Risk factors for prostate cancer: An umbrella review of prospective observational studies and mendelian randomization analyses.

Briefly, robust evidence for causality was assigned based on nominally significant P value and directional concordant effect across all methods performed; probable evidence was assigned based on nominally significant P value in at least 1 method (main or sensitivity analyses) and concordant effect direction among all methods performed; suggestive evidence was assigned when at least 1 method had a nominally significant P value but the direction of the effect estimates differed between methods; insufficient evidence was assigned for significant associations based only on 1 main analysis while no sensitivity analysis was available ( Table 1 ).

also in 7,732 other papers