A repeated measures model for analysis of continuous outcomes in sequential parallel comparison design studies

Area of Science:

• Clinical Trials Methodology
• Statistical Analysis in Psychiatry
• Biostatistics

Background:

• Sequential Parallel Comparison Design (SPCD) addresses high placebo response rates in clinical trials.
• SPCD, initially for binary outcomes, is now applied to continuous outcomes, common in fields like psychiatry.
• Existing analytic methods for SPCD continuous data include seemingly unrelated regression and ordinary least squares.

Purpose of the Study:

• To propose a novel repeated measures linear model for analyzing continuous data from SPCD trials.
• To develop a method that accounts for all collected outcome data, including data missing at random.
• To provide a robust statistical approach for hypothesis testing in SPCD trials.

Main Methods:

• A repeated measures linear model is proposed.
• The model incorporates all outcome data and handles missing data using an 'at random' assumption.
• A contrast is formulated post-model fitting for primary hypothesis testing.

Main Results:

• Simulations demonstrate the proposed model preserves Type I error rates, even with small sample sizes.
• The new approach offers adequate statistical power.
• The method yields the smallest mean squared error across various assumptions compared to existing techniques.

Conclusions:

• The proposed repeated measures linear model is a superior method for analyzing SPCD trial data with continuous outcomes.
• This approach offers improved accuracy, power, and efficiency over existing methods.
• Researchers conducting SPCD trials are recommended to consider this advanced statistical methodology.