Robust inference methods for meta‐analysis involving influential outlying studies

Hisashi Noma; Shonosuke Sugasawa

doi:10.1002/sim.10157

journal article Open Access Jun 20, 2024

Robust inference methods for meta‐analysis involving influential outlying studies

Hisashi Noma

Shonosuke Sugasawa

Statistics in Medicine Vol. 43 No. 20 pp. 3778-3791 · Wiley

View at Publisher Save 10.1002/sim.10157

Abstract

Meta‐analysis is an essential tool to comprehensively synthesize and quantitatively evaluate results of multiple clinical studies in evidence‐based medicine. In many meta‐analyses, the characteristics of some studies might markedly differ from those of the others, and these outlying studies can generate biases and potentially yield misleading results. In this article, we provide effective robust statistical inference methods using generalized likelihoods based on the density power divergence. The robust inference methods are designed to adjust the influences of outliers through the use of modified estimating equations based on a robust criterion, even when multiple and serious influential outliers are present. We provide the robust estimators, statistical tests, and confidence intervals via the generalized likelihoods for the fixed‐effect and random‐effects models of meta‐analysis. We also assess the contribution rates of individual studies to the robust overall estimators that indicate how the influences of outlying studies are adjusted. Through simulations and applications to two recently published systematic reviews, we demonstrate that the overall conclusions and interpretations of meta‐analyses can be markedly changed if the robust inference methods are applied and that only the conventional inference methods might produce misleading evidence. These methods would be recommended to be used at least as a sensitivity analysis method in the practice of meta‐analysis. We have also developed an R package, robustmeta, that implements the robust inference methods.

Topics

No keywords indexed for this article. Browse by subject →

References

36

[1]

10.1002/9781119536604

[2]

The PRISMA 2020 statement: an updated guideline for reporting systematic reviews

Matthew J Page, Joanne E McKenzie, Patrick M Bossuyt et al.

BMJ 10.1136/bmj.n71

[3]

10.1002/0471725153

[4]

Outlier and influence diagnostics for meta-analysis

Wolfgang Viechtbauer, Mike W.-L. Cheung

Research Synthesis Methods 10.1002/jrsm.11

[5]

10.1007/s10729-007-9041-8

[6]

10.1002/jrsm.1191

[7]

Flexible parametric models for random‐effects distributions

Katherine J. Lee, Simon G. Thompson

Statistics in Medicine 10.1002/sim.2897

[8]

10.2188/jea.je20200376

[9]

10.1002/0471725382

[10]

10.1007/978-4-431-56922-0

[11]

10.1093/biomet/85.3.549

[12]

On Information and Sufficiency

S. Kullback, R. A. Leibler

The Annals of Mathematical Statistics 10.1214/aoms/1177729694

[13]

Ghosh A "Robust estimation for independent non‐homogeneous observations using density power divergence with applications to linear regression" Electron J Statist (2016)

[14]

10.1007/s11749-015-0445-3

[15]

10.1002/sim.8377

[16]

An Optimum Property of Regular Maximum Likelihood Estimation

V. P. Godambe

The Annals of Mathematical Statistics 10.1214/aoms/1177705693

[17]

10.1136/bmj.l689

[18]

10.1136/bmj.h1109

[19]

Meta-analysis in clinical trials

Rebecca DerSimonian, Nan Laird

Controlled Clinical Trials 10.1016/0197-2456(86)90046-2

[20]

10.3390/e23091147

[21]

10.1080/01621459.2018.1518237

[22]

Hyvärinen A "Estimation of non‐normalized statistical models by score matching" J Mach Learn Res (2005)

[23]

Burden RL (2010)

[24]

Choosing a robustness tuning parameter

J. Warwick, M. C. Jones

Journal of Statistical Computation and Simulation 10.1080/00949650412331299120

[25]

On the ‘optimal’ density power divergence tuning parameter

Sancharee Basak, Ayanendranath Basu, M. C. Jones

Journal of Applied Statistics 10.1080/02664763.2020.1736524

[26]

10.1093/biomet/asz075

[27]

10.1002/jrsm.1455

[28]

10.32614/rj-2016-001

[29]

Methods to calculate uncertainty in the estimated overall effect size from a random‐effects meta‐analysis

Areti Angeliki Veroniki, Dan Jackson, Ralf Bender et al.

Research Synthesis Methods 10.1002/jrsm.1319

[30]

10.1002/sim.7140

[31]

10.1002/sim.4350

[32]

10.1201/9780429246593

[33]

10.1002/sim.4172

[34]

10.1002/jrsm.1037

[35]

10.1016/j.jmva.2021.104846

[36]

10.1136/bmj.c221

Metrics

3

Citations

36

References

Details

Published: Jun 20, 2024
Vol/Issue: 43(20)
Pages: 3778-3791
License: View

Authors

H

Hisashi Noma

Department of Interdisciplinary Statistical Mathematics The Institute of Statistical Mathematics Tokyo Japan; The Graduate Institute for Advanced Studies The Graduate University for Advanced Studies (SOKENDAI) Tokyo Japan

S

Shonosuke Sugasawa

Faculty of Economics Keio University Tokyo Japan

Funding

Japan Society for the Promotion of Science Award: JP21K03105

Cite This Article

Hisashi Noma, Shonosuke Sugasawa (2024). Robust inference methods for meta‐analysis involving influential outlying studies. Statistics in Medicine, 43(20), 3778-3791. https://doi.org/10.1002/sim.10157

Robust inference methods for meta‐analysis involving influential outlying studies

You May Also Like