Inverse-Intensity-Weighted Generalized Estimating Equations With Irregularly Measured Longitudinal Data and...

Area of Science:

• Biostatistics
• Longitudinal Data Analysis
• Biomedical Research

Background:

• Longitudinal data are crucial in biomedical research, often collected at scheduled visits.
• Irregularities in visit schedules and participant dropout are common challenges.
• Existing methods like inverse intensity weighting generalized estimating equations (IIW-GEE) may be biased if dropout is related to the outcome.

Purpose of the Study:

• To extend the IIW-GEE framework to account for informative dropout in longitudinal studies.
• To reduce bias in parameter estimation when participant dropout is influenced by the health outcome.
• To provide a more accurate analysis of longitudinal biomedical data.

Main Methods:

• Developed an extended IIW-GEE framework to adjust for informative dropout.
• Utilized simulation studies to evaluate the performance of the proposed method.
• Applied the method to real-world clinical trial data (STAR*D).

Main Results:

• The extended IIW-GEE framework significantly reduced bias caused by informative dropout.
• Simulation studies confirmed the effectiveness of the adjusted method.
• Analysis of STAR*D data revealed that disease trajectories were overestimated when informative dropout was ignored.

Conclusions:

• The proposed extension to IIW-GEE provides a robust method for analyzing longitudinal data with informative dropout.
• Accounting for informative dropout is essential for accurate estimation of health trajectories in biomedical research.
• This method improves the reliability of findings from studies like the STAR*D trial.