Comparison of Effect Sizes Associated With Biomarkers Reported in Highly Cited Individual Articles and in Subsequent Meta-analyses

摘要: Context Many biomarkers are proposed in highly cited studies as determinants of disease risk, prognosis, or response to treatment, but few eventually transform clinical practice. Objective To examine whether the magnitude effect sizes is accurate overestimated. Data Sources We searched ISI Web Science and MEDLINE until December 2010. Study Selection included biomarker that had a relative risk presented their abstract. Eligible articles were those received more than 400 citations been published any 24 biomedical journals. also for subsequent meta-analyses on same associations (same outcome). Extraction In studies, data extraction was focused disease/outcome, under study, first reported From each meta-analysis, we extracted overall largest study. performed independently by 2 investigators. Results evaluated 35 associations. For 30 (86%), stronger estimate study; 3 study only twice size 29 (83%) corresponding meta-analysis found smaller estimate. Only 15 nominally statistically significant based 7 point greater 1.37. Conclusion Highly often report larger estimates postulated evaluating