RESUMO
OBJECTIVE: Various HIV testing models have been described, but widespread implementation has lagged. We describe a clinical HIV testing program notable for its use of conventional (nonrapid) assays, native hospital personnel, and an electronic system to aid targeted patient selection. METHODS: Clinical HIV testing program records and hospital emergency department (ED) and laboratory records were reviewed and linked for the period from January 2007 until November 5, 2008. RESULTS: There were 103,475 visits to the ED, for which 1,258 (1.2%) resulted in HIV testing, and 54 (4.3%) were positive for HIV antibody. Result notification was successful for 52 (96%) individuals with positive test results. After reporting to the health department, it was determined that 28 (2.2%) individuals had received a new diagnosis, of whom 89% were linked with care. Mean baseline CD4 counts for new diagnoses for periods 1 through 3, respectively, were (1) unavailable, (2) 138 cells/µL (95% confidence interval [CI] 34 to 242 cells/µL), and (3) 222 cells/µL (95% CI 119 to 325 cells/µL). Overall, mean calculated to be 180 cells/µL (95% CI 16 to 345 cells/µL). CONCLUSION: This ED HIV testing model successfully expanded new patient identification, result notification, and linkage to care. Although this effort falls short of Centers for Disease Control and Prevention recommendations, the model can be implemented widely without external funding for parallel staffing or rapid assays.
