Principal quantile treatment effect estimation using principal scores.

Area of Science:

• Biostatistics
• Clinical Trial Methodology
• Epidemiology

Background:

• Intercurrent events complicate the precise definition of treatment effects in clinical trials.
• Estimands and principal strata are crucial for accurately defining and analyzing treatment effects.
• Traditional analyses often yield biased results when dealing with intercurrent events.

Purpose of the Study:

• To propose novel principal quantile treatment effect estimators.
• To enable nonparametric estimation of potential outcome distributions using principal score weighting.
• To provide accurate treatment effect estimates in the presence of intercurrent events, without the exclusion restriction assumption.

Main Methods:

• Development of principal quantile treatment effect estimators.
• Application of principal score weighting for nonparametric estimation.
• Simulation studies to evaluate estimator performance.
• Illustration using data from a nonerosive reflux disease randomized controlled trial.

Main Results:

• The proposed method accurately estimates quantiles of outcomes within principal strata.
• Simulation studies confirm the validity of the estimators, especially when quantiles are preferred over means.
• The method effectively handles intercurrent events without bias.

Conclusions:

• Principal quantile treatment effect estimators offer a robust approach for analyzing treatment effects in randomized trials with intercurrent events.
• The proposed method is particularly useful when population-level summaries like medians or quantiles are of primary interest.
• This technique enhances the precision and reliability of treatment effect estimation in complex clinical trial settings.