To determine the proportion of HIV-infected sexually transmitted disease (STD) clinic patients identified during routine, voluntary HIV counseling and testing and to characterize patients accepting and refusing counseling and testing, we linked data from a blinded HIV seroprevalence survey to data from the HIV counseling and testing program. This study characterizes patients accepting and refusing routine HIV counseling and testing in two public STD clinics. A cross-sectional, blinded HIV seroprevalence survey was conducted of 1,232 STD clinic patients offered HIV counseling and testing. HIV seroprevalence was higher among patients who refused voluntary testing (7.8% versus 3.6%, P = 0.001). Patients who refused testing were more likely to report a prior HIV test (45.6% versus 27.2%; P < 0.001). Among patients reporting a prior HIV test, differences were noted between reported prior results, both positive and negative, and blinded results. HIV-infected STD patients may not be detected by routine HIV testing, and self-reported HIV results should be confirmed.