Table 2.

Evidence classification criteria^{46 47 48}

Evidence class	Description
Class I: convincing evidence	>1000 cases (or >20 000 participants for continuous outcomes); statistical significance at P<10−6 (random effects); no evidence of small study effects and excess significance bias; 95% prediction interval excluded null value; no large heterogeneity (I2<50%)
Class II: highly suggestive evidence	>1000 cases (or >20 000 participants for continuous outcomes), statistical significance at P<10−6 (random effects), and largest study with 95% confidence interval excluding null value
Class III: suggestive evidence	>1000 cases (or >20 000 participants for continuous outcomes) and statistical significance at P<0.001
Class IV: weak evidence	Remaining significant associations with P<0.05
NS: non-significant	P>0.05