Determinants of loss to follow-up in patients receiving antiretroviral therapy and variations among health facilities in Oromia, Ethiopia: Application of the Generalized Estimating Equation model

Abstract

Background: Loss to follow-up (LTFU) of HIV positive patients from antiretroviral therapy (ART) reduces treatment benefits and leads to treatment failures. Determinants of LTFU in ART patients were investigated in Oromia region, but various studies had insufficient sample sizes and may not be generalizable to the entire population of the region. The objectives of this study was to investigate the determinants of LTFU and variations in health facilities among HIV patients receiving ART in Oromia Region. Methods: A retrospective cohort study based on facility records was utilized and the data were extracted from the dataset gathered collaboratively by Jimma University Health Institute and Health Bureau of Oromia. Data cleaning and descriptive analysis was done using SPSS while R software was used for inferencial analysis. Generalized Estimating Equations was used to identify determinants of LTFU and variations between health facilities was handled by considering confounding by cluster (CBC). Result and Conclusion: A total of 875 patients from 19 facilities were included, of whom 432 (49.40%) were lost. The predictors identified were being male (AOR = 1.669; 95% CI: 1.110, 2.510), female sex workers (AOR = 6.931; 95% CI: 1.560, 30.800), not being enroled in ASM (AOR = 3.925; 95% CI: 2.720, 5.670) and drug regimen change/switch (AOR = 4.027; 95% CI: 2.700, 6.010). Low income level, WHOclinical stage I and Within-cluster effect of poor adherence also have statistically significant association with LTFU. The result also indicate that there is confounding by cluster among the Health facilities. Therefore, health professionals working in ART centers and potential stakeholders in HIV/AIDS care and treatment should consider patients with these characteristics to prevent or minimize the rate of LTFU of patients on ART.