A Re-Evaluation of Random-Effects Meta-Analysis

Julian P. T. Higgins1, Simon G. Thompson1, David J. Spiegelhalter1
1Medical Research Council Biostatistics Unit, Cambridge, UK

Tóm tắt

Summary

Meta-analysis in the presence of unexplained heterogeneity is frequently undertaken by using a random-effects model, in which the effects underlying different studies are assumed to be drawn from a normal distribution. Here we discuss the justification and interpretation of such models, by addressing in turn the aims of estimation, prediction and hypothesis testing. A particular issue that we consider is the distinction between inference on the mean of the random-effects distribution and inference on the whole distribution. We suggest that random-effects meta-analyses as currently conducted often fail to provide the key results, and we investigate the extent to which distribution-free, classical and Bayesian approaches can provide satisfactory methods. We conclude that the Bayesian approach has the advantage of naturally allowing for full uncertainty, especially for prediction. However, it is not without problems, including computational intensity and sensitivity to a priori judgements. We propose a simple prediction interval for classical meta-analysis and offer extensions to standard practice of Bayesian meta-analysis, making use of an example of studies of ‘set shifting’ ability in people with eating disorders.

Từ khóa


Tài liệu tham khảo

Abrams, 1998, Approximate Bayesian inference in random effects meta-analysis, Statist. Med., 17, 201, 10.1002/(SICI)1097-0258(19980130)17:2<201::AID-SIM736>3.0.CO;2-9

Ades, 2005, The interpretation of random-effects meta-analysis in decision models, Med. Decsn Mak., 25, 646, 10.1177/0272989X05282643

Aitkin, 1999, A general maximum likelihood analysis of variance components in generalized linear models, Biometrics, 55, 117, 10.1111/j.0006-341X.1999.00117.x

Aitkin, 1999, Meta-analysis by random effect modelling in generalized linear models, Statist. Med., 18, 2343, 10.1002/(SICI)1097-0258(19990915/30)18:17/18<2343::AID-SIM260>3.0.CO;2-3

Anello, 1995, Exploratory or analytic meta-analysis: should we distinguish between them?, J. Clin. Epidem., 48, 109, 10.1016/0895-4356(94)00084-4

Bailey, 1987, Inter-study differences—how should they influence the interpretation and analysis of results, Statist. Med., 6, 351, 10.1002/sim.4780060327

Berkey, 1995, A random-effects regression model for meta-analysis, Statist. Med., 14, 395, 10.1002/sim.4780140406

Bernardo, 1994, Bayesian Theory, 10.1002/9780470316870

Biggerstaff, 1997, Incorporating variability in estimates of heterogeneity in the random effects model in meta-analysis, Statist. Med., 16, 753, 10.1002/(SICI)1097-0258(19970415)16:7<753::AID-SIM494>3.0.CO;2-G

Böhning, 2000, Computer-assisted Analysis of Mixtures and Applications: Meta-analysis, Disease Mapping and Others, 10.1080/00401706.2000.10485740

Böhning, 2005, Meta-analysis: a unifying meta-likelihood approach framing unobserved heterogeneity, study covariates, publication bias, and study quality, Meth. Inform. Med., 44, 127, 10.1055/s-0038-1633931

Bradburn, 2006, Much ado about nothing: a comparison of the performance of meta-analytical methods with rare events, Statist. Med., 26, 53, 10.1002/sim.2528

Burr, 2005, A Bayesian semiparametric model for random-effects meta-analysis, J. Am. Statist. Ass., 100, 242, 10.1198/016214504000001024

Burr, 2003, A meta-analysis of studies on the association of the platelet PlA polymorphism of glycoprotein IIIa and risk of coronary heart disease, Statist. Med., 22, 1741, 10.1002/sim.1375

Cochrane Injuries Group Albumin Reviewers, 1998, Human albumin administration in critically ill patients: systematic review of randomised controlled trials, Br. Med. J., 317, 235, 10.1136/bmj.317.7153.235

Colditz, 1995, Heterogeneity in meta-analysis of data from epidemiologic studies: Commentary, Am. J. Epidem., 142, 371, 10.1093/oxfordjournals.aje.a117644

Cox, 2006, Encyclopedia of Statistical Sciences, 1074

DerSimonian, 2007, Random-effects model for meta-analysis of clinical trials: an update, Contemp. Clin. Trials, 28, 105, 10.1016/j.cct.2006.04.004

DerSimonian, 1986, Meta-analysis in clinical trials, Contr. Clin. Trials, 7, 177, 10.1016/0197-2456(86)90046-2

