Abstract
Evidence-based medicine uses a hierarchy of publication types according to their vulnerability to bias. A widely used measure of journal “quality” is its impact factor, which describes the citation rate of its publications. We investigated the relationship between impact factor for eight anaesthesia journals and publication type with respect to their level of evidence 1-4 using Spearman rank correlation (rho). There were 1418 original publications during 2001 included in the analysis. The number (%) of publication types according to evidence-based medicine level were: level 1: 6 (0.4%), level 2: 533 (38%) level 3: 329 (23%), level 4: 550 (39%). There was no correlation between journal ranking according to impact factor and publication type (rho=–0.03, P=0.25). The correlation between journal rank and the proportion of publications that were randomized trials was –0.35 (P<0.001). The correlation between journal rank and number of publications was 0.65 (P<0.001). The correlation between journal rank and number of level 1 or 2 studies was 0.58 (P<0.001). The overall level of evidence published in anaesthesia journals was high. Journal rank according to impact factor is related to the number of publications, but not the proportion of publications that are evidence-based medicine level 1 or 2.