Extending multivariate Student's- semiparametric mixed models for longitudinal data with censored responses and...

Area of Science:

• Biostatistics
• Longitudinal Data Analysis
• Statistical Modeling

Background:

• Longitudinal studies track changes over time, often involving censored data.
• Standard models may lack flexibility or robustness for complex dependencies.
• Semiparametric mixed models offer a framework for such data.

Purpose of the Study:

• To extend semiparametric mixed models for longitudinal censored data.
• To incorporate Student's t-distribution for enhanced robustness.
• To utilize nonparametric regression for flexible covariate dependence.

Main Methods:

• Developed a semiparametric mixed model using Student's t-distribution.
• Employed nonparametric regression for functional covariate dependence.
• Used penalized likelihood and smoothing splines within an EM-type algorithm.
• Assessed robustness against outlying observations using Mahalanobis distance.

Main Results:

• The proposed model effectively handles longitudinal censored data.
• Maximum likelihood estimates demonstrated robustness to outliers.
• Nonparametric functions were successfully estimated using smoothing splines.
• The approach showed good performance in simulations and real-world data.

Conclusions:

• The extended model provides a flexible and robust approach for longitudinal censored data analysis.
• It is suitable for complex data structures and can accommodate non-Gaussian errors.
• The methodology is validated through simulations and application to acquired immunodeficiency syndrome (AIDS) clinical trial data.