A Study of Survival Modelling in Dialysis Patients Applying Different Statistical Tools

Diana-Silvia ZILISTEANU
Ion Radu ZILISTEANU
Mihai VOICULESCU

Keywords

statistics, SPSS, survival analysis, Cox proportional hazard regression, mathematical model

Abstract

The aim of this study is to develop a model for survival probabilities of incident dialysis patients based on demographic, clinical and biological characteristics. We used statistical methods, mainly Cox regression, and 2 statistical tools: SPSS version 21 and Excel. During the first stage, data were analysed using SPSS software, edition 21. We performed survival analysis using Cox proportional hazard regression test, to assess the relationship of explanatory variables with survival time. Second stage of analysis was performed using Excel computations based on the results provided by Cox analysis. Starting from the basal risk curve, and applying the coefficients derived from the Cox regression analysis, the hazard curve was calculated for any combination of values for the variables included in the equation. Based on these elements, we constructed an Excel model for survival simulation.