METAMISS: Stata module to perform meta-analysis with missing data
Ian R. White, Julian P. T. Higgins · RePEc: Research Papers in Economics · 2007
metamiss performs meta-analysis with a binary outcome, using data on the number of successes, number of failures and number of missing values by arm. A variety of imputation methods are available, including imputing failures, imputing successes, worst- and best-case. Different imputation schemes may be applied to subgroups with different reported reasons for missing data. The degree of informative missingness may be specified via the informative missingness odds ratio (IMOR) in each group. Finally, uncertainty about the IMORs may be taken into account in a Bayesian analysis. This command should be especially useful for sensitivity analysis. metamiss and metamiss2 are different commands with overlapping functions. Users with binary outcomes including reasons for missing data should use metamiss. Other users should use metamiss2.