In the US only 30% of all HIV-infected (HIV+) persons are engaged in HIV care and on ART. Accordingly, key to addressing the US HIV epidemic is to identify out-of-care HIV+ persons and to link them into sustainable care. Criminal justice involvement (CJI) is endemic to communities most affected by the HIV epidemic and it creates barriers to HIV care. On the other hand, CJI?and specifically incarceration?presents opportunities to identify out-of-care HIV+ persons and facilitate their return to care. Most incarcerations occur in jails: county- or municipal-operated facilities that house pretrial detainees and inmates sentenced to ? 1 year. Research examining continuity of care among jail inmates has been limited to large urban jails, predominantly in the Northeast. Little HIV research among jail inmates has been conducted in the South, a region with small and mid-sized rural jails and a disproportionately high number of HIV deaths. As a result, a substantial gap exists in elucidating the epidemiology of HIV in jails and identifying opportunities to maintain and re-establish care in this setting. Statewide repositories of jail medical records could facilitate the analysis of HIV care across representative samples of jails, but given jails? local, independent operation, such repositories do not exist. In response, we propose a Big Data solution: creation of a de-identified database integrating 1) individual- level, inmate records published daily on jail websites and retrieved using a web-scraping program that automates data collection and 2) confidential HIV diagnosis and care records maintained by the state department of public health (DPH). The integrated data will be used to assess the burden of HIV among jail inmates and their continuity of HIV care as they transition between jail and the community. Beyond population- level surveillance, a real-time version of our database that retains personal identifiers could be used by DPH to identify HIV+ jail inmates not engaged in care prior to or during incarceration, and state disease specialists could then intervene to coordinate receipt of HIV care in jail and linkage to community care at release. However, use of a surveillance database with identifiers raises several ethical considerations. Accordingly, for HIV+ inmates incarcerated in southern state with a large and heterogeneous array of jails, we will: 1) Develop a highly novel, de-identified database combining jail and state HIV records; 2) Use the de- identified database to examine burden of known HIV in county jails, assess inmates? use of HIV services before, during and after incarceration, and identify inmate and facility factors associated with (dis)continuity of care; and 3) Explore HIV+ inmates?, jailors?, and other stakeholders? perspectives on the use of inmates? anonymized data for HIV surveillance and the potential use of their identifiable data to enhance receipt of care during incarceration and after release. Our proposed Big Data project will identify jails which would benefit from greater HIV resources, and explore the potential pitfalls and strengths of leveraging identifiable, sensitive data to improve continuity of care and reduce HIV infectivity among jail inmates.
Existing research examining the epidemiology of HIV among jail inmates and their use of healthcare has been limited, and largely conducted in high capacity urban jails. This project proposes to create a de-identified HIV surveillance database to better understand the burden of HIV in jails and inmates? utilization of HIV care as they transition between correctional and community settings in a southern state populated with small and mid-sized rural jails. We will also explore the ethics of creating a version of our surveillance database that retains identifiers, so that state outreach workers can identify and contact out-of-care HIV-positive inmates to facilitate their re-entry into HIV care during incarceration and after release.
