Bias in meta-analysis detected by a simple, graphical test

TLDR

Funnel plots, which plot effect estimates against sample size, are used to detect bias in meta‑analyses that may later be contradicted by large trials. The study examined whether asymmetry in funnel plots predicts discordance between meta‑analyses and large trials and assessed bias prevalence in published meta‑analyses. The authors searched Medline for pairs of a meta‑analysis and a single large trial, defined concordance as same direction and within 30 % of the trial estimate, and measured funnel‑plot asymmetry via the intercept from a regression of standard normal deviates against precision. Among eight meta‑analysis/large‑trial pairs, four were concordant and four discordant; discordant pairs showed larger effects and funnel‑plot asymmetry in three of four, whereas concordant pairs showed none; funnel‑plot asymmetry indicated bias in 38 % of journal meta‑analyses and 13 % of Cochrane reviews, suggesting that a simple funnel‑plot test can flag potential bias but is limited when few small trials are included.

Abstract

<h3>Abstract</h3> <b>Objective:</b> Funnel plots (plots of effect estimates against sample size) may be useful to detect bias in meta-analyses that were later contradicted by large trials. We examined whether a simple test of asymmetry of funnel plots predicts discordance of results when meta-analyses are compared to large trials, and we assessed the prevalence of bias in published meta-analyses. <b>Design:</b> Medline search to identify pairs consisting of a meta-analysis and a single large trial (concordance of results was assumed if effects were in the same direction and the meta-analytic estimate was within 30% of the trial); analysis of funnel plots from 37 meta-analyses identified from a hand search of four leading general medicine journals 1993-6 and 38 meta-analyses from the second 1996 issue of the <i>Cochrane Database of Systematic Reviews</i>. <b>Main outcome measure:</b> Degree of funnel plot asymmetry as measured by the intercept from regression of standard normal deviates against precision. <b>Results:</b> In the eight pairs of meta-analysis and large trial that were identified (five from cardiovascular medicine, one from diabetic medicine, one from geriatric medicine, one from perinatal medicine) there were four concordant and four discordant pairs. In all cases discordance was due to meta-analyses showing larger effects. Funnel plot asymmetry was present in three out of four discordant pairs but in none of concordant pairs. In 14 (38%) journal meta-analyses and 5 (13%) Cochrane reviews, funnel plot asymmetry indicated that there was bias. <b>Conclusions:</b> A simple analysis of funnel plots provides a useful test for the likely presence of bias in meta-analyses, but as the capacity to detect bias will be limited when meta-analyses are based on a limited number of small trials the results from such analyses should be treated with considerable caution. <h3>Key messages</h3> Systematic reviews of randomised trials are the best strategy for appraising evidence; however, the findings of some meta-analyses were later contradicted by large trials Funnel plots, plots of the trials9 effect estimates against sample size, are skewed and asymmetrical in the presence of publication bias and other biases Funnel plot asymmetry, measured by regression analysis, predicts discordance of results when meta-analyses are compared with single large trials Funnel plot asymmetry was found in 38% of meta-analyses published in leading general medicine journals and in 13% of reviews from the <i>Cochrane Database of Systematic Reviews</i> Critical examination of systematic reviews for publication and related biases should be considered a routine procedure