Sample size and power calculations in repeated measurement analysis

Area of Science:

• Neuropsychopharmacology
• Clinical Trial Design
• Statistical Analysis

Background:

• Controlled clinical trials frequently use parallel-groups designs with repeated measurements.
• Hypotheses often focus on differences in change patterns or magnitudes from baseline.

Purpose of the Study:

• Introduce formulas and a computer program for sample size and power calculations.
• Facilitate two-stage analyses in clinical trials.

Main Methods:

• The study details a two-stage analytical approach.
• It incorporates three definitions of change and considers baseline scores as a covariate.
• It accounts for homogeneous or heterogeneous autoregressive correlation patterns.

Main Results:

• Formulas and a computer program are presented for sample size and power calculations.
• The program includes empirical adjustments for projected dropout rates.
• The methods accommodate various change definitions and correlation structures.

Conclusions:

• The developed tools support accurate sample size and power determination for neuropsychopharmacology trials.
• The program enhances the statistical rigor of analyzing repeated measures data.
• It offers flexibility for different analytical approaches and data characteristics.