Figure 4
From: Testing the treatment effect on competing causes of death in oncology clinical trials
Empirical power and size of the tests in scenario 3. Empirical power (1 - β) and size (α) of the tests for CD and NCD in scenario 3 (HR CD = 0.8, HR NCD = 1). The data are simulated without censored observations. The plots on the first line concern CD, those on the second line NCD. The plots on the left are for data with correct causes of death (P(misclass) = 0%), those on the right for data with 20% of misclassified causes of death (P(misclass) = 20%). The bold grey horizontal line corresponds to the 0.05 level and ρ to the correlation.