Use of electronic data to assist with long-term follow-up in an urban HIV clinic

Abstract

BackgroundThe University of Chicago Medicine (UCM) HIV clinic follows up 565 patients, with case management by one social worker. This large cohort makes active case follow-up difficult. Our clinic is located in a neighbourhood on the south side of Chicago. This area includes four of the six communities with the highest annual HIV diagnosis rates in the city and contains neighbourhoods with an HIV prevalence of 1% or more. We describe a programme that used medical informatics to re-link patients lost to follow-up into care and to assist in the long-term management of this group. MethodsPatients at the UCM HIV clinic were eligible for the study if they had attended at least one appointment but had not been seen in clinic for 6 months or more, our criterion for lost to follow-up. We used an electronic clinic database and electronic medical records to identify these patients. Those patients classified as lost to follow-up were re-linked to care by clinic social workers by phone or, if that failed, by follow-up letters. Patients were contacted from February to June, 2013, by the social worker and a social work intern. In an open programme, patients will become eligible for re-linkage on a rolling basis. FindingsAmong 565 patients, 56 (10%) had not been seen for 6 months to 1 year, 40 (7%) had not been seen for 1–2 years, and 27 (5%) had not been seen for more than 2 years. Overall, 442 (78%) of our patients were retained in care and 123 (22%) patients were determined to be lost to follow-up. Age, sex, race, ethnicity, and insurance coverage did not predict lost to follow-up status. Ten patients (8%) were dead. Among 113 living patients, 64 (57%) were contacted by phone. 32 had moved or were in care elsewhere. 30 of the remaining 32 (94%) scheduled a clinic visit. Ten patients scheduled a visit in response to a letter or voicemail, but 49 (9% of the clinic cohort) remain lost to follow-up. InterpretationOur retention rate is much higher than the US national retention rate of about 40%. Nonetheless, with a large cohort, it is important to implement measures to track retention. With this aim, we found electronic data to be useful for identification of patients lost to follow-up and to aid in efforts towards retention in care. We were very successful at collecting data from or re-linking patients who we could contact by phone, but less so when we had to send a letter. In addition to re-engaging patients into care, it is important for us to know if patients have died or transferred care to adequately assess our local cascade of engagement in care. With this study we hope to show the feasibility and sustainability of clinic-wide re-linkage initiatives. FundingNone.