Sample size and power issues in estimating incremental cost-effectiveness ratios from clinical trials data

Area of Science:

• Health Economics
• Clinical Trial Design
• Statistical Inference

Background:

• Prospective economic evaluations are increasingly integrated into randomized controlled trials.
• Trial-based cost-effectiveness analysis allows for statistical inference to quantify uncertainty in the incremental cost-effectiveness ratio (ICER).
• Existing literature compares various methods for calculating confidence intervals for the ICER.

Purpose of the Study:

• To address power and sample size considerations in trial-based cost-effectiveness analysis.
• To determine the sample size needed to achieve a confidence interval narrow enough to distinguish between cost-effective and non-cost-effective regions.
• To establish criteria for adequate sample size based on the cost-effectiveness plane.

Main Methods:

• Developing an approach for sample size calculation in trial-based cost-effectiveness analysis.
• Utilizing the cost-effectiveness plane to define regions of cost-effectiveness.
• Defining an ellipse on the cost-effectiveness plane to separate adequate from inadequate sample sizes.

Main Results:

• The proposed method links sample size to the precision of the ICER estimate.
• For a given sample size, the cost-effectiveness plane is divided into two regions by an ellipse.
• Adequate sample size is achieved when the true ICER lies on or outside this ellipse.

Conclusions:

• The study provides a framework for determining sample size in trial-based cost-effectiveness studies.
• This approach ensures sufficient statistical power to make definitive conclusions about cost-effectiveness.
• Accurate sample size calculation is crucial for reliable economic evaluations within clinical trials.