatis. RESULTS: C trachomatis was diagnosed in 4.89% of the women. The WHO algorithm for use in settings where no vaginal examination could be performed had a sensitivity of 9% and a specificity of 96%. The corresponding figures for the WHO algorithm incorporating the findings of a speculum examination were 47% and 56% respectively. Algorithms incorporating symptoms or signs other than those suggested by the WHO did not yield satisfactory standards of validity. CONCLUSIONS: The findings of this study do not support the widespread introduction of the use of clinical decision models for screening of women for chlamydia infection in primary health care settings such as family planning or antenatal clinics. The large number of false positive results with the use of the clinical algorithms tested in this study would cause unnecessary costs to the health system and unnecessary interventions to the women treated.