Statistical Comparison of Survival Models for Analysis of Cancer Data

Abstract

<b>Background:</b> The Cox Proportional Hazard model is the most popular technique to analysis the effects ofcovariates on survival time but under certain circumstances parametric models may offer advantages overCox’s model. In this study we use Cox regression and alternative parametric models such as: Weibull, Exponentialand Lognormal models to evaluate prognostic factors affecting survival of patients with stomach cancer.Comparisons were made to find the best model. <br/><b>Methods</b>: To determine independent prognostic factors reducingsurvival time for stomach cancer, we compared parametric and semi-parametric methods applied to patientswho registered in one cancer registry center located in southern Iran using the Akaike Information Criterion.<br/><b>Results</b>: Of a total of 442 patients, 266 (60.2%) died. The results of data analysis using Cox and parametricmodels were approximately similar. Patients with ages 60-75 and >75 years at diagnosis had an increased riskfor death followed by those with poor differentiated grade and presence of distant metastasis (P<0.05). <br/><b>Conclusion</b>:Although the Hazard Ratio in Cox model and parametric ones are approximately similar, according to AkaikeInformation Criterion, the Weibull and Exponential models are the most favorable for survival analysis.