This article describes an implementation in Stata of the MICE method for multiple imputation of missing data, developed by van Buuren, Boshuizen, and Knook (1999). The method, called "multivariate imputation by chained equations," involves creating multiple copies of the dataset with missing values imputed, analyzing each copy separately, and combining the results. The article introduces five Stata ado-files: mvis for creating multiple imputations, uvis for univariate imputation, micombine for fitting regression models to imputed data, and misplit and mijoin for converting datasets between different formats. The methods are illustrated using a breast cancer dataset for prognostic modeling. The article also discusses the importance of using multiple imputation over complete-case analysis, which can lead to significant loss of data. It highlights the need for proper imputation methods that account for the uncertainty in missing data and the use of Rubin's rules for calculating standard errors. The article also provides guidance on selecting the number of imputations (m) to ensure acceptable precision in confidence intervals. It emphasizes the importance of preserving the multivariate structure of the original data in imputations and the potential issues with using random imputation methods. The article concludes with a discussion on the limitations of current methods and the need for further research in this area.This article describes an implementation in Stata of the MICE method for multiple imputation of missing data, developed by van Buuren, Boshuizen, and Knook (1999). The method, called "multivariate imputation by chained equations," involves creating multiple copies of the dataset with missing values imputed, analyzing each copy separately, and combining the results. The article introduces five Stata ado-files: mvis for creating multiple imputations, uvis for univariate imputation, micombine for fitting regression models to imputed data, and misplit and mijoin for converting datasets between different formats. The methods are illustrated using a breast cancer dataset for prognostic modeling. The article also discusses the importance of using multiple imputation over complete-case analysis, which can lead to significant loss of data. It highlights the need for proper imputation methods that account for the uncertainty in missing data and the use of Rubin's rules for calculating standard errors. The article also provides guidance on selecting the number of imputations (m) to ensure acceptable precision in confidence intervals. It emphasizes the importance of preserving the multivariate structure of the original data in imputations and the potential issues with using random imputation methods. The article concludes with a discussion on the limitations of current methods and the need for further research in this area.