BACKGROUND: The enhanced HIV/AIDS Reporting System (eHARS), the primary HIV surveillance database, is used to monitor health outcomes for persons living with HIV (PLWH). It contains information on HIV diagnosis and patient demographics, as well as CD4 and viral load test results, which are used to assess health outcomes (e.g. viral suppression). However, not all indicators of care receipt are captured in eHARS. In response, Virginia created the Care Markers database (CMDB), which compiles multiple data sources, including eHARS, into a single, integrated system. We aim to quantify the impact of data integration via CMDB on viral suppression estimates.
METHODS: CMDB compiles routine HIV surveillance data (eHARS) and other care marker data, including CD4 cell count, viral load test, HIV-related medical visit, and receipt of antiretroviral prescription. Data sources include Ryan White, the Medical Monitoring Project, Medicaid, and HIV surveillance data from other jurisdictions. Vital status data come from the Department of Health’s Division of Vital Statistics, the National Death Index, the Social Security Death Master File, routine updates from other jurisdictions, and matching with LexisNexis ® Accurint ®. Our population of interest was PLWH in Virginia as of December 31, 2014 according to eHARS. Viral suppression was assessed with 2 different data sources among our study population: 1) eHARS alone; 2) CMDB. We tabulated and compared frequency and percentage of 2014 viral suppression for the two data sources.
RESULTS: We identified 23,043 PLWH in Virginia as of December 31, 2014 according to eHARS. With eHARS data alone, the percentage of PLWH virally suppressed was estimated at 29.9% (n=6,880), while 38.9% (n=8,695) of PLWH were estimated to be virally suppressed using the full CMDB.
CONCLUSIONS: Incorporation of multiple data sources via CMDB resulted in increased viral suppression estimates for PLWH in Virginia, demonstrating the importance of data integration for monitoring health outcomes among PLWH. CMDB has the potential to increase patient history completeness and health outcome accuracy. Future analyses will estimate the contribution of each data source in CMDB.