Leveraging routine viral load testing to integrate diabetes screening among patients on antiretroviral therapy in Malawi

BACKGROUND: People living with HIV are at an increased risk of diabetes mellitus due to HIV infection and exposure to antiretroviral therapy (ART). Despite this, integrated diabetes screening has not been implemented commonly in African HIV clinics. Our objective was to explore the feasibility of integrating diabetes screening into existing routine HIV viral load (VL) monitoring and to determine a group of HIV patients that benefit from a targeted screening for diabetes. METHODS: A mixed methods study was conducted from January to July 2018 among patients on ART aged≥18 y and healthcare workers at an urban HIV clinic in Zomba Central Hospital, Malawi. Patients who were due for routine VL monitoring underwent a finger-prick for simultaneous point-of-care glucose measurement and dried blood spot sampling for a VL test. Diabetes was diagnosed according to WHO criteria. We collected demographic and medical history information using an interviewer-administered questionnaire and electronic medical records. We conducted focus group discussions among healthcare workers about their experience and perceptions regarding the integrated diabetes screening program. RESULTS: Of patients undergoing routine VL monitoring, 1316 of 1385 (95%) had simultaneous screening for diabetes during the study period. The median age was 44 y (IQR: 38–53); 61% were female; 28% overweight or obese; and median ART duration was 83 mo (IQR: 48–115). At baseline, median CD4 count was 199 cells/mm(3) (IQR: 102–277) and 50% were in WHO clinical stages I or II; 45% were previously exposed to stavudine and 88% were virologically suppressed (<1000 copies/mL). Diabetes prevalence was 31/1316 (2.4%). Diabetes diagnosis was associated with age ≥40 y (adjusted OR [aOR] 7.44; 95% CI: 1.74 to 31.80), being overweight and/or obese (aOR 2.46; 95% CI: 1.13 to 5.38) and being on a protease inhibitor-based ART regimen (aOR 5.78; 95% CI: 2.30 to 14.50). Healthcare workers appreciated integrated diabetes screening but also reported challenges including increased waiting time, additional workload and inadequate communication of results to patients. CONCLUSIONS: Integrating diabetes screening with routine VL monitoring (every 2 y) seems feasible and was valued by healthcare workers. The additional cost of adding diabetes screening into VL clinics requires further study and could benefit from a targeted approach prioritizing patients aged ≥40 y, being overweight/obese and on protease inhibitor-based regimens.