RESUMEN
Background and Aim: Various methods are used to analyze the data in survival studies. The aim of this study was to compare the analysis results of penalized splines in Cox model and Cox proportional hazards model in relation to the effects of prognostic factors related to the survival of renal transplant patients
Material and Methods: This retrospective study included data of 876 renal transplant patients in Kermanshah, in Iran, from 2001 to 2015. Cox proportional hazards model and penalized splines in cox model were used. We used R version 3.3.2 for data analysis. The significance level was set at 0.05 and Akaike information criterion was used to compare the efficacy rates of these models
Results: Using Cox proportional hazards model, one, three and five year survival rates of the patients were 94.9, 92.3 and 90.2 percent, respectively. But the corresponding figures for penalized spline model were 95.6, 93.3 and 91.4 percent. The recipient's age, postoperative creatinine, urea after surgery and underlying diabetes were significantly associated with the patients' survival rates [p < 0.05]. Akaike's criteria for Cox proportional hazards model and penalized spline model were 1016.3 and 984.1, respectively
Conclusion: Considering lower Akaike information criterion and suitability of penalized spline model for continuous variables affecting the survival of renal transplant patients, this model can be used instead of Cox proportional hazards model for more accurate estimate of survival rate of the renal transplant patients