TABLE 2.

Subdistribution hazard ratio and odds ratio (OR) at time 28, and event‐specific hazard ratio with respect to recovery derived from cumulative event probabilities under the constant event‐specific hazard assumption and resulting sample size when chosen as parameter for study planning with two‐sided type I error rate of 0.05 and power 0.8. Sample sizes $N_{ES}$ and $N_{SD}$ were calculated using Equations (13) and (14) and $N_{OR}$ was calculated using Equation (2) in Hsieh et al. (1998), all with $p=0.5$

				${\theta }_{ES}$	$N_{ES}$	${\theta }_{ES-CE}$	${\theta }_{SD}(28)$	$N_{SD}$	OR (28)	$N_{OR}$
0.7	0.55	0.10	0.10	1.59	237	1.25	1.51	300	1.91	325
0.7	0.55	0.15	0.15	1.65	200	1.30	1.51	300	1.91	325
0.7	0.55	0.20	0.20	1.76	157	1.38	1.51	300	1.91	325
0.7	0.55	0.10	0.20	1.39	474	0.54	1.51	300	1.91	325
0.7	0.55	0.15	0.20	1.54	274	0.91	1.51	300	1.91	325