Semiparametric Trend Analysis for Stratified Recurrent Gap Times Under Weak Comparability Constraint

Area of Science:

• Biostatistics
• Longitudinal Data Analysis
• Survival Analysis

Background:

• Recurrent event data, common in longitudinal studies, presents analytical challenges due to multiple events per individual.
• Existing methods like Wang and Chen's (2000) comparability constraint impose independence, leading to potential information loss.
• The accelerated failure time model is a framework for analyzing time-to-event data.

Purpose of the Study:

• To propose a novel comparability constraint for recurrent event data analysis.
• To overcome the limitations of existing methods, specifically the unnecessary independence assumption.
• To improve the efficiency and reduce the variance of time trend estimations in longitudinal studies.

Main Methods:

• Developed a new comparability constraint for gap time pairs under the accelerated failure time model.
• The proposed constraint allows for dependent gap time pairs while maintaining the same conditional distribution.
• Compared the performance of the new method against Wang and Chen's (2000) estimator through simulations.

Main Results:

• The proposed comparability constraint utilizes a larger set of gap time data pairs compared to the strong comparability method.
• Simulation studies demonstrated that the new estimator has a smaller variance than Wang and Chen's (2000) estimator.
• The method was successfully applied to real-world data from the HIV Prevention Trial Network 052 study.

Conclusions:

• The new comparability constraint offers a more efficient approach to analyzing recurrent event data.
• By relaxing the independence assumption, the method reduces information loss and improves estimation accuracy.
• This advancement has practical implications for analyzing longitudinal data in various research fields, including clinical trials.