Methods for detecting microbial antigens in clinical specimens offer an alternative to culture in the diagnosis of some sexually transmitted diseases. Developers of the immunologic methods are faced with a number of problems in evaluating the new tests. Traditionally, these tests are compared to culture as the "gold standard." Unfortunately, culture for Neisseria gonorrhoeae or Chlamydia trachomatis--the two agents most commonly sought--is considerably less sensitive than 100 percent. Immunologic methods may appear to produce false positives when the paired specimens are actually false-negative cultures. Another source of discordant results is sampling variation. These considerations, however, will not account for all false-positive results. Even the best non-culture methods have a low rate of false-positive results. If a new test has a specificity of 97 percent, it, by definition, yields approximately 3 percent false-positive reactions. In low-prevalence settings this false-positive rate will create problems in interpreting the results. For example, in a population with 3 percent prevalence of infection, a positive result in a 97 percent specificity test could only have a predictive value of 50 percent. Most testing for STD agents is performed in low-prevalence settings. None of the currently available immunodiagnostic procedures has a performance profile that suggests it will be satisfactory for diagnostic use in the low-prevalence setting.