Multiple imputation of covariates by substantive-model compatible fully conditional specification
Jonathan W. Bartlett
Department of Medical Statistics
London School of Hygiene and Tropical Medicine
London, UK
[email protected]
|
Tim P. Morris
MRC Clinical Trials Unit at UCL
Institute of Clinical Trials and Methodology
and
London School of Hygiene and Tropical Medicine
London, UK
[email protected]
|
Abstract. Multiple imputation is a practical, principled approach to handling missing
data. When used to impute missing values in covariates of regression models,
imputation models may be misspecified if they are not compatible with the
substantive model of interest for the outcome. In this article, we introduce
the smcfcs command, which imputes covariates by substantive-model
compatible fully conditional specification. This modifies the popular fully
conditional specification or chained-equations approach to multiple imputation
by imputing each