Issues in performing a network meta-analysis

Senn, Stephen; Gavini, François; Magrez, David; Scheen, André

doi:10.1177/0962280211432220

Cited by 68 publications

(94 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This can be broadly related to the issue of publication bias, the concern that studies with significant results are more likely to be published, and published studies (especially those in the meta-analyst’s own language) are more likely to be included in an NMA [35]. Another limitation is that our focus here has been the detection of outlyingness under the Bayesian framework, but there is a large frequentist literature on the subject [8][9][10]. For instance, Senn et al [8] proposed frequentist statistical algorithms using the technique of Aitken estimator [36], and argued that it was inappropriate to treat the main effects of trial as random, and that the generalization from a classic random-effect meta-analysis to a network meta-analysis involved strong assumptions about the variance components.…”

Section: Discussion and Future Workmentioning

confidence: 99%

“…Another limitation is that our focus here has been the detection of outlyingness under the Bayesian framework, but there is a large frequentist literature on the subject [8][9][10]. For instance, Senn et al [8] proposed frequentist statistical algorithms using the technique of Aitken estimator [36], and argued that it was inappropriate to treat the main effects of trial as random, and that the generalization from a classic random-effect meta-analysis to a network meta-analysis involved strong assumptions about the variance components. Our detection measures RD (ARD) will still work in a frequentist setting, but STR (ASTR), Bayesian p -values and SMN are inapplicable.…”

Section: Discussion and Future Workmentioning

confidence: 99%

See 1 more Smart Citation

Detecting outlying trials in network meta‐analysis

Zhang

Carlin

2015

Statistics in Medicine

View full text Add to dashboard Cite

Summary Network meta-analysis (NMA) expands the scope of a conventional pairwise meta-analysis to simultaneously handle multiple treatment comparisons. However, some trials may appear to deviate markedly from the others, and thus be inappropriate to be synthesized in the NMA. In addition, the inclusion of these trials in evidence synthesis may lead to bias in estimation. We call such trials trial-level outliers. To the best of our knowledge, while heterogeneity and inconsistency in NMA have been extensively discussed and well addressed, few previous papers have considered the proper detection and handling of trial-level outliers. In this paper, we propose several Bayesian outlier detection measures, which are then applied to a diabetes data set. Simulation studies comparing our approaches in both arm- and contrast-based model settings are provided in two supporting appendices.

show abstract

Section: Discussion and Future Workmentioning

confidence: 99%

Section: Discussion and Future Workmentioning

confidence: 99%

Detecting outlying trials in network meta‐analysis

Zhang

Carlin

2015

Statistics in Medicine

View full text Add to dashboard Cite

show abstract

“…When summary measures are available, it is customary to model the response by a linear (mixed) model assuming normality and accounting for possible heterogeneity in precision by weighting. In the diabetes example by Senn et al [7], we have at our disposal mean responses per treatment and trial as well as the associated sample standard deviations and sample sizes, from which the variance of a mean can be computed. Thus, the models used for our analyses are of the form…”

Section: Methodsmentioning

confidence: 99%

“…Specifically, we will explore ways to detect inconsistency using standard procedures for linear models available in most statistical packages. The methods will be illustrated using the diabetes example published by Senn et al [7]. This example has also been used by Krahn et al [1] to illustrate their proposed methods for detection of inconsistency using a baseline contrast parameterization, so our results can be compared directly to that paper in order to appreciate the degree of agreement between both model formulations and the resulting tests and diagnostic checks for inconsistency.…”

Section: Introductionmentioning

confidence: 99%

Network-meta analysis made easy: detection of inconsistency using factorial analysis-of-variance models

Piepho

2014

BMC Med Res Methodol

View full text Add to dashboard Cite

BackgroundNetwork meta-analysis can be used to combine results from several randomized trials involving more than two treatments. Potential inconsistency among different types of trial (designs) differing in the set of treatments tested is a major challenge, and application of procedures for detecting and locating inconsistency in trial networks is a key step in the conduct of such analyses.MethodsNetwork meta-analysis can be very conveniently performed using factorial analysis-of-variance methods. Inconsistency can be scrutinized by inspecting the design × treatment interaction. This approach is in many ways simpler to implement than the more common approach of using treatment-versus-control contrasts.ResultsWe show that standard regression diagnostics available in common linear mixed model packages can be used to detect and locate inconsistency in trial networks. Moreover, a suitable definition of factors and effects allows devising significance tests for inconsistency.ConclusionFactorial analysis of variance provides a convenient framework for conducting network meta-analysis, including diagnostic checks for inconsistency.

show abstract

“…7 A sparsely connected network of 10 treatments for the treatment of diabetes is used by Senn et al to make points about approaches to analysis. 8 Graphical approaches, both of the network and of the results, are again described to summarise the data.…”

mentioning

confidence: 99%

Meta-analysis in clinical research

Julious

2013

Stat Methods Med Res

View full text Add to dashboard Cite

Meta-analysis is important in the evaluation of therapies 1,2 for integrating the results of individual studies in order to advance or inform clinical research. In this special issue, different aspects of metaanalysis are reviewed using case studies to illustrate the methodologies being described.There can be issues when evaluating evidence when the events being observed are rare. 3 Lane compares available methods for binary data considering different summary measures such as riskdifference, relative-risk and odds-ratio scale as well as fixed and random-effect methods in the assessment. The paper also highlights the benefit of how graphical approaches can add value. 4A case study is presented by Julious to highlight issues in evaluating evidence for safety and efficacy, which updates previous analyses with additionally reported trials.5 These data also consider the problem of the event of interest being rare. Graphical approaches are highlighted to investigate the assumptions in a meta-analysis. 6 Often, however, a direct comparison of two therapies is not possible and indirect comparisons need to be made through a network meta-analysis. 7 A sparsely connected network of 10 treatments for the treatment of diabetes is used by Senn et al. to make points about approaches to analysis.8 Graphical approaches, both of the network and of the results, are again described to summarise the data.Indirect comparisons are also used in the assessment, and design, of non-inferiority studies. 9,10In non-inferiority clinical trials, a test treatment is compared to an active-control rather than to placebo -when randomising to placebo is unethical or not feasible. A critical question is whether the test treatment would have been superior to placebo, had placebo been used in the non-inferiority trial. 11,12 This question, as highlighted by Schmidli et al., 13 can only be addressed indirectly, based on information from relevant historical trials with data on active-control and placebo.As highlighted multivariate meta-analysis is becoming increasingly popular. The advantages and limitations of multivariate meta-analysis have been discussed.14 The main limitation being computational complexity and hence, in a timely paper, Mavridis and Salanti 15 review the statistical methods and the related software for multivariate meta-analysis.Pooling evidence from individual trials is important in the evaluation of research. For the individual trials adaptive designs have been recommended. 16,17 However, there can be issues in pooling evidence when trials have stopped early. 18,19 Bassler et al. 20 review controversies associated with randomised controlled trials stopped early for apparent benefit and discussed how pooled effects from meta-analyses including trials stopped early could potentially overestimate an effect.

show abstract

Issues in performing a network meta-analysis

Cited by 68 publications

References 37 publications

Detecting outlying trials in network meta‐analysis

Detecting outlying trials in network meta‐analysis

Network-meta analysis made easy: detection of inconsistency using factorial analysis-of-variance models

Meta-analysis in clinical research

Contact Info

Product

Resources

About