Long-Term Survival of Patient with End-Stage Renal Disease Using Bayesian Mixture Cure Rate Frailty Models

Area of Science:

• Nephrology
• Biostatistics
• Epidemiology

Background:

• End-stage renal disease (ESRD) is rising globally, particularly in developing nations.
• Accurate statistical modeling is essential for understanding patient survival and heterogeneity.
• Controlling for cure patterns is crucial in survival analyses for ESRD patients.

Purpose of the Study:

• To apply advanced statistical models for analyzing survival data in hemodialysis patients.
• To identify factors influencing mortality risk in ESRD patients.
• To compare the efficacy of a mixed cure Weibull model with the Cox proportional hazards model.

Main Methods:

• A longitudinal study analyzed data from 170 hemodialysis patients (2006-2016).
• A gamma frailty mixed cure Weibull model (MC-WG) using Bayesian inference was employed for time-to-event (death) analysis.
• Cox proportional hazards (Cox PH) model was used for comparison.

Main Results:

• Mortality was observed in approximately 49% of patients, with a median survival time of 37.5 months.
• Increased mortality risk was associated with older age, female sex, and higher serum urea levels.
• Higher creatinine levels correlated with decreased mortality risk.

Conclusions:

• The MC-WG Bayesian model identified significant survival predictors (diabetes, AST, calcium, phosphorus, uric acid) not found by the Cox PH model.
• The MC-WG Bayesian model provides more consistent results with existing literature.
• Advanced statistical models offer more robust survival estimates for hemodialysis patients.