Positive-outcome bias: comparison of emergency medicine and general medicine literatures.

Abstract

The existence of positive-outcome bias in the medical literature is well established. Positive-outcome bias in two emergency medicine journals was compared with that found in two general medicine journals. Published original contributions from Annals of Emergency Medicine, American Journal of Emergency Medicine, JAMA, and New England Journal of Medicine were reviewed. Articles were categorized as demonstrating a positive or negative outcome or showing no difference using new criteria. Descriptive articles were excluded. Of 700 articles reviewed, 177 emergency medicine and 211 general medicine articles met the study criteria. The emergency medicine journals had 142 articles (80%) with positive outcomes, 27 (15%) with negative outcomes, and 8 (5%) with no difference. The general medicine journals had 169 articles (80%) with positive outcomes, 33 (16%) with negative outcomes, and 9 (4%) with no difference. There was no significant difference between journal groups (chi-square; p = 0.99). The power of the study was 0.80 to detect a difference of 15% between groups with alpha set at 0.05. There was no significant difference in the proportions of positive-outcome studies published in this sample of the emergency medicine literature compared with the general medicine literature. The potential impact of positive-outcome bias and methods of dealing with the problem are reviewed.