A re-evaluation of fixed effect(s) meta-analysis

Published on Jan 1, 2018in Journal of The Royal Statistical Society Series A-statistics in Society2.483
· DOI :10.1111/RSSA.12275
Kenneth Rice89
Estimated H-index: 89
(UW: University of Washington),
Julian P T Higgins134
Estimated H-index: 134
(UoB: University of Bristol),
Thomas Lumley104
Estimated H-index: 104
(University of Auckland)
Sources
Abstract
Meta†analysis is a common tool for synthesizing results of multiple studies. Among methods for performing meta†analysis, the approach known as ‘fixed effects’ or ‘inverse variance weighting’ is popular and widely used. A common interpretation of this method is that it assumes that the underlying effects in contributing studies are identical, and for this reason it is sometimes dismissed by practitioners. However, other interpretations of fixed effects analyses do not make this assumption, yet appear to be little known in the literature. We review these alternative interpretations, describing both their strengths and their limitations. We also describe how heterogeneity of the underlying effects can be addressed, with the same minimal assumptions, through either testing or meta†regression. Recommendations for the practice of meta†analysis are given; it is hoped that these will foster more direct connection of the questions that meta†analysts wish to answer with the statistical methods they choose.
Figures & Tables
Download
📖 Papers frequently viewed together
References81
Newest
#1David C. Hoaglin (UMMS: University of Massachusetts Medical School)H-Index: 37
Many meta-analyses report using ‘Cochran's Q test' to assess heterogeneity of effect-size estimates from the individual studies. Some authors cite work by W. G. Cochran, without realizing that Cochran deliberately did not use Q itself to test for heterogeneity. Further, when heterogeneity is absent, the actual null distribution of Q is not the chi-squared distribution assumed for ‘Cochran's Q test'. This paper reviews work by Cochran related to Q. It then discusses derivations of the asymptotic ...
Source
#1David C. Hoaglin (UMMS: University of Massachusetts Medical School)H-Index: 37
Keywords: inverse-variance weights; heterogeneity; random-effects model; Q statistic; softwareInspired by the papers of Gene Glass, Robert Rosenthal, Frank Schmidt, and Will Shadish and Jesse Lecy, I examinethe connection between research on methods and the practice of meta-analysis. Surprisingly, most results in theliterature on methods have only limited relevance to actual practice. The reasons for this situation seem to formfive categories: incorrect assumptions, inaccurate approximations, pre...
Source
#1Andreas BujaH-Index: 40
#2Richard A. BerkH-Index: 65
Last. Kai ZhangH-Index: 12
view all 8 authors...
Abstract. More than thirty years ago Halbert White inaugurated a “modelrobust” form of statistical inference based on the “sandwich estimator” of standard error. It is asymptotically correct even under “model misspecification,” that is, when models are approximations rather than generative truths. It is well-known to be “heteroskedasticity-consistent”, but it is less well-known to be “nonlinearity-consistent” as well. Nonlinearity, however, raises fundamental issues: When fitted models are appro...
#1David SpiegelhalterH-Index: 115
#2Keith R. AbramsH-Index: 92
Last. Jonathan P. MylesH-Index: 1
view all 3 authors...
Preface. List of examples. 1. Introduction. 1.1 What are Bayesian methods? 1.2 What do we mean by 'health--care evaluation'? 1.3 A Bayesian approach to evaluation. 1.4 The aim of this book and the intended audience. 1.5 Structure of the book. 2. Basic Concepts from Traditional Statistical Analysis. 2.1 Probability. 2.1.1 What is probability? 2.1.2 Odds and log--odds. 2.1.3 Bayes theorem for simple events. 2.2 Random variables, parameters and likelihood. 2.2.1 Random variables and their distribut...
#1Martin Gögele (European Academy of Bozen)H-Index: 23
#2Cosetta MinelliH-Index: 52
Last. John R. ThompsonH-Index: 88
view all 9 authors...
There has been a steep increase in the number of meta-analyses of genome-wide association (GWA) studies aimed at identifying genetic variants with increasingly smaller effects, but pressure to publish findings of new genetic associations has limited the time available for careful consideration of all of their methodological aspects. The authors surveyed the literature (2007–2010) to provide empirical evidence on the methods used in GWAmetaanalyses, including their organization, requirements abou...
Source
This paper investigates the use of likelihood methods for meta-analysis, within the random-effects models framework. We show that likelihood inference relying on first-order approximations, while improving common meta-analysis techniques, can be prone to misleading results. This drawback is very evident in the case of small sample sizes, which are typical in meta-analysis. We alleviate the problem by exploiting the theory of higher-order asymptotics. In particular, we focus on a second-order adj...
Source
#1Elena Kulinskaya (UEA: University of East Anglia)H-Index: 33
#2Michael B. Dollinger (PLU: Pacific Lutheran University)H-Index: 8
Last. Kirsten Bjørkestøl (University of Agder)H-Index: 4
view all 3 authors...
W. G. Cochran's Q statistic was introduced in 1937 to test for equality of means under heteroscedasticity. Today, the use of Q is widespread in tests for homogeneity of effects in meta-analysis, but often these effects (such as risk differences and odds ratios) are not normally distributed. It is common to assume that Q follows a chi-square distribution, but it has long been known that this asymptotic distribution for Q is not accurate for moderate sample sizes. In this paper, the effect and wei...
Source
#1Elena Kulinskaya (UEA: University of East Anglia)H-Index: 33
#2Michael B. Dollinger (PLU: Pacific Lutheran University)H-Index: 8
Last. Kirsten Bjørkestøl (University of Agder)H-Index: 4
view all 3 authors...
