Background:  In the United States, confidential (name-based) positive HIV test results are reportable to an HIV case surveillance system, but anonymous (code-based) positive HIV results are not. A person’s risk behavior or other characteristics may influence whether s/he chooses an anonymous or a confidential HIV test; therefore, confidential testers may not be representative of anonymous testers. We compared HIV prevalence and incidence among people seeking anonymous and confidential HIV testing.

Methods:   Client characteristics and laboratory results were collected for people seeking voluntary HIV testing at publicly funded HIV testing sites in King County, Washington, between 2001 and 2004. We used χ² or Fisher’s exact statistics to assess associations of demographic characteristics with registration type. To calculate HIV incidence, we applied the Serological Testing Algorithm for Recent HIV Seroconversion (STARHS). In STARHS, HIV⁺ specimens are tested with a modified, less-sensitive enzyme immunoassay (LS-EIA) that is non-reactive early in the course of HIV infection; a non-reactive result indicates a probable recent HIV infection. 

Results:  Of the 7575 anonymous testers, 53% were men who have sex with men (MSM) or injection drug users (IDU). Of the 27,553 confidential testers, 79% were non-MSM-non-IDU.  Overall HIV prevalence was higher among anonymous testers than among confidential testers (1.9% vs 0.6%, respectively). Overall HIV incidence was higher among anonymous testers (1.3 per 100 person-years, 95%CI 0.8 to 2.0) than among confidential testers (0.4 per 100 person-years, 95%CI 0.3 to 0.6). This relationship was true for nearly all subgroups. However, HIV prevalence and incidence among MSM-non-IDU did not differ by registration type, and prevalence among MSM-IDU did not differ by registration type. 

Conclusions:  In this cohort, HIV prevalence and HIV incidence were greater among anonymous testers than among confidential testers. This is largely explained by the much higher proportion of MSM and MSM-IDU who test anonymously and the large proportion of lower risk non-MSM-non-IDU who make up the confidential testers. Measures of HIV among reportable, confidential tests may underestimate true rates, except when limited to subgroups that carry the greatest disease burden.