Adjustment of survival analysis model for tuberculosis data from the city of Patos de Minas - Minas Gerais

Abstract

Tuberculosis has been one of the most studied their biological epidemiology, diagnosis, prophylactic and therapeutic aspects diseases. However this disease stubbornly persists worldwide, causing human suffering and death in the 21st century. Thus there is a need for epidemiological studies associated with statistical techniques such as Analysis of Survival have to study data from experiments in which the response variable is usually the time to occurrence of an event of interest being applied extensively in the medical field . In this paper the AFT parametric model was used to model the time until healing function of the covariates conside-
33 red in determining the distribution that best fits the data , to analyze the effect of covariates on time to occurrence of an event , death from tuberculosis or healing in 437 patients detected with tuberculosis treated by the Regional Department of Health - Patos de Minas - MG , in the period 2003 - 2012. We conclude that in relation to the covariate sex, a female patient is 109.38 % more likely to cure that in masculine patients . The results also suggest patients from Patos de Minas city is 259.18 % more in the probability of cure than patients from other regions.