Modeling Time to Death using the application of Survival model among HIV/TB Co-infected Patients who are under ART Follow-up in Axsum-Hospital, Ethiopia.



ART, Axum Hospital, HIV/TB co-infected, Risk factors, Survival Analysis


The relationship between Human Immune Deficiency Virus (HIV) and Tuberculosis (TB) is complex, health problem resulting in the synergistic increases in patients’’ morbidity and mortality. The probability of being infected by both infections is attention seeking health issue worldwide. The application of combination of antiretroviral therapy (ARV) in 1996 has immense effect in extending the life span of infected patients by slowing the wasting period, and by boosting the CD4 cell count of an infected patient. This study aimed to identify factors that influence the survival status of HIV/TB co-infected patients who are under ART follow-up and to assess the effectiveness of Cox PH model compared to parametric models in modelling time-to-death. The resulting data set comprises 210 cases all of them HIV-infected TB patients who are above the age of 15 years, and who have started anti-TB treatment between the years 2011 and 2015. The Cox proportional hazard model used to compare parametric models in modelling time-to-death. A total 210 participants 169 (80.5%) were died due to the disease, and 41 (19.5%) were not presented to follow-up during the time of data collection. From the total 210 ART followers, 88 (41.9%) were male while the rest were female. The study showed that the WHO clinical stage III is 1.27 (p = 0.023) indicating that WHO clinical stage III has the tendency to prolong the survival time of HIV/TB co-infected patient compared to stage IV. The study also revealed that the Accelerated failure time model has the best predictive power compared to the Cox model based on the AIC values. The best fitted model for survival analysis is the Generalized Gamma Accelerated failure time model. Among the several prognostic factors Age, CD4 cell count, and WHO Clinical Stages II and III were identified as significant prognostic factors.

Additional Files