Electronic data sharing between public health department and clinical providers improves accuracy of HIV retention data

Abstract

BackgroundRetention in care is critical for treatment and prevention of HIV. Many HIV care clinics measure retention rate, but data are often incomplete for patients who are classified as lost to follow-up but may be actually in care elsewhere, moved, or died. The Data to Care (D2C) initiative supports data sharing between health departments and HIV providers to confirm patient care status and facilitate reengagement efforts for out of care HIV patients.MethodsThe University of Chicago Medicine (UCM) provided an electronic list to the Chicago Department of Public Health (CDPH) of adult HIV-positive patients whose retention status was not certain. Retention in care was defined as at least 2 visits >90 days apart within the prior 12 months. CDPH matched this list of patients with data from the Chicago electronic HIV surveillance database. Matches were based on patient name, including alternative spellings and phonetics, and birth date. CDPH also cross-checked patient names with the CDC’s national enhanced HIV-AIDS Reporting System (eHARS) database. CDPH provided UCM with patient current care status, i.e., patient was in care elsewhere (as verified by lab data), moved out of state, or deceased.Results780 HIV-positive patients received care in the UCM adult HIV clinic from January 1, 2013 to March 31, 2017. Of these, 360 were retained in care as of March 2017. We shared data with CDPH for 492 patients. Of these, 294 (59.8%) were matched, and 168 (34.1%) had a date of last medical care provided. See Table 1 for patient dispositions, before and after data sharing. 24 (13.4%) of patients believed to be lost to follow up according to UCM records were confirmed either transferred care or deceased according to health department data.ConclusionData sharing between the health department and HIV providers can improve data accuracy regarding retention in care among people living with HIV.Table 1. HIV Patient disposition data before and after data sharing with health departmentPatient dispositionBefore data sharing, n (%)After data sharing, n (%)Currently in care at UCM360 (46.2%)360 (46.2%)Transferred139 (17.8%)143 (18.3%)Deceased63 (8.1%)83 (10.6%)Incarcerated1 (0.1%)1 (0.1%)Moved out of jurisdiction39 (5.0%)39 (5.0%)Lost to follow up178 (22.8%)154 (19.7%)Disclosures J. P. Ridgway, Gilead FOCUS: Grant Investigator, Grant recipient; D. Pitrak, Gilead Sciences FOCUS: Grant Investigator, Grant recipient

Development of a predictive model for retention in HIV care using natural language processing of clinical notes

“…5 Public health agencies and researchers have recently shown interest in utilizing electronic medical record (EMR) data to improve retention in care and other HIV care continuum outcomes. [6][7][8] Electronic medical records have rapidly expanded in the last decade, with over 95% of US hospitals and 86% of US ambulatory clinics utilizing EMRs. 9 EMRs contain vast amounts of data that can be analyzed to understand and predict HIV outcomes.…”