Aplicação do modelo de regressão de Cox

Abstract

The main purpose of this research was study the survival analyses by statistics techniques to censored data. In this case the outcome variable is the time until the event occur, called failure time. In this research we studied the Cox Proportional Hazard Model with application in medical data. For the analyses we used data from Hospital Universitário de Santa Maria using a follow-up of 69 patients enrolled in the study from 1991 to 1999. After a confirmed diagnosis of system digestive cancer, members were followed until death or until the end of the study or until the subject was lost to follow-up. The covariates analyzed were: age, type of surgery, disease stage, type of treatment and date of the death or censored. The survival function was estimated by Kaplan-Meier method and Log-rank and Chi-square tests were used to compare the survival curves. The Cox Proportional Hazard Model was used to model the survival time data. The mean age was 60 years and 71% died because the cancer. The median survival time was 12 month. The Kaplan-Meier and Log-rank analysis showed that the survival curves by surgery type were not significantly different (p>0.05) and the survival curves by disease stage and treatment type was significantly different (p<0.05). The Cox regression model obtained included just the covariate disease stage (p=0.001), tested by partial likelihood ratio and Wald tests. The conclusion was that the death risk increase with the increased of thedisease stage and this is expected.