Research and Reporting Methods18 February 2014
    Author, Article, and Disclosure Information

    This article has an additional interactive example appended as a Supplement. Please visit the Supplement tab on this page to access the presentation.

    A primary goal of meta-analysis is to improve the estimation of treatment effects by pooling results of similar studies. This article explains how the most widely used method for pooling heterogeneous studies—the DerSimonian–Laird (DL) estimator—can produce biased estimates with falsely high precision. A classic example is presented to show that use of the DL estimator can lead to erroneous conclusions. Particular problems with the DL estimator are discussed, and several alternative methods for summarizing heterogeneous evidence are presented. The authors support replacing universal use of the DL estimator with analyses based on a critical synthesis that recognizes the uncertainty in the evidence, focuses on describing and explaining the probable sources of variation in the evidence, and uses random-effects estimates that provide more accurate confidence limits than the DL estimator.

    References

    • 1. DerSimonian RLaird NMeta-analysis in clinical trials. Control Clin Trials1986;7:177-88. [PMID: 3802833] CrossrefMedlineGoogle Scholar
    • 2. Weseley ACDouglas GWContinuous use of chlorothiazide for prevention of toxemia of pregnancy. Obstet Gynecol1962;19:355-8. [PMID: 14006267] MedlineGoogle Scholar
    • 3. Flowers CEGrizzle JEEasterling WEBonner OBChlorothiazide as a prophylaxis against toxemia of pregnancy. A double-blind study. Am J Obstet Gynecol1962;84:919-29. [PMID: 13893680] CrossrefMedlineGoogle Scholar
    • 4. Menzies DNControlled trial of chlorothiazide in treatment of early pre-eclampsia. Br Med J1964;1:739-42. [PMID: 14102017] CrossrefMedlineGoogle Scholar
    • 5. Fallis NEPlauche WCMosey LMLangford HGThiazide versus placebo in prophylaxis of toxemia of pregnancy in primigravid patients. Am J Obstet Gynecol1964;88:502-4. [PMID: 14123429] CrossrefMedlineGoogle Scholar
    • 6. Cuadros ATatum HJThe prophylactic and therapeutic use of bendroflumethiazide in pregnancy. Am J Obstet Gynecol1964;89:891-7. [PMID: 14207556] CrossrefMedlineGoogle Scholar
    • 7. Landesman RAguero OWilson KLaRussa RCampbell WPenaloza OThe prophylactic use of chlorthalidone, a sulfonamide diuretic, in pregnancy. Br J Obstet Gynaecol1965;72:1004-10. CrossrefGoogle Scholar
    • 8. Kraus GWMarchese JRYen SSProphylactic use of hydrochlorothiazide in pregnancy. JAMA1966;198:1150-4. [PMID: 5332983] CrossrefMedlineGoogle Scholar
    • 9. Tervilä LVartiainen EThe effects and side effects of diuretics in the prophylaxis of toxaemia of pregnancy. Acta Obstet Gynecol Scand1971;50:351-6. [PMID: 4945572] CrossrefMedlineGoogle Scholar
    • 10. Campbell DMMacGillivray IThe effect of a low calorie diet or a thiazide diuretic on the incidence of pre-eclampsia and on birth weight. Br J Obstet Gynaecol1975;82:572-7. [PMID: 1096930] CrossrefMedlineGoogle Scholar
    • 11. Collins RYusuf SPeto ROverview of randomised trials of diuretics in pregnancy. Br Med J (Clin Res Ed)1985;290:17-23. [PMID: 3917318] CrossrefMedlineGoogle Scholar
    • 12. Biggerstaff BJTweedie RLIncorporating variability in estimates of heterogeneity in the random effects model in meta-analysis. Stat Med1997;16:753-68. [PMID: 9131763] CrossrefMedlineGoogle Scholar
    • 13. Hardy RJThompson SGA likelihood approach to meta-analysis with random effects. Stat Med1996;15:619-29. [PMID: 8731004] CrossrefMedlineGoogle Scholar
    • 14. Higgins JPThompson SGQuantifying heterogeneity in a meta-analysis. Stat Med2002;21:1539-58. [PMID: 12111919] CrossrefMedlineGoogle Scholar
    • 15. Review Manger (RevMan) [computer program]. Version 5.2. Copenhagen, Denmark: Nordic Cochrane Center, Cochrane Collaboration; 2012. Google Scholar
    • 16. DerSimonian RKacker RRandom-effects model for meta-analysis of clinical trials: an update. Contemp Clin Trials2007;28:105-14. [PMID: 16807131] CrossrefMedlineGoogle Scholar
    • 17. Knapp GHartung JImproved tests for a random effects meta-regression with a single covariate. Stat Med2003;22:2693-710. [PMID: 12939780] CrossrefMedlineGoogle Scholar
    • 18. Normand SLMeta-analysis: formulating, evaluating, combining, and reporting. Stat Med1999;18:321-59. [PMID: 10070677] CrossrefMedlineGoogle Scholar
    • 19. Whitehead AMeta-analysis of Controlled Clinical Trials. New York: J Wiley; 2002. Google Scholar
    • 20. Higgins JPThompson SGSpiegelhalter DJA re-evaluation of random-effects meta-analysis. J R Stat Soc Ser A Stat Soc2009;172:137-59. [PMID: 19381330] CrossrefMedlineGoogle Scholar
    • 21. Sterne JACMeta-analysis in Stata: An Updated Collection from the Stata Journal. 1st ed. College Station, Texas: Stata Pr; 2009. Google Scholar
    • 22. Viechtbauer WConducting meta-analysis in R with the metafor package. J Stat Softw2010;36:1-48. CrossrefGoogle Scholar
    • 23. Brockwell SEGordon IRA comparison of statistical methods for meta-analysis. Stat Med2001;20:825-40. [PMID: 11252006] CrossrefMedlineGoogle Scholar
    • 24. Kontopantelis EReeves Dmetaan: Random-effects meta-analysis. Stata J2010;10:395-407. CrossrefGoogle Scholar
    • 25. Guolo AVarin CThe R package metaLik for likelihood inference in meta-analysis. J Stat Softw2012;50:1-14. CrossrefMedlineGoogle Scholar
    • 26. Guolo AHigher-order likelihood inference in meta-analysis and meta-regression. Stat Med2012;31:313-27. [PMID: 22173666] CrossrefMedlineGoogle Scholar
    • 27. Warn DEThompson SGSpiegelhalter DJBayesian random effects meta-analysis of trials with binary outcomes: methods for the absolute risk difference and relative risk scales. Stat Med2002;21:1601-23. [PMID: 12111922] CrossrefMedlineGoogle Scholar
    • 28. Lambert PCSutton AJBurton PRAbrams KRJones DRHow vague is vague? A simulation study of the impact of the use of vague prior distributions in MCMC using WinBUGS. Stat Med2005;24:2401-28. [PMID: 16015676] CrossrefMedlineGoogle Scholar
    • 29. Higgins JPWhitehead ABorrowing strength from external trials in a meta-analysis. Stat Med1996;15:2733-49. [PMID: 8981683] CrossrefMedlineGoogle Scholar