On power and sample size calculation for QT studies with recording replicates at given time point

Area of Science:

• Pharmacology and Clinical Trials
• Biostatistics and Statistical Modeling

Background:

• Routine QT studies are crucial for drug safety assessment.
• Electrocardiogram (ECG) recording replicates are used to improve data reliability.
• Sample size and power calculations are critical for efficient study design.

Purpose of the Study:

• To examine the impact of ECG recording replicates on power and sample size calculations in QT studies.
• To derive formulas for sample size calculations considering covariates and different study designs.
• To propose an approach for optimizing the allocation of subjects and replicates under budget constraints.

Main Methods:

• Formulas for sample size calculations were derived for parallel-group and crossover designs.
• The impact of correlation between replicate ECGs on sample size was analyzed.
• An optimization approach was developed for subject and replicate allocation.

Main Results:

• Replicate ECGs can reduce required sample size when replicates are nearly independent (correlation near 0).
• Minimal sample size gains are observed when replicate ECGs are highly correlated (correlation near 1).
• An optimal allocation strategy was proposed to maximize power or minimize cost.

Conclusions:

• The use of replicate ECGs in QT studies can impact sample size and power calculations.
• The correlation between replicate ECGs is a key factor in determining the efficiency of this approach.
• An optimized subject and replicate allocation strategy can enhance study efficiency and cost-effectiveness.