Spatial Modelling Of Tb Prevalence In Kenya Using Integrated Nested Laplace Approximation (INLA)

Abstract

Despite there being nationwide reduction in TB Cases incidence, there has been an increase in TB infection cases in various areas in the country The use of registered numbers of reported new cases of TB is not enough to explain the distribution pattern and the determinants of the disease. Spatial statistics as method of analysis is a possible way to have the knowledge on the areas where the disease infection and distribution pattern is high. In this study we apply Integrated Nested Laplace Approximation (INLA). INLA is a new tool for Bayesian inference on Latent Gaussian fields which substitute Markov Chain Monte Carlos (MCMC) simulations with deterministic approximations to the posterior marginals of interest. Its main benefits are the extreme speed and the high accuracy of the results. Moreover INLA can be easily implemented through the INLA-program and its R-interface. The study shows that TB prevalence is distributed different within the different provinces with Nyanza province having the largest number of the population with TB with a proportion of over 15% followed by Rift Valley province with a prevalence of between 10% to 15%, the province with the least number of people with TB is North Eastern province with a prevalence of less than 2%. The study shows that there some areas in the country that have lesser health facilities, with North Eastern recording a negligible number of health facilities, this could imply that there may be more TB cases in the region, but due to lack of health facilities not all the active cases are registered The main factors that affect the spread of TB from the study include; poverty, proportion taught on spread and control of TB, urban residence, all these factors are significant each lying in the 0.5 quant. This implies that the government needs to put in place policy to improve the living standards of urban residents as well as the per capita income of her population.