Detsky, 1992, Incorporating variations in the quality of individual randomized trials into meta-analysis, J. Clin. Epidem., 45, 255, 10.1016/0895-4356(92)90085-2

DuMouchel, 1994, Hierarchical Bayes linear models for meta-analysis

DuMouchel, 2000, Statistical Methodology in the Pharmaceutical Sciences, 127

Eddy, 1992, Meta-analysis by the Confidence Profile Method

Egger, 1997, Bias in meta-analysis detected by a simple, graphical test, Br. Med. J., 315, 629, 10.1136/bmj.315.7109.629

Feinstein, 1995, Meta-analysis: statistical alchemy for the 21st century, J. Clin. Epidem., 48, 71, 10.1016/0895-4356(94)00110-C

Follmann, 1999, Valid inference in random effects meta-analysis, Biometrics, 55, 732, 10.1111/j.0006-341X.1999.00732.x

Gail, 1985, Testing for qualitative interaction between treatment effects and patient subsets, Biometrics, 41, 361, 10.2307/2530862

Gelman, 2006, Prior distributions for variance parameters in hierarchical models, Bayes. Anal., 1, 515

Gelman, 1995, Bayesian Data Analysis, 10.1201/9780429258411

Glasziou, 2002, Investigating causes of heterogeneity in systematic reviews, Statist. Med., 21, 1503, 10.1002/sim.1183

Goodman, 1999, Toward evidence-based medical statistics, 2: the Bayes factor, Ann. Intern. Med., 130, 1005, 10.7326/0003-4819-130-12-199906150-00019

Greenland, 2005, Multiple-bias modelling for analysis of observational data, J. R. Statist. Soc. A, 168, 267, 10.1111/j.1467-985X.2004.00349.x

Greenland, 2001, On the bias produced by quality scores in meta-analysis, and a hierarchical view of proposed solutions, Biostatistics, 2, 463, 10.1093/biostatistics/2.4.463

Hardy, 1996, A likelihood approach to meta-analysis with random effects, Statist. Med., 15, 619, 10.1002/(SICI)1097-0258(19960330)15:6<619::AID-SIM188>3.0.CO;2-A

Hardy, 1998, Detecting and describing heterogeneity in meta-analysis, Statist. Med., 17, 841, 10.1002/(SICI)1097-0258(19980430)17:8<841::AID-SIM781>3.0.CO;2-D

Hartung, 2001, A refined method for the meta-analysis of controlled clinical trials with binary outcome, Statist. Med., 20, 3875, 10.1002/sim.1009

Hartung, 2001, On tests of the overall treatment effect in meta-analysis with normally distributed responses, Statist. Med., 20, 1771, 10.1002/sim.791

Hedges, 1987, Commentary, Statist. Med., 6, 381, 10.1002/sim.4780060333

Hedges, 1985, Statistical Methods for Meta-analysis

Higgins, 2002, Quantifying heterogeneity in a meta-analysis, Statist. Med., 21, 1539, 10.1002/sim.1186

Higgins, 2002, Statistical heterogeneity in systematic reviews of clinical trials: a critical appraisal of guidelines and practice, J. Hlth Serv. Res. Poly, 7, 51, 10.1258/1355819021927674

Higgins, 1996, Borrowing strength from external trials in a meta-analysis, Statist. Med., 15, 2733, 10.1002/(SICI)1097-0258(19961230)15:24<2733::AID-SIM562>3.0.CO;2-0

Jüni, 2001, Assessing the quality of controlled clinical trials, Br. Med. J., 323, 42, 10.1136/bmj.323.7303.42

Kass, 1995, Bayes Factors, J. Am. Statist. Ass., 90, 773, 10.1080/01621459.1995.10476572

Knapp, 2006, Assessing the amount of heterogeneity in random-effects meta-analysis, Biometr. J., 48, 271, 10.1002/bimj.200510175

Laird, 1978, Nonparametric maximum likelihood estimation of a mixing distribution, J. Am. Statist. Ass., 73, 805, 10.1080/01621459.1978.10480103

Laird, 1989, Empirical Bayes confidence intervals for a series of related experiments, Biometrics, 45, 481, 10.2307/2531491

Lambert, 2005, How vague is vague?: a simulation study of the impact of the use of vague prior distributions in MCMC using WinBUGS, Statist. Med., 24, 2401, 10.1002/sim.2112

Larholt, 1989, Statistical methods and heterogeneity in meta-analysis

Lee, 2007, Flexible parametric models for random-effects distributions, Statist. Med., 27, 418, 10.1002/sim.2897

Light, 1987, Accumulating evidence from independent studies—what we can win and what we can lose, Statist. Med., 6, 221, 10.1002/sim.4780060304

