TABLE 1.

For Scenario 1 with ordered categorical outcome, performance of estimators for $\beta$ under (a) the null hypothesis $\beta =0$ and (b) the alternative $\beta =\log (1.5)=0.405$ at each interim analysis time $(t_1,\dots ,t_4)=(150,195,240,285)$ days and at the final analysis at $t_{end}=330$ days

	MC mean	MC SD	Ave MC SE	MSE ratio	MC mean	MC SD	Ave MC SE	MSE ratio
	(a) Null hypothesis
	${\widehat{\beta }}_{{𝒯}_F}(t)$				${\widehat{\beta }}_{IPW}(t)$
$t_1$	$-$0.002	0.294	0.294	1.000	$-$0.004	0.232	0.232	1.603
$t_2$	$-$0.002	0.221	0.221	1.000	$-$0.002	0.189	0.189	1.330
$t_3$	$-$0.003	0.184	0.185	1.000	$-$0.002	0.166	0.164	1.239
$t_4$	$-$0.001	0.162	0.162	1.000	$-$0.001	0.152	0.151	1.139
$t_{end}$	0.000	0.146	0.146	1.000	0.000	0.147	0.146	0.991
	${\widehat{\beta }}_{AIPW1}(t)$				${\widehat{\beta }}_{AIPW2}(t)$
$t_1$	$-$0.004	0.221	0.221	1.775	$-$0.005	0.203	0.198	2.095
$t_2$	$-$0.002	0.178	0.178	1.534	$-$0.002	0.168	0.165	1.717
$t_3$	$-$0.002	0.156	0.154	1.399	$-$0.002	0.149	0.145	1.542
$t_4$	$-$0.001	0.141	0 .140	1.327	$-$0.001	0.138	0.136	1.380
$t_{end}$	0.000	0.135	0.135	1.169	0.000	0.135	0.135	1.169

	(b) Alternative hypothesis
	${\widehat{\beta }}_{{𝒯}_F}(t)$				${\widehat{\beta }}_{IPW}(t)$
$t_1$	0.408	0.294	0.294	1.000	0.406	0.235	0.235	1.566
$t_2$	0.406	0.220	0.221	1.000	0.406	0.191	0.191	1.336
$t_3$	0.404	0.185	0.185	1.000	0.405	0.167	0.165	1.221
$t_4$	0.406	0.163	0.162	1.000	0.406	0.153	0.152	1.131
$t_{end}$	0.406	0.147	0.146	1.000	0.406	0.148	0.147	0.985

	${\widehat{\beta }}_{AIPW1}(t)$				${\widehat{\beta }}_{AIPW2}(t)$
$t_1$	0.405	0.224	0.224	1.733	0.406	0.204	0.200	2.078
$t_2$	0.406	0.180	0.180	1.508	0.408	0.169	0.167	1.702
$t_3$	0.405	0.158	0.155	1.378	0.408	0.150	0.147	1.523
$t_4$	0.406	0.142	0.141	1.314	0.407	0.139	0.137	1.373
$t_{end}$	0.406	0.137	0.136	1.159	0.406	0.137	0.136	1.159

Note: MC mean is the mean of 10 000 Monte Carlo estimates; MC SD is the Monte Carlo SD, Ave MC SE is the mean of Monte Carlo SEs, and MSE ratio is the ratio of Monte Carlo mean square error for ${\widehat{\beta }}_{{𝒯}_F}(t)$ divided by that for the indicated estimator.