A new command, metamiss, performs meta-analysis with binary outcomes when some or all studies have missing data. Missing values can be imputed as successes, as failures, according to observed event rates, or by a combination of these according to reported reasons for the data being missing. Alternatively, the user can specify the value of, or a prior distribution for, the informative missingness odds ratio.
ForsterJ. J., and SmithP. W. F.1998. Model-based inference for categorical survey data subject to non-ignorable nonresponse. Journal of the Royal Statistical Society, Series B (Statistical Methodology) 60: 57–70.
2.
GambleC., and HollisS.2005. Uncertainty method improved on best–worst case analysis in a binary meta-analysis. Journal of Clinical Epidemiology58: 579–588.
3.
HigginsJ. P. T., WhiteI. R., and WoodA. M.2008. Imputation methods for missing outcome data in meta-analysis of clinical trials. Clinical Trials5: 225–239.
4.
JoyC. B., AdamsC. E., and LawrieS. M.2006. Haloperidol versus placebo for schizophrenia. Cochrane Database of Systematic Reviews4: CD003082.
5.
LittleR. J. A., and RubinD. B.2002. Statistical Analysis with Missing Data. 2nd ed. Hoboken, NJ: Wiley.
6.
RubinD. B.1977. Formalizing subjective notions about the effect of nonrespondents in sample surveys. Journal of the American Statistical Association72: 538–543.
7.
WhiteI. R., HigginsJ. P. T., and WoodA. M.2008. Allowing for uncertainty due to missing data in meta-analysis - Part 1: Two-stage methods. Statistics in Medicine27: 711–727.
