Modelling Survival Among Tuberculosis Patients in Kenya

Abstract

The airborne illness tuberculosis (TB), which mostly affects the lungs, is a substantial source of mortality when the right controls are not implemented. Tb is preventable and treatable, but despite this, close to 10 million individuals contract it annually throughout the world. This study examined factors related to survival time in tuberculosis patients from Kenya. This study compared survival rates across Treatment phases and several subcounties. The study also sought to determine factors that predict survival among Kenyan patients with tuberculosis. The study employed secondary population-based tuberculosis data from electronic database Treatment Information from Basic Unit (TIBU). To compare survival function for Treatment stages and the different sub-counties, the Kaplan-Meier estimation method was utilized. Independent variables against survival time were modeled using the Cox proportional hazards regression model . The results of the Cox proportional hazards regression examining the relationship between predictors and survival time were as follows: Age, HR = 1.0241 (CI: 1.0028 – 1.046). Consequently, the risk of death for someone with tuberculosis is 2.4% higher for every additional year. HIV, HR = 3.0812(CI: 1.6508 – 5.751). When diagnosed with tuberculosis, HIV positive people had a 3 times higher mortality rate than HIV negative individuals. TB type and Alcohol consumptionwere significant predictors of death in TB patients. Kaplan- Meier estimation method shows that survival between Intensive and Continuation Phase were significantly different as it had a p-value of <0.0001. Results from Kaplan-Meier data, however, did not reveal any significant variations in survival in the four sub-counties