Adaptive propensity score procedure improves matching in prospective observational trials

Table 2 Mean matching rate/mean lower CI limit of the matching rate at interim and final analysis for the naive approach and the resampling CI method for different numbers of patients in the control group

	Interim	Final	Interim	Final
n _control	Naiv		99%-CI
50	0.89	0.79	0.49	0.92
75	0.93	0.81	0.58	0.92
100	0.95	0.82	0.64	0.92
125	0.96	0.83	0.67	0.92
150	0.97	0.84	0.70	0.92
175	0.98	0.84	0.72	0.92
200	0.98	0.84	0.73	0.92
225	0.98	0.85	0.74	0.92
250	0.99	0.85	0.76	0.92
275	0.99	0.85	0.76	0.92
300	0.99	0.86	0.77	0.92
	95%-CI		90%-CI
50	0.54	0.91	0.56	0.90
75	0.62	0.91	0.64	0.90
100	0.67	0.91	0.68	0.91
125	0.70	0.91	0.71	0.91
150	0.72	0.91	0.73	0.91
175	0.74	0.91	0.75	0.91
200	0.75	0.91	0.76	0.91
225	0.76	0.91	0.77	0.91
250	0.77	0.91	0.78	0.91
275	0.78	0.91	0.79	0.91
300	0.79	0.91	0.80	0.91

The resampling CI method is applied for different confidence levels (α_CI∈{0.01,0.05,0.1})

ISSN: 1471-2288