Figure 3.
Power of APPLE, SEPPLE methods, and regular log-rank test ignoring the delayed treatment effect given sample size and hazard ratio under different values of proportions of subjects who can survive beyond the delayed phase p = 70%, 80%, and 90%. Total accrual duration A = 1 year; total study duration τ = 3 years; the treatment effect delayed duration t* = 6 months; Type I error rate α = 0.05.