Bayesian semiparametric growth models for measurement error and missing data in CD4/CD8 ratio: Application to AIDS...

Area of Science:

• Biostatistics
• Clinical Epidemiology
• Immunology

Background:

• Assessing immune recovery after antiretroviral therapy (ART) is crucial for HIV/AIDS management.
• Time-varying predictors, such as the CD4/CD8 ratio, are key indicators but often suffer from measurement errors and missing values.
• Traditional statistical methods may yield biased results when ignoring these data complexities.

Purpose of the Study:

• To develop robust statistical models for analyzing immune recovery in HIV patients undergoing ART.
• To address challenges posed by measurement errors and missing data in time-varying predictors like the CD4/CD8 ratio.
• To differentiate between patients progressing to AIDS and those who are not, using a Bayesian approach.

Main Methods:

• Introduction of semiparametric mixed-effect models designed to handle measurement errors and missing values in predictors.
• Development of a fully Bayesian framework for model fitting and parameter estimation.
• Application of the models to discriminate between potential progressors and non-progressors to AIDS.

Main Results:

• The proposed semiparametric models effectively account for measurement errors and missing data, reducing bias in the analysis of immune recovery.
• The Bayesian approach successfully discriminates between patient groups based on their risk of progressing to AIDS.
• Demonstrated utility of the methods using real-world data from an AIDS clinical study.

Conclusions:

• Semiparametric mixed-effect models offer a flexible and reliable approach for analyzing complex longitudinal data in HIV/AIDS research.
• The Bayesian methodology provides a powerful tool for risk stratification and understanding disease progression.
• Accurate analysis of immune recovery markers is essential for optimizing patient management and treatment strategies in HIV care.