Table 1.

Simulation results for WMR and comparison with semi-parametric method. We assumed (log(T),M)∼N₂(0,0,1,1, − 0.7). An independent censoring time was generated such that either 20% or 40% subjects were censored. The results are based on 1000 such simulations. The MLE and semi-parametric estimates are cited from Heagerty and Zheng (2005). MCSD denotes the Monte-Carlo SD. The nominal coverage is 95.0. Estimate of integrated AUC (iAUC) and its variance is based on 10 equi-spaced quantiles of time. The variance of integrated AUC is estimated assuming a constant weight

		MLE		Semi-parametric		WMR (n = 200)
Log time	AUC(t)	Mean	SD	Mean	SD	Mean	MCSD	EstSD	Coverage
20% censoring
− 2.0	0.884	0.884	0.018	0.881	0.044	0.876	0.055	0.050	90.2
− 1.5	0.833	0.834	0.019	0.829	0.035	0.828	0.040	0.041	93.4
− 1.0	0.782	0.782	0.019	0.771	0.033	0.780	0.037	0.036	93.2
− 0.5	0.734	0.734	0.019	0.720	0.033	0.734	0.035	0.034	94.1
0.0	0.693	0.693	0.018	0.686	0.034	0.695	0.037	0.037	94.0
0.5	0.660	0.660	0.016	0.657	0.040	0.664	0.047	0.045	94.9
1.0	0.634	0.634	0.015	0.637	0.041	0.638	0.066	0.064	93.0
iAUC	0.741	0.741	0.016	0.740	0.018	0.738	0.020	0.017	89.6
40% censoring
− 2.0	0.884	0.884	0.019	0.875	0.048	0.876	0.056	0.050	89.9
− 1.5	0.833	0.834	0.021	0.827	0.037	0.828	0.041	0.042	92.9
− 1.0	0.782	0.782	0.021	0.772	0.035	0.781	0.038	0.037	92.7
− 0.5	0.734	0.734	0.020	0.722	0.039	0.735	0.039	0.038	94.3
0.0	0.693	0.693	0.019	0.687	0.041	0.696	0.044	0.043	93.5
0.5	0.660	0.660	0.018	0.655	0.043	0.666	0.061	0.059	92.6
1.0	0.634	0.635	0.016	0.637	0.048	0.640	0.108	0.100	91.8
iAUC	0.741	0.741	0.017	0.742	0.021	0.739	0.023	0.019	88.3