Bayesian analysis of survival data with missing censoring indicators

Area of Science:

• Biostatistics
• Clinical Trials
• Survival Analysis

Background:

• Large clinical studies often face challenges in collecting complete data, particularly regarding the precise timing of events.
• Missing censoring indicators in survival data can arise when physical examinations are impractical at the final monitoring time.
• The probability of missing data may be influenced by monitoring time and patient covariates.

Purpose of the Study:

• To develop a robust statistical method for analyzing survival data with missing censoring indicators.
• To estimate regression parameters within the Cox proportional hazards model framework.
• To address the complexities introduced by non-random missingness in clinical trial data.

Main Methods:

• A fully Bayesian semi-parametric approach was employed for survival data analysis.
• The method specifically targets situations with missing censoring indicators.
• Regression parameters of the Cox proportional hazards model were estimated.

Main Results:

• Theoretical investigations demonstrated the efficacy of the proposed Bayesian method.
• Simulation studies confirmed superior performance compared to existing methods.
• The method was successfully applied to real-world data from the Orofacial Pain study.

Conclusions:

• The proposed Bayesian semi-parametric method offers a reliable solution for survival data with missing censoring indicators.
• This approach enhances the accuracy of regression parameter estimation in large clinical studies.
• The findings have significant implications for the analysis of complex clinical trial data.