Investigating underlying risk as a source of heterogeneity in meta-analysis

Abstract

In a meta-analysis of clinical trials, an important issue is whether the treatment benefit varies according to the underlying risk of the patients in the different trials. The usual naive analyses employed to investigate this question use either the observed risk of events in the control groups, or the average risk in the control and treatment groups, as a measure of underlying risk. These analyses are flawed and can produce seriously misleading results. We show how their biases depend on three components of variability, the within-trial and between-trial variances of the control group risks, and the between-trial variance of the treatment effects. We propose a Bayesian solution to the problem which can be carried out using the BUGS implementation of Gibbs sampling. The analysis is illustrated for a meta-analysis of bleeding and mortality data in trials of sclerotherapy for patients with cirrhosis, and the results contrasted with those from the naive approaches. Comparisons with other methods recently proposed for this problem are also made. We conclude that the Bayesian solution presented in this paper is not only more appropriate than other proposed methods, but is also sufficiently easy to implement that it can be used by applied researchers undertaking meta-analyses.

Comparison of Certolizumab Pegol with Other Anticytokine Agents for Treatment of Rheumatoid Arthritis: A Multiple-treatment Bayesian Metaanalysis

“…Since RA is a progressive disease, disease duration is linked in some way to disease severity [as attested by the related differences in C-reactive protein (CRP) level], and this source of heterogeneity between studies performed at different times should be taken into account when assessing the relative efficacy of treatments in comparison with a common comparator (in our case, placebo) 41 . Such declines in time of disease duration at entry to clinical studies of antirheumatic biotherapies probably reflect the progressively increasing confidence of physicians in the safety of these novel treatments, while being also a probable consequence of international recommendations encouraging institution of active disease management as soon as possible after RA diagnosis 42,43 .…”

A comparison of the active surveillance of scrapie in the European Union

“…The two types of heterogeneity that can be found in a meta-analysis have been described previously [15]: the effect heterogeneity resulting from the different interventions across the studies (countries in our case) and the baseline heterogeneity. The latter is a well-known and commonly reported source of variability in the meta-analysis literature [19,21]. Baseline heterogeneity, in our setting, arises when the risks of scrapie observed in the abattoir survey vary in a significant manner across the EU countries.…”

“…Meta-analysis methodologies are well suited to: (1) demonstrate the occurrence of heterogeneity in the way scrapie surveillance is conducted across the EU, and (2) evaluate the performance of the two surveys in the detection of the two types of scrapie. The explanation of the heterogeneity between studies is a logical step and one of increasing importance when conducting meta-analyses [17,19]. For the application of these techniques, we need some measure of effect: odds-ratios, risk-ratios or rateratios.…”

Bayesian statistics in medicine: a 25 year review

“…Heterogeneity was a developing theme, starting with testing for heterogeneity using Bayes factors [308]. A paper on underlying risk as a source of heterogeneity in meta-analysis in trials of sclerotherapy for cirrhosis continued this theme, but used the freely available general purpose package BUGS which the authors claimed made the technology 'realistically available to applied researchers undertaking meta-analysis' [309]. This work was then developed further [310] with a tutorial on meta-analysis comparing different approaches [311] and a method comparison for explaining heterogeneity in meta analysis [312].…”