. Author manuscript; available in PMC: 2020 May 13.

Published in final edited form as: Ann Intern Med. 2019 Oct 8;171(8):555–567. doi: 10.7326/M19-1152

Table 2.

Predicted significance vs observed significance for primary (or first listed, when primary was not specified) effect sizes among all included studies, and stratified by quality metric category.

		P < 0.05 threshold			P < 0.005 threshold
	Number of studies	Predicted significance	Observed significance	P-value	Predicted significance	Observed significance	P-value
Full cohort	142	62%	104 (73%)	0.011	46%	96 (68%)	<0.001
Best Practices	14	12%	9 (64%)	0.001	2%	8 (57%)	0.001
Communication	5	43%	3 (60%)	0.67	40%	3 (60%)	0.63
Medical Errors	32	50%	20 (62%)	0.169	33%	15 (47%)	0.182
Patient Outcomes	17	64%	9 (53%)	NP^a	54%	9 (53%)	NP^a
Quality & Safety	74	65%	62 (84%)	<0.001	50%	60 (81%)	<0.001

NP: not pertinent, observed smaller than predicted