Study and Analysis of the Chest Cancer Data Using Survival Models

Abstract

This research aimed to estimate the effects of prognostic factors on chest cancer survival, the research studied two models in survival analysis; the Cox-Proportional Hazard (PH) model is most usable method in present time of survival data in the occurrence covariate or prognosticates aspects, and the Accelerated Failure Time (AFT) model is another substitute way for analysis of survival data. KaplanMeier method has been applied to survival function and hazard function for estimation, the log-rank test was used to test the differences in the survival analysis. The data was obtained from Nanakali Hospital in the period from 1st January 2013 to 31st December 2017 with follow up period until 1st April 2018. The results for Kaplan-Meier and log-rank test showed the significant difference in survival or death by chest cancer for all presented related prognostic factors. The Cox-PH and AFT model does not identify the same prognostic factors that influenced in chest cancer survival. The Cox Proportional Hazards model displays a significant lack of fit while the accelerated failure time model describes the data well. AFT with Weibull
distribution was chosen to be the best model for our data by using Tow model selection criterion; Akaike Information Criterion (AIC) and Bayesian information criterion (BIC). Also, the results performed by the statistical package in Mat-lab, Stat-graphic and SPSS, which was used to analyze the data.