Louis, 1991, Using Empirical Bayes methods in biopharmaceutical research, Statist. Med., 10, 811, 10.1002/sim.4780100604

Morris, 1983, Parametric empirical Bayes inference: theory and applications, J. Am. Statist. Ass., 78, 47, 10.1080/01621459.1983.10477920

Morris, 1992, Hierachical models for combining information and for meta-analysis, Bayes. Statist., 4, 321

O’Hagan, 1994, Kendall’s Advanced Theory of Statistics, vol. 2B, Bayesian Inference

Ohlssen, 2007, Flexible random-effects models using Bayesian semi-parametric models: applications to institutional comparisons, Statist. Med., 26, 2088, 10.1002/sim.2666

Pan, 1997, Test for qualitative interaction of clinical significance, Statist. Med., 16, 1645, 10.1002/(SICI)1097-0258(19970730)16:14<1645::AID-SIM596>3.0.CO;2-G

Pauler, 2000, Meta-analysis in Medicine and Health Policy, 205

Peto, 1987, Why do we need systematic overviews of randomised trials?, Statist. Med., 6, 233, 10.1002/sim.4780060306

Piantadosi, 1993, A comparison of the power of two tests for qualitative interactions, Statist. Med., 12, 1239, 10.1002/sim.4780121305

Raghunathan, 1993, Analysis of binary data from a multicenter clinical-trial, Biometrika, 80, 127, 10.1093/biomet/80.1.127

Raudenbush, 1985, Empirical Bayes meta-analysis, J. Educ. Statist., 10, 75, 10.3102/10769986010002075

Robbins, 1983, Some thoughts on empirical Bayes estimation, Ann. Statist., 11, 713, 10.1214/aos/1176346239

Roberts, 2007, A systematic review and meta-analysis of set-shifting ability in eating disorders, Psychol. Med., 37, 1075, 10.1017/S0033291707009877

Skene, 1990, Hierarchical models for multicentre binary response studies, Statist. Med., 9, 919, 10.1002/sim.4780090808

Smith, 1995, Bayesian approaches to random-effects meta-analysis: a comparative study, Statist. Med., 14, 2685, 10.1002/sim.4780142408

Spiegelhalter, 2003, Bayesian approaches to multiple sources of evidence and uncertainty in complex cost-effectiveness modelling, Statist. Med., 22, 3687, 10.1002/sim.1586

Spiegelhalter, 2002, Bayesian measures of model complexity and fit (with discussion), J. R. Statist. Soc. B, 64, 583, 10.1111/1467-9868.00353

Spiegelhalter, 2003, WinBUGS User Manual, Version 1.4

Stein, 1956, Proc. 3rd Berkeley Symp, 197

Stijnen, 1990, Empirical Bayes methods in clinical trials meta-analysis, Biometr. J., 32, 335, 10.1002/bimj.4710320316

Thompson, 1994, Why sources of heterogeneity in meta-analysis should be investigated, Br. Med. J., 309, 1351, 10.1136/bmj.309.6965.1351

Thompson, 1998, Encyclopedia of Biostatistics, 2570

Thompson, 2002, How should meta-regression analyses be undertaken and interpreted?, Statist. Med., 21, 1559, 10.1002/sim.1187

Tjur, 1998, Nonlinear regression, quasi likelihood, and overdispersion in generalized linear models, Am. Statistn, 52, 222

Turner, 2008, Bias modelling in evidence synthesis, J. R. Statist. Soc. A

Vangel, 1999, Maximum likelihood analysis for heteroscedastic one-way random effects ANOVA in interlaboratory studies, Biometrics, 55, 129, 10.1111/j.0006-341X.1999.00129.x

Van Houwelingen, 1993, A bivariate approach to meta-analysis, Statist. Med., 12, 2273, 10.1002/sim.4780122405

Viechtbauer, 2007, Confidence intervals for the amount of heterogeneity in meta-analysis, Statist. Med., 26, 37, 10.1002/sim.2514

Warn, 2002, Bayesian random effects meta-analysis of trials with binary outcomes: methods for the absolute risk difference and relative risk scales, Statist. Med., 21, 1601, 10.1002/sim.1189

Welton, 2008, Models for potentially biased evidence in meta-analysis using empirically based priors, J. R. Statist. Soc. A

Whitehead, 2002, Meta-analysis of Controlled Clinical Trials, 10.1002/0470854200

Whitehead, 1991, A general parametric approach to the meta-analysis of randomised clinical trials, Statist. Med., 10, 1665, 10.1002/sim.4780101105