Meta-analysis seeks to combine the results of several experiments in order to improve the accuracy of decisions. It is common to use a test for homogeneity to determine if the results of the several experiments are sufficiently similar to warrant their combination into an overall result. Cochran'sQstatistic is frequently used for this homogeneity test. It is often assumed thatQfollows a chi-square distribution under the null hypothesis of homogeneity, but it has long been known that this asympto...
Source
#1Gerta Rücker (University Medical Center Freiburg)H-Index: 51
#2Guido Schwarzer (University Medical Center Freiburg)H-Index: 59
Last. Martin Schumacher (University Medical Center Freiburg)H-Index: 102
view all 4 authors...
Source
#1Danyu Lin (UNC: University of North Carolina at Chapel Hill)H-Index: 89
#2Donglin Zeng (UNC: University of North Carolina at Chapel Hill)H-Index: 50
Meta-analysis is widely used to synthesize the results of multiple studies. Although meta-analysis is traditionally carried out by combining the summary statistics of relevant studies, advances in technologies and communications have made it increasingly feasible to access the original data on individual participants. In the present paper, we investigate the relative efficiency of analyzing original data versus combining summary statistics. We show that, for all commonly used parametric and semi...
Source
Cited By86
Newest
Source
#1R. J. Fredericksen (Harborview Medical Center)H-Index: 2
#4R. M. Nance (Harborview Medical Center)H-Index: 1
Last. C. Del Rio (Emory University)H-Index: 12
view all 24 authors...
Source
Meta-analyses combine the estimators of individual means to estimate the common mean of a population. However, the common mean could be undefined or uninformative in some scenarios where individual means are “ordered” or “sparse”. Hence, assessments of individual means become relevant, rather than the common mean. In this article, we propose simultaneous estimation of individual means using the James–Stein shrinkage estimators, which improve upon individual studies’ estimators. We also propose i...
Source
#1Jingwen Zhang (University of Melbourne)H-Index: 2
#2Jennifer L. Perret (University of Melbourne)H-Index: 22
Last. Shyamali C. Dharmage (University of Melbourne)H-Index: 76
view all 10 authors...
Despite the challenges of diagnosing and managing adult patients with chronic cough, a systematic synthesis of evidence on aetiological risk factor is lacking. We systematically searched PubMed and EMBASE to synthesize the current evidence for longitudinal associations between a wide range of risk factors and chronic cough in the general adult population, following the meta-analysis of observational studies in epidemiology (MOOSE) guidelines. The Newcastle-Ottawa scale was used to assess the qua...
Source
#1Cullen Soares (UMB: University of Maryland, Baltimore)H-Index: 1
#2Amjad Samara (WashU: Washington University in St. Louis)H-Index: 5
Last. Sebhat Erqou (Providence Regional Medical Center Everett)H-Index: 22
view all 10 authors...
Background Studies have reported that people living with HIV have higher burden of subclinical cardiovascular disease, but the data are not adequately synthesized. We performed meta‐analyses of stu...
Source
#1Spencer Hansen (UW: University of Washington)H-Index: 1
#2Kenneth Rice (UW: University of Washington)H-Index: 89
Meta-analysis of proportions is conceptually simple: Faced with a binary outcome in multiple studies, we seek inference on some overall proportion of successes/failures. Under common effect models, exact inference has long been available, but is not when we more realistically allow for heterogeneity of the proportions. Instead a wide range of nonexact fixed-effects methods are used, the interpretation of some of which is challenging. In this paper, we present methods for exact statistical tests ...
Source
#1Christina Vogel (University Hospital Southampton NHS Foundation Trust)H-Index: 10
#2Sarah R Crozier (Southampton General Hospital)H-Index: 49
Last. Janis Baird (University Hospital Southampton NHS Foundation Trust)H-Index: 30
view all 8 authors...
Background Previous product placement trials in supermarkets are limited in scope and outcome data collected. This study assessed the effects on store-level sales, household-level purchasing, and dietary behaviours of a healthier supermarket layout. null null Methods and findings This is a prospective matched controlled cluster trial with 2 intervention components: (i) new fresh fruit and vegetable sections near store entrances (replacing smaller displays at the back) and frozen vegetables repos...
Source
#1William L. Baker (UConn: University of Connecticut)H-Index: 35
#2Diana M Sobieraj (UConn: University of Connecticut)H-Index: 23
Source
#1Doretta Caramaschi (UoB: University of Bristol)H-Index: 13
#2James Jungius (UoB: University of Bristol)H-Index: 1
Last. Siri E. HåbergH-Index: 35
view all 13 authors...
STUDY QUESTION Is cord blood DNA methylation associated with having been conceived by medically assisted reproduction? SUMMARY ANSWER This study does not provide strong evidence of an association of conception by medically assisted reproduction with variation in infant blood cell DNA methylation. WHAT IS KNOWN ALREADY Medically assisted reproduction consists of procedures used to help infertile/subfertile couples conceive, including ART. Due to its importance in gene regulation during early deve...
Source
#1Sacha Epskamp (UvA: University of Amsterdam)H-Index: 33
#2Adela-Maria Isvoranu (UvA: University of Amsterdam)H-Index: 7
Last. Mike W.-L. Cheung (NUS: National University of Singapore)H-Index: 38
view all 3 authors...
A growing number of publications focus on estimating Gaussian graphical models (GGM, networks of partial correlation coefficients). At the same time, generalizibility and replicability of these highly parameterized models are debated, and sample sizes typically found in datasets may not be sufficient for estimating the underlying network structure. In addition, while recent work emerged that aims to compare networks based on different samples, these studies do not take potential cross-study hete...
Source
This website uses cookies.
We use cookies to improve your online experience. By continuing to use our website we assume you agree to the placement of these cookies.
To learn more, you can find in our Privacy Policy.