Skip navigation
   
 
Scholarly Communication
Contacts

Bivariate random-effects meta-analysis and the estimation of between-study correlation

Riley, Richard D.; Abrams, Keith R.; Sutton, Alexander J.; Lambert, Paul C. and Thompson, John R. (2007) Bivariate random-effects meta-analysis and the estimation of between-study correlation. BMC Medical Research Methodology, 7 . Article Number: 3. ISSN 1471-2288

[img]
Preview
PDF
Available under License Creative Commons Attribution.

395Kb

Cited 44 times in WoS

Abstract

Background: When multiple endpoints are of interest in evidence synthesis, a multivariate meta-analysis can jointly synthesise those endpoints and utilise their correlation. A multivariate random-effects metaanalysis must incorporate and estimate the between-study correlation (ρB). Methods: In this paper we assess maximum likelihood estimation of a general normal model and a generalised model for bivariate random-effects meta-analysis (BRMA). We consider two applied examples, one involving a diagnostic marker and the other a surrogate outcome. These motivate a simulation study where estimation properties from BRMA are compared with those from two separate univariate randomeffects meta-analyses (URMAs), the traditional approach. Results: The normal BRMA model estimates ρB as -1 in both applied examples. Analytically we show this is due to the maximum likelihood estimator sensibly truncating the between-study covariance matrix on the boundary of its parameter space. Our simulations reveal this commonly occurs when the number of studies is small or the within-study variation is relatively large; it also causes upwardly biased betweenstudy variance estimates, which are inflated to compensate for the restriction on B. Importantly, this does not induce any systematic bias in the pooled estimates and produces conservative standard errors and mean-square errors. Furthermore, the normal BRMA is preferable to two normal URMAs; the meansquare error and standard error of pooled estimates is generally smaller in the BRMA, especially given data missing at random. For meta-analysis of proportions we then show that a generalised BRMA model is better still. This correctly uses a binomial rather than normal distribution, and produces better estimates than the normal BRMA and also two generalised URMAs; however the model may sometimes not converge due to difficulties estimating ρB. Conclusion: A BRMA model offers numerous advantages over separate univariate synthesises; this paper highlights some of these benefits in both a normal and generalised modelling framework, and examines the estimation of between-study correlation to aid practitioners.

Item Type:Article
Additional Information:Published: 12 January 2007. 15 pages (page numbers not for citation purposes).
Uncontrolled Keywords:MULTIVARIATE APPROACH; MULTIPLE OUTCOMES; SENSITIVITY; REGRESSION; REVIEWS; MARKERS; BIAS
Subjects:R Medicine > R Medicine (General)
Departments, Research Centres and Related Units:Academic Faculties, Institutes and Research Centres > Research Centres > Medical Statistics and Health Evaluation, Centre for
DOI:10.1186/1471-2288-7-3
Publisher's Statement:© 2007 Riley et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Related URLs:
Refereed:Yes
Status:Published
ID Code:746
Deposited On:27 Jun 2008 15:17
Last Modified:29 Feb 2012 09:34

Repository Staff Only: item control page

   
Search


Full text only
Peer reviewed only

Browse
Cross Archive Search
Find
Top 50 authors
Top 50 items
[more statistics]
 
   

These pages are maintained by Library Staff @ University of Liverpool Library

 

All pages © The University of Liverpool, 2004 | Disclaimer | Accessibility | Staff | Students