A joint survival model for estimating the association between viral load outcome and survival time to death among HIV/AIDS patients attending health care and treatment centers in Tanzania

Abstract

Abstract Background Globally, HIV/AIDS is one of the diseases that have a huge burden in terms of cost and health of individuals; and Sub-Sahara Africa is the highly affected region by the pandemic. Tanzania is among the countries that have a higher prevalence of HIV/AIDS-related mortality. This study aimed at using the joint survival model to estimate the association between viral load outcome and survival outcome to death adjusting for age, sex, adherence, and visit date. Methods Secondary data from a retrospective cohort of HIV patients attending health care and treatment centers were used to analyze the association between the longitudinal viral load and time-to-death outcomes. The three-step analysis was based on the individual mixed effects linear model and the Cox proportional hazards models to estimate the significance of the independent outcomes, and the joint survival model as a final model. The joint model was used to estimate the factors affecting the average change in log viral load over time and the risk factors for the survival time of HIV patients. The exposures for both models were ART adherence status, age, male, and visit date whereas the outcome for the LMM was log viral load and the outcome for the Cox PH model was time-to-death in years. Results The joint survival model results revealed that a 10-year increase in age was associated with a 37% increased risk of death (HR = 1.369, 95% CI: 1.253–1.844), and being male was associated with a 49% higher risk of death (HR = 1.489, 95% CI: 1.202–1.844) compared to females. The results also provided evidence of an association between the longitudinal log viral load and the survival time to death ) whereby a unit increase in the log viral load was associated with a 26% increase in the risk of death (HR = 1.262, 95% CI: 1.226–1.301). Conclusion The joint survival model analysis provided valuable insights into the associations between time to death and log viral load with adherence to ART, age, visit date, and sex of the patients. This implies that viral load suppression, as well as sex and age-specific interventions, are necessary for reducing HIV/AIDS-related deaths.