Table 1.

Results of simulations using clinical and administrative data to estimate the sample size needed to detect a 25% reduction in the suicide attempt rate in the 18 months following randomization.

Number per arm	Number of sites	Percent with prior PHQ item 9 response of 2 or 3	Event rate, usual care arm	Event rate, intervention arm	Power, unadjusted	Power, adjusted for covariates
5000	1	12.7	0.048	0.036	0.76	0.76
6000	1	11.4	0.048	0.036	0.83	0.83
7000	1	10.5	0.048	0.036	0.88	0.88
8000	1	9.8	0.048	0.036	0.91	0.91
9000	1	9.3	0.048	0.036	0.95	0.95
10000	1	8.9	0.048	0.036	0.96	0.96
5000	2	20.4	0.048	0.036	0.76	0.76
6000	2	17.8	0.048	0.036	0.83	0.83
7000	2	16.0	0.048	0.036	0.88	0.88
8000	2	14.6	0.048	0.036	0.93	0.92
9000	2	13.6	0.048	0.036	0.95	0.95
10000	2	12.7	0.048	0.036	0.96	0.97
5000	3	28.1	0.048	0.036	0.76	0.76
6000	3	24.5	0.048	0.036	0.83	0.83
7000	3	21.4	0.048	0.036	0.89	0.88
8000	3	19.4	0.048	0.036	0.93	0.93
9000	3	17.8	0.048	0.036	0.94	0.94
10000	3	16.5	0.048	0.036	0.97	0.97