Meta-analysis with missing study-level sample variance data

News

  • Author: Statistics Views, Amit K. Chowdhry, Robert H. Dworkin, Michael P. McDermott
  • Date: 10 May 2016
  • Copyright: Image appears courtesy of Getty Images

The authors of an article just published in Statistics in Medicine have considered a study-level meta-analysis with a normally distributed outcome variable and possibly unequal study-level variances, where the object of inference is the difference in means between a treatment and control group.

They clarify their paper in further detail below.

thumbnail image: Meta-analysis with missing study-level sample variance data

Meta-analysis is used widely in a variety of fields to combine results from multiple independent studies to arrive at an overall conclusion concerning the effect of a treatment. This paper considers a study-level meta-analysis with a normally distributed outcome variable and possibly unequal study-level variances, where the object of inference is the difference in means between a treatment and a parallel control group.

A common complication in such an analysis is missing sample variances (or standard deviations) for some studies. A frequently-used approach is to impute, or fill in, the missing values with the weighted (by sample size) mean of the observed variances (mean imputation). Another approach is to include only those studies with non-missing variances (complete case analysis). Both mean imputation and complete case analysis, however, are only valid under the missing-completely-at-random assumption (MCAR), which states that whether or not an observation is missing is independent of the value of the observation.

Even in this case the weights used in mean imputation are not necessarily optimal. In this paper, a multiple imputation method is proposed employing gamma meta-regression to impute the missing sample variances. This method takes advantage of study-level predictors that may be used to provide information about the missing data. Through simulation studies, it is shown that the multiple imputation approach, when the imputation model is correctly specified, is superior to competing methods in terms of type I error and confidence interval coverage probabilities. Finally, a similar approach to handling missing variances in cross-over studies is described.

To read the paper in full, please visit the link below where the article is available in Early View:

Meta-analysis with missing study-level sample variance data

Amit K. Chowdhry, Robert H. Dworkin, Michael P. McDermott

DOI: 10.1002/sim.6908

Related Topics

Related Publications

Related Content

Site Footer

Address:

This website is provided by John Wiley & Sons Limited, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ (Company No: 00641132, VAT No: 376766987)

Published features on StatisticsViews.com are checked for statistical accuracy by a panel from the European Network for Business and Industrial Statistics (ENBIS)   to whom Wiley and StatisticsViews.com express their gratitude. This panel are: Ron Kenett, David Steinberg, Shirley Coleman, Irena Ograjenšek, Fabrizio Ruggeri, Rainer Göb, Philippe Castagliola, Xavier Tort-Martorell, Bart De Ketelaere, Antonio Pievatolo, Martina Vandebroek, Lance Mitchell, Gilbert Saporta, Helmut Waldl and Stelios Psarakis.