A quality framework to improve routine rapid HIV screening, diagnosis, and linkage to care at a high-volume, urban emergency department in New York City

Abstract

BackgroundEarly identification and treatment of HIV infection reduces morbidity and mortality and the likelihood of transmission to others. In 2010, the Spencer Cox Center for Health at St Luke's and Roosevelt Hospitals partnered with the emergency department to move from a counsellor-based to an integrated model of oral rapid HIV antibody (RHIV) testing in the hospital's two emergency rooms in New York City. For this descriptive study, we collected data for patients seen in the emergency department, and new HIV-positive patients linked to care between 2011 and 2012, to understand barriers to programme implementation and improve quality of routine rapid HIV testing services. MethodsOver 24 months between Jan 1, 2011, and Dec 31, 2012, we implemented the integrated HIV testing model and conducted several monitoring and quality improvement projects. Every month, we measured the number of patients eligible, triaged, offered, accepted, and completed the RHIV test, and the number who tested positive and were linked to care. Data were collected from emergency department, outpatient, and inpatient electronic health records, and were compiled as a necessary part of quality management and the provision of linkage to care. Additional analyses included acceptance rate by site and by triage nurse. Nurses with fewer than 20 triage visits were excluded. Acceptance rate data were extracted from the emergency department electronic health record in aggregate form on a monthly basis, and entered into SPSS for statistical analysis. FindingsModel change resulted in a six-fold increase in testing within 6 months. Of the 339 449 triaged visits, patients from 323 575 (95·3%) visits were eligible for screening (>13 years, triage acuity level III or higher), 305 791 (90·1%) patients were offered the RHIV test, and 34 598 (11·3%) eligible people accepted HIV testing. Among these, 25 690 (74·3%) tests were completed. 81 (0·32%) new HIV cases were identified, of which 61 (75·3%) had an HIV primary care visit within 90 days of preliminary test. In subanalyses, we identified significant differences in individual and site performance for test acceptance and test completion. Among the 105 nurses included in the analysis, acceptance rates ranged from 0·4% to 30·8% (mean 10·8%). Nurses at one emergency room had a significantly higher mean acceptance rate (0·13, SD 0·06) than did nurses at the other (0·08, SD 0·07; p<0·0001). Linkage to care improved over time. InterpretationIntegrating routine, near-universal screening into a high-volume emergency department is feasible. From these results, five quality gaps were identified for targeted intervention: eligibility, test offer, acceptance rate, test completion, and linkage to care. These gaps are likely to exist in other routine screening programmes, and each can be targeted with additional measurement and quality interventions. Interventions that we have evaluated include changes in the emergency medical room, dissemination of individual performance reports, and collaboration with the local health department to reach patients lost to follow-up. FundingThe St Luke's and Roosevelt Hospitals routine RHIV testing programme receives grant funding from the New York City Department of Health and Mental Hygiene for costs associated with RHIV testing of uninsured patients.