Predictors of Mortality Among Adult Patients on Antiretroviral Therapy at Hawassa City Administration Public Hospitals, Ethiopia, 2015–2021: A Retrospective Cohort Study

Background Globally human immunodeficiency virus (HIV)/acquired immune deficiency syndrome (AIDS) continued as a public health concern, and millions of people died, and many were newly infected every year. Patients who enrolled in antiretroviral (ARV) therapy are still dying, and the predictors of mortality from study to study were inconsistent. This study aimed to assess the mortality rate and its predictors among adult patients on ARV therapy at Hawassa city administration public hospitals. Methods A retrospective cohort study was conducted in two hospitals at Hawassa city administration on 402 patients’ charts selected by simple random sampling methods using computer‐generated random numbers. The data were collected by reviewing patient charts. Data coding, entry, and verification were done using Epidata 3.1 and exported to STATA 14.1 for data analysis. Kaplan–Meier survival curve coupled with the log‐rank test was used to compare survival between different categories of predictors. The Cox proportional hazard regression model was used, and Cox–Snell residual model checked the model’s goodness of fit. The final model was interpreted by using the adjusted hazard ratio (AHR) with a 95% confidence interval (CI) and p ‐value < 0.05. Results The median follow‐up time in this study was 49 months (IQR = 29–69) At the end of follow‐up, followed for 18,835 person‐months of follow‐up, 53 patients died, and the mortality rate was 2.81 per 1000 person‐months observation. Fair/poor ART adherence (AHR 4.39, 95%CI = 2.36–8.17), WHO clinical stage III/IV (AHR = 2.4, 95% CI = 1.3–4.5), CD4 count less than 200 cells/µL (AHR 4.39, 95%CI = 2.36–8.17) ambulatory and bedridden functional status (AHR 2.5, 95%CI = 1.35–4.63) and not receiving isoniazid prophylaxis (AHR 4.3, 95%CI = 2.99–9.5) were statistically significant predictors of mortality. Conclusion The mortality rate in this study was relatively high. Advanced WHO clinical stages, lower CD4 cell count, poor/fair adherence, not receiving isoniazid prophylaxis therapy, and ambulatory/bedridden functionality were independent predictors of mortality.