An additive-multiplicative model for longitudinal data with informative observation times.

Area of Science:

• Biostatistics
• Longitudinal Data Analysis
• Clinical Trial Methodology

Background:

• Standard statistical models (e.g., generalized linear mixed models) are often insufficient for complex longitudinal data.
• Challenges include informative observation processes and non-additive patient influences on outcomes.
• Existing methods may not fully capture the interplay between outcome and observation dynamics.

Purpose of the Study:

• To extend standard longitudinal models to handle informative observation processes.
• To incorporate non-additive (multiplicative) effects of patient characteristics on outcomes.
• To develop a flexible modeling framework for complex longitudinal data analysis.

Main Methods:

• Proposed a novel modeling structure with additive-multiplicative components.
• Developed theoretical underpinnings for statistical inference within this new framework.
• Validated the method through simulation studies and application to a real-world observational study.

Main Results:

• The proposed additive-multiplicative models effectively accommodate informative observation processes.
• The method demonstrated good performance in finite-sample simulation scenarios.
• Successfully applied the model to analyze data from an alcohol-associated hepatitis study.

Conclusions:

• The developed flexible longitudinal models offer a robust approach for analyzing complex clinical data.
• This framework enhances the ability to draw valid inferences when standard assumptions are violated.
• The methodology provides valuable tools for observational studies and clinical trial data analysis.