Bayesian inference on mixed-effects varying-coefficient joint models with skew- t distribution for longitudinal data...

Area of Science:

• Biostatistics
• Epidemiology
• Clinical Trials

Background:

• HIV viral load and CD4 cell counts are critical biomarkers in AIDS clinical studies.
• The relationship between these biomarkers is inverse but time-varying.
• Existing statistical models often fail to account for real-world data complexities.

Purpose of the Study:

• To develop a joint statistical model for analyzing the time-varying relationship between HIV viral load and CD4 cell counts.
• To address challenges including asymmetric data, limit of detection (LOD), covariate measurement error, and missing data.
• To provide a robust modeling framework for AIDS clinical studies.

Main Methods:

• A mixed-effects varying-coefficient model was adapted to capture time-varying relationships.
• The model incorporates asymmetric distributions, LOD data, covariate measurement error, and missing data.
• Bayesian inference procedures were developed for parameter estimation.

Main Results:

• The proposed joint model effectively handles complex data issues common in AIDS clinical studies.
• The model provides accurate estimation of the time-varying relationship between HIV viral load and CD4 cell counts.
• Comparisons with other models demonstrated the superiority of the proposed approach.

Conclusions:

• The developed joint model offers a more realistic and accurate approach to analyzing AIDS clinical trial data.
• This methodology improves the understanding of HIV disease progression and treatment response.
• The Bayesian inference procedure provides a reliable tool for parameter estimation in complex biomedical data.