TABLE 6.

Results for difference in RMST at 14 days estimand in hospitalized population, when the adjusted estimator uses all six baseline variables from Section 4.1.3

Sample size	Estimator type	Effect	P(reject H 0)	MSE	Bias	Variance	Rel. Eff.
100	Unadjusted	0.000	0.011	76.296	0.014	76.304	1.000
100	Adjusted	0.000	0.035	73.480	0.013	73.488	0.963
100	Unadjusted	0.507	0.025	67.882	−0.938	67.008	1.000
100	Adjusted	0.507	0.063	62.857	−0.668	62.418	0.926
100	Unadjusted	1.004	0.087	53.738	−2.030	49.622	1.000
100	Adjusted	1.004	0.154	50.988	−1.804	47.738	0.949
200	Unadjusted	0.000	0.044	95.651	−0.131	95.644	1.000
200	Adjusted	0.000	0.055	85.260	−0.176	85.238	0.891
200	Unadjusted	0.507	0.108	80.512	−0.332	80.410	1.000
200	Adjusted	0.507	0.131	71.970	−0.187	71.943	0.894
200	Unadjusted	1.004	0.330	62.739	−1.014	61.718	1.000
200	Adjusted	1.004	0.399	56.397	−0.770	55.810	0.899
500	Unadjusted	0.000	0.051	100.299	−0.042	100.307	1.000
500	Adjusted	0.000	0.054	83.466	−0.008	83.474	0.832
500	Unadjusted	0.507	0.226	87.159	0.085	87.160	1.000
500	Adjusted	0.507	0.274	71.673	0.155	71.656	0.822
500	Unadjusted	1.004	0.735	72.850	−0.032	72.856	1.000
500	Adjusted	1.004	0.816	62.236	0.150	62.220	0.854
1000	Unadjusted	0.000	0.052	99.702	0.113	99.700	1.000
1000	Adjusted	0.000	0.053	81.859	0.144	81.846	0.821
1000	Unadjusted	0.507	0.411	87.420	0.282	87.349	1.000
1000	Adjusted	0.507	0.492	71.611	0.329	71.510	0.819
1000	Unadjusted	1.004	0.958	76.466	0.282	76.394	1.000
1000	Adjusted	1.004	0.980	63.461	0.360	63.339	0.830

Confidence intervals and hypothesis tests are Wald‐based. “Effect” denotes the true estimand value; “MSE” denotes mean squared error; “Rel. Eff.” denotes relative efficiency, which we approximate as the ratio of the MSE of the estimator under consideration to the MSE of the unadjusted estimator. In each block of four rows, the first two rows involve no treatment effect and the last two rows involve a benefit from treatment. MSE and variance are scaled by n; bias is scaled by $n^{1/2}$.