Background:  South Carolina had a 2004 AIDS case rate ranked 10^th in the nation at 18.1 per 100,000. We sought to determine to what extent the South Carolina HIV/AIDS cases in the recent past could have been diagnosed earlier at a prior healthcare encounter.

Methods:  Our strategy was to link HIV/AIDS case reports in South Carolina to prior healthcare encounters. HIV data from individuals first diagnosed with HIV infection between January 2001 and February 2006 were matched using several variables, such as gender, race/ethnicity, and county of patient’s residence, with encounters in 60 emergency rooms, 62 inpatient facilities, 63 outpatient surgical facilities, and 19 free medical clinics around the state. Medical encounters were categorized to distinguish visits that were likely versus unlikely to have prompted an HIV test. Diagnostic codes likely to suggest the need for an HIV test were:  sexually transmitted diseases, acute retroviral syndrome, 37 diagnostic codes possibly or probably related to HIV (eg, lymphadenopathy, fevers, tuberculosis, pneumonia), and codes related to intravenous drug use. All remaining codes were considered unlikely to suggest the need for an HIV test. Odds ratios and 95% confidence intervals were used to compare the significance of gender, race, exposure group, and number of visits as determinants of late versus early testing.

Results:  Of the 4345 HIV⁺ individuals diagnosed between 2001 and 2006, 3157 (72.7%) had visited a South Carolina healthcare facility 1 or more times prior to testing HIV⁺. Of the 3157 persons, 1316 (41.6%) developed AIDS within 1 year of testing (late testers).  Fewer females than males were late testers (OR 0.63, 95%CI 0.55 to 0.74). None of the other factors was significant. Overall, 20,271 separate visits were recorded for these 3157 persons, and 11,746 visits (57.9%) occurred no more than 3 years prior to HIV testing. Diagnostic codes for 15,469 visits (76.3%) were for diagnoses unlikely to prompt an HIV test, and diagnostic codes for 4802 visits (23.7%) were likely to suggest the need for an HIV test.

Conclusions:  In South Carolina, with an HIV/AIDS prevalence ranked 10^th of 50 states, there were many prior healthcare visits among HIV-infected persons that could have provided an opportunity for earlier HIV diagnosis. However, most of the diagnostic codes from these healthcare visits would not have prompted an HIV test. These data present a persuasive argument for routine HIV screening in healthcare settings to detect undiagnosed HIV in South Carolina.