Abstract

At first, we review a sample size estimation method for log-rank test in survival analysis. Although it is widely used these days, it has a weakness for practical use. We propose a modification method to avoid the weakness.