Individual participant data meta-analyses should not ignore clustering

Objectives Individual participant data (IPD) meta-analyses often analyze their IPD as if coming from a single study. We compare this approach with analyses that rather account for clustering of patients within studies. Study Design and Setting Comparison of effect estimates from logistic re...

Full description

Bibliographic Details
Main Authors: Abo-Zaid, Ghada, Guo, Boliang, Deeks, Jonathan J., Debray, Thomas P.A., Steyerberg, Ewout W., Moons, Karel G.M., Riley, Richard David
Format: Article
Published: Elsevier 2013
Online Access:https://eprints.nottingham.ac.uk/31498/
_version_ 1848794214953910272
author Abo-Zaid, Ghada
Guo, Boliang
Deeks, Jonathan J.
Debray, Thomas P.A.
Steyerberg, Ewout W.
Moons, Karel G.M.
Riley, Richard David
author_facet Abo-Zaid, Ghada
Guo, Boliang
Deeks, Jonathan J.
Debray, Thomas P.A.
Steyerberg, Ewout W.
Moons, Karel G.M.
Riley, Richard David
author_sort Abo-Zaid, Ghada
building Nottingham Research Data Repository
collection Online Access
description Objectives Individual participant data (IPD) meta-analyses often analyze their IPD as if coming from a single study. We compare this approach with analyses that rather account for clustering of patients within studies. Study Design and Setting Comparison of effect estimates from logistic regression models in real and simulated examples. Results The estimated prognostic effect of age in patients with traumatic brain injury is similar, regardless of whether clustering is accounted for. However, a family history of thrombophilia is found to be a diagnostic marker of deep vein thrombosis [odds ratio, 1.30; 95% confidence interval (CI): 1.00, 1.70; P = 0.05] when clustering is accounted for but not when it is ignored (odds ratio, 1.06; 95% CI: 0.83, 1.37; P = 0.64). Similarly, the treatment effect of nicotine gum on smoking cessation is severely attenuated when clustering is ignored (odds ratio, 1.40; 95% CI: 1.02, 1.92) rather than accounted for (odds ratio, 1.80; 95% CI: 1.29, 2.52). Simulations show models accounting for clustering perform consistently well, but downwardly biased effect estimates and low coverage can occur when ignoring clustering. Conclusion Researchers must routinely account for clustering in IPD meta-analyses; otherwise, misleading effect estimates and conclusions may arise.
first_indexed 2025-11-14T19:12:39Z
format Article
id nottingham-31498
institution University of Nottingham Malaysia Campus
institution_category Local University
last_indexed 2025-11-14T19:12:39Z
publishDate 2013
publisher Elsevier
recordtype eprints
repository_type Digital Repository
spelling nottingham-314982020-05-04T20:18:56Z https://eprints.nottingham.ac.uk/31498/ Individual participant data meta-analyses should not ignore clustering Abo-Zaid, Ghada Guo, Boliang Deeks, Jonathan J. Debray, Thomas P.A. Steyerberg, Ewout W. Moons, Karel G.M. Riley, Richard David Objectives Individual participant data (IPD) meta-analyses often analyze their IPD as if coming from a single study. We compare this approach with analyses that rather account for clustering of patients within studies. Study Design and Setting Comparison of effect estimates from logistic regression models in real and simulated examples. Results The estimated prognostic effect of age in patients with traumatic brain injury is similar, regardless of whether clustering is accounted for. However, a family history of thrombophilia is found to be a diagnostic marker of deep vein thrombosis [odds ratio, 1.30; 95% confidence interval (CI): 1.00, 1.70; P = 0.05] when clustering is accounted for but not when it is ignored (odds ratio, 1.06; 95% CI: 0.83, 1.37; P = 0.64). Similarly, the treatment effect of nicotine gum on smoking cessation is severely attenuated when clustering is ignored (odds ratio, 1.40; 95% CI: 1.02, 1.92) rather than accounted for (odds ratio, 1.80; 95% CI: 1.29, 2.52). Simulations show models accounting for clustering perform consistently well, but downwardly biased effect estimates and low coverage can occur when ignoring clustering. Conclusion Researchers must routinely account for clustering in IPD meta-analyses; otherwise, misleading effect estimates and conclusions may arise. Elsevier 2013-08 Article NonPeerReviewed Abo-Zaid, Ghada, Guo, Boliang, Deeks, Jonathan J., Debray, Thomas P.A., Steyerberg, Ewout W., Moons, Karel G.M. and Riley, Richard David (2013) Individual participant data meta-analyses should not ignore clustering. Journal of Clinical Epidemiology, 66 (8). 865-873.e4. ISSN 1878-5921 http://www.sciencedirect.com/science/article/pii/S0895435613000723 doi:10.1016/j.jclinepi.2012.12.017 doi:10.1016/j.jclinepi.2012.12.017
spellingShingle Abo-Zaid, Ghada
Guo, Boliang
Deeks, Jonathan J.
Debray, Thomas P.A.
Steyerberg, Ewout W.
Moons, Karel G.M.
Riley, Richard David
Individual participant data meta-analyses should not ignore clustering
title Individual participant data meta-analyses should not ignore clustering
title_full Individual participant data meta-analyses should not ignore clustering
title_fullStr Individual participant data meta-analyses should not ignore clustering
title_full_unstemmed Individual participant data meta-analyses should not ignore clustering
title_short Individual participant data meta-analyses should not ignore clustering
title_sort individual participant data meta-analyses should not ignore clustering
url https://eprints.nottingham.ac.uk/31498/
https://eprints.nottingham.ac.uk/31498/
https://eprints.nottingham.ac.uk/31498/