The Kaplan-Meier Method for Estimating and Comparing Proportions in a Randomized Controlled Trial with Dropouts

Area of Science:

• Biostatistics
• Clinical Trials
• Epidemiology

Background:

• Estimating event proportions in randomized controlled trials (RCTs) is crucial for assessing treatment efficacy.
• Standard methods like intent-to-treat and completers-only can be biased or inefficient, particularly with participant dropout.

Purpose of the Study:

• To propose and evaluate the Kaplan-Meier method for estimating and comparing event proportions in RCTs.
• To address the limitations of traditional methods in the presence of participant attrition.

Main Methods:

• Application of the Kaplan-Meier method, typically used in survival analysis, to a non-survival endpoint setting.
• Extensive simulation studies to compare the performance (bias and efficiency) of Kaplan-Meier against standard methods.
• Demonstration of methods for single-sample proportion estimation and two-sample proportion comparison.

Main Results:

• Simulation studies indicate that the Kaplan-Meier method exhibits reduced bias and increased efficiency compared to intent-to-treat and completers-only methods.
• The proposed method effectively estimates and compares proportions in both single and two-sample scenarios.

Conclusions:

• The Kaplan-Meier method provides a statistically robust and more accurate approach for estimating and comparing event proportions in RCTs, especially when participant dropout occurs.
• This method enhances the reliability of trial results, as demonstrated by its application to a Parkinson's disease trial dataset.