Review papers : The statistical basis of meta-analysis

Two models for study-to-study variation in a meta-analysis are presented, critiqued and illustrated. One, the fixed effects model, takes the studies being analysed as the universe of interest; the other, the random effects model, takes these studies as representing a sample from a larger population of possible studies. With emphasis on clinical trials, this paper illustrates in some detail the application of both models to three summary measures of the effect of an experimental intervention versus a control: the standardized difference for comparing two means, and the relative risk and odds ratio for comparing two proportions.

[1]  B. Woolf.  ON ESTIMATING THE RELATION BETWEEN BLOOD GROUP AND DISEASE , 1955, Annals of human genetics.

[2]  W. Haenszel,et al.  Statistical aspects of the analysis of data from retrospective studies of disease. , 1959, Journal of the National Cancer Institute.

[3]  J. Gart,et al.  On the bias of various estimators of the logit and its variance with application to quantal bioassay. , 1967, Biometrika.

[4]  P C Elwood,et al.  A Randomized Controlled Trial of Acetyl Salicyclic Acid in the Secondary Prevention of Mortality from Myocardial Infarction , 1974, British medical journal.

[5]  J. Fleiss.  Statistical methods for rates and proportions , 1974 .

[6]  E. Veys,et al.  Letter: HL-A and infective sacroileitis. , 1974, Lancet.

[7]  W. Gruen.  Effects of brief psychotherapy during the hospitalization period on the recovery process in heart attacks. , 1975, Journal of consulting and clinical psychology.

[8]  R. R. Hart.  Recovery of Open Heart Surgery Patients as a function of a taped hypnotic induction procedure , 1975 .

[9]  G. Glass.  Primary, Secondary, and Meta-Analysis of Research1 , 1976 .

[10]  D P Byar,et al.  Tests for homogeneity of effect in an epidemiologic investigation. , 1977, American journal of epidemiology.

[11]  P C Elwood,et al.  Aspirin and secondary mortality after myocardial infarction. , 1979, Lancet.

[12]  K Uberla,et al.  Secondary Prevention of Myocardial Infarction Comparison of Acetylsalicylic Acid, Phenprocoumon and Placebo , 1979, Thrombosis and Haemostasis.

[13]  Aspirin in coronary heart disease. The Coronary Drug Project Research Group. , 1980, Circulation.

[14]  Lippincott Williams Wilkins,et al.  Persantine and aspirin in coronary heart disease. The Persantine-Aspirin Reinfarction Study Research Group. , 1980, Circulation.

[15]  W. Grove.  Statistical Methods for Rates and Proportions, 2nd ed , 1981 .

[16]  T C Chalmers,et al.  A method for assessing the quality of a randomized control trial. , 1981, Controlled clinical trials.

[17]  Norman E. Breslow,et al.  Odds ratio estimators when the data are sparse , 1981 .

[18]  The precision of the Mantel-Haenszel estimator in case-control studies with multiple matching. , 1982, American journal of epidemiology.

[19]  J. Fleiss.  The Mantel-Haenszel estimator in case-control studies with varying numbers of controls matched to each case. , 1984, American journal of epidemiology.

[20]  G. Glass,et al.  A new look at evidence about reduced cost of medical utilization following mental health treatment. , 1984, The American journal of psychiatry.

[21]  R. Peto,et al.  Beta blockade during and after myocardial infarction: an overview of the randomized trials. , 1985, Progress in cardiovascular diseases.

[22]  L. Hedges,et al.  Statistical Methods for Meta-Analysis , 1987 .

[23]  S Greenland,et al.  A general estimator for the variance of the Mantel-Haenszel odds ratio. , 1986, American journal of epidemiology.

[24]  N. Laird,et al.  Meta-analysis in clinical trials. , 1986, Controlled clinical trials.

[25]  N Breslow,et al.  Estimators of the Mantel-Haenszel variance consistent in both sparse data and large-strata limiting models. , 1986, Biometrics.

[26]  J. Simes,et al.  Total parenteral nutrition and cancer clinical trials , 1986, Cancer.

[27]  K. Bailey,et al.  Inter-study differences: how should they influence the interpretation and analysis of results? , 1987, Statistics in medicine.

[28]  T. Chalmers,et al.  Meta-analyses of randomized controlled trials. , 1987, The New England journal of medicine.

[29]  R Peto,et al.  Why do we need systematic overviews of randomized trials? , 1987, Statistics in medicine.

[30]  D L Demets,et al.  Methods for combining randomized clinical trials: strengths and limitations. , 1987, Statistics in medicine.

[31]  R. Brand,et al.  Total parenteral nutrition and sepsis. , 1988, Archives of disease in childhood.

[32]  L. Bolognese,et al.  RANDOMISED TRIAL OF INTRAVENOUS STREPTOKINASE, ORAL ASPIRIN, BOTH, OR NEITHER AMONG 17 187 CASES OF SUSPECTED ACUTE MYOCARDIAL INFARCTION: ISIS-2 , 1988, The Lancet.

[33]  M. Jenicek,et al.  Meta-analysis in medicine. Where we are and where we want to go. , 1989, Journal of clinical epidemiology.

[34]  S Greenland,et al.  Bias in the one-step method for pooling study results. , 1990, Statistics in medicine.

[35]  K. Dickersin,et al.  Identification of meta-analyses. The need for standard terminology. , 1990, Controlled clinical trials.

[36]  S. Pocock,et al.  Can meta-analyses be trusted? , 1991, The Lancet.

[37]  J. Fleiss,et al.  Meta-analysis in epidemiology, with special reference to studies of the association between exposure to environmental tobacco smoke and lung cancer: a critique. , 1991, Journal of clinical epidemiology.

[38]  C D Naylor,et al.  Incorporating variations in the quality of individual randomized trials into meta-analysis. , 1992, Journal of clinical epidemiology.

[39]  J. Cutler,et al.  Variance imputation for overviews of clinical trials with continuous response. , 1992, Journal of clinical epidemiology.

[40]  D. Felson,et al.  Bias in meta-analytic research. , 1992, Journal of clinical epidemiology.