Correlates of ART attrition among adults under antiretroviral therapy in Southern Ethiopia, retrospective cohort study.

Abstract

High attrition rates from ART are the primary contributors to morbidity, death, hospitalisation, rising transmission rates, treatment failure, rising burden of opportunistic infections (OIs), and the evolution of HIV-virus resistance (HIVDR). In Sub-Saharan Africa, more than two-thirds of ART patients will not receive continuous care. There is little information about the correlates that contribute to attrition from ART services among ART patients in Southern Ethiopia. Hence, this study aims to identify correlates of attrition from antiretroviral therapy services for adults under antiretroviral therapy at Otona Teaching and Referral Hospital, Wolaita Zone, Southern Ethiopia. From 1 January 2013 to 31 December 2017, a retrospective cohort analysis was performed. The pre-determined 328 medical records were chosen using a simple random sampling technique using computer-generated random numbers. Epi Info version 3.5.3 was used to enter and clean the data, which were then exported to STATA version 11 for analysis. The Cox proportional hazards model, both bivariate and multivariable, was used. Variables with p-values less than 0.25 in bivariate analysis were considered candidates for multivariable analysis, and variables with p-values less than 0.05 were deemed statistically important in multivariable analysis. The intensity of the correlation and statistical significance were determined using the CHR, AHR, and 95 per cent confidence intervals. The magnitude of attrition from ART service was 21.60% (95% CI: 17.10, 26.10). The distance between home and hospital is more than five kilometres (AHR:3.84;95% CI: 1.99,7.38), no registered phone number (AHR:2.47;95%CI:1.32,4.09), have not taken isoniazid prophylaxis (AHR:2.23;95%CI:1.30,4.09), alcohol consumption (AHR: 1.77; 95% CI:1.01, 3.12), and had no caregiver (AHR: 2.11; 95% CI:1.23, 3.60) were statistically significant in the Cox proportional hazard model. Distance between home and hospital, phone number registration on follow-up chart, having a history of alcohol consumption, isoniazid prophylaxis provision, and having family support were independent correlates of attrition from antiretroviral treatment services.