Bayesian semiparametric partially linear cure models with partly interval-censored data.

Area of Science:

• Biostatistics
• Survival Analysis
• Epidemiology

Background:

• Partly interval-censored data with cure fractions are common in biomedical and epidemiological studies.
• Conventional mixture cure models often assume linear covariate effects, limiting flexibility for nonlinear relationships.

Purpose of the Study:

• To propose a flexible semiparametric mixture cure model accommodating both parametric and nonparametric covariate structures.
• To address limitations of conventional models by incorporating nonlinear covariate effects.

Main Methods:

• Utilized spline-based techniques for approximating unspecified functions.
• Implemented a four-stage data augmentation approach for model and data complexities.
• Developed a Bayesian approach for posterior estimation of model parameters.

Main Results:

• The proposed semiparametric mixture cure model demonstrated flexibility in handling nonlinear covariate effects.
• Simulation studies confirmed the finite-sample performance of the novel method.
• The approach was successfully applied to child mortality data, showing practical utility.

Conclusions:

• The flexible semiparametric mixture cure model offers an improved approach for analyzing survival data with cure fractions and nonlinear covariate effects.
• This method enhances the applicability of cure models in epidemiological and biomedical research.
• The Bayesian framework provides a computationally convenient way to estimate model parameters.