Table 1.

Clinical significance and precision of the log-hazard ratio according to the initial and final sample sizes

Δ	log e (Δ)	b λ c	N		Pr( LCL > δ /2 | H 1 )	e CIW 1	Pr( UCL < δ /2 | H 0 )	d CIW 0
1.25	0.22	1.12	aInitial	1264	0.2925	0.322	0.2651	0.314
			cFinal	5402	0.8241	0.155	0.8016	0.151
1.50	0.41	1.22	aInitial	384	0.2759	0.602	0.2658	0.577
			cFinal	1694	0.8349	0.285	0.8039	0.273
1.75	0.56	1.32	aInitial	204	0.2766	0.850	0.2536	0.804
			cFinal	938	0.8496	0.392	0.8021	0.371
2.00	0.69	1.41	aInitial	132	0.2700	1.087	0.2344	1.018
			cFinal	632	0.8503	0.487	0.8052	0.457

The ^ainitial N calculated using equation (5), Schoenfeld’s [14] formula, is the total sample size required to detect a hazard ratio Δ at the 5% level with 80% power, assuming equal subject allocation and a 0.5 overall censoring proportion. ^b λ_c is the hazard rate for the exponential censoring time given by equation (7), and δ. = log_e(Δ). The ^cfinal N is the total sample size such that both Pr(LCL > δ /2 | H ₁) and Pr(UCL < δ /2 | H ₀) are at least 0.8 as estimated by the proportion of times LCL and UCL are bounded by δ /2 in 10,000 iterations. ^dCIW₀ and ^eCIW₁ are the mean width of the 95% confidence intervals under H ₀ and H ₁, respectively.