Tekleweyni Haftu2025-06-212024-11-28https://repository.mu.edu.et/handle/123456789/67710.82589/muir-592Background: Human immunodeficiency virus (HIV) is an infection that attacks the body’s immune system. Acquired immunodeficiency syndrome (AIDS) is the most advanced stage of the disease. HIV remains one of the world's most significant public health challenges, particularly in low- and middle-income countries. Objectives: This study aimed to determine the potential risk factors affecting the time-to-death of HIV-TB co-infected patients in Mekelle and Ayder Comprehensive Referral hospitals and also demonstrate the application of parametric survival models in analyzing these survival outcomes. Methods: Parametric shared frailty models have been used with four baseline hazard functions (Exponential, Weibull, Log-logistic, Log-normal) and two frailty distributions (Gamma, InverseGaussian). 215 HIV-TB co-infected patients whose age was 18 and above who started ART from January 2015 to December 2016 and followed up to January 2020 were included in the study from Mekelle and Ayder hospitals. Data were analyzed using statistical software R version 4.4.1 and STATA version 14.0. Results: Out of 215 patients, about 60(28%) died while 155(72%) were censored during the follow-up period with the median death time of HIV-TB co-infected patients at 30 months. The clustering effect was significant, and the Weibull-Inverse Gaussian shared frailty model was preferred over the rest parametric shared frailty models based on the Akaike Information Criterion (AIC) and Bayesian Information Criteria (BIC). The result shows patients’ educational level, baseline Weight, TB type, baseline CD4, and place of residence were significant, whereas sex, WHO clinical stage, marital status, functional status, regimen type, and age were not significant covariates for HIV-TB co-infected patients in Mekelle and Ayder comprehensive Referral hospitals. The clustering effect was significant for modeling the time-to-death of HIV-TB patients’ dataset and there was heterogeneity among the hospitals on the death of patients (θ=040.). Conclusion: This study showed that there was a clustering (frailty) effect on modeling time to Death among patients treated in Mekelle and Ayder comprehensive hospitals because of heterogeneity in hospitals from which the patients are treated, assuming patients treated in the same hospital share similar risk factors related to death.enSurvival analysisParametric shared frailtyHIV-TB co-infectedTime-to-deathClustering effectThesis