Clinical algorithms for the screening of women for gonococcal and chlamydial infection: evaluation of pregnant women and prostitutes in Zaire.

A substantial proportion of women with gonococcal and/or chlamydial infection are asymptomatic. Thus active case detection is problematical, particularly in developing countries, where facilities and materials for laboratory testing are limited. We assessed the diagnostic validity of the hierarchical clinical algorithms recommended by the World Health Organization as well as that of a nonhierarchical scoring system, using data for 1,160 pregnant women (a low-prevalence group) and 1,222 prostitutes (a high-prevalence group) in Kinshasa, Zaire. Neisseria gonorrhoeae and/or Chlamydia trachomatis was detected in 6.5% and 31.0% of pregnant women and prostitutes, respectively. No single variable that was both sensitive (> 60%) and specific (> 60%) was associated with infection. A simple hierarchical algorithm based only on reported symptoms had a sensitivity of 48.0% and 54.9% and a specificity of 75.2% and 52.2% for the screening of pregnant women and prostitutes, respectively. A second algorithm that included a speculum examination had a sensitivity of only 29.3% but a specificity of 85.3% in pregnant women. When a nonhierarchical scoring system was used, the sensitivity was 72.0% and 71.0% and the specificity was 73.5% and 55.8% for pregnant women and prostitutes, respectively. Scoring systems that incorporate risk markers as well as symptoms and signs may represent affordable alternative methods of screening for gonococcal and/or chlamydial infections among women in resource-poor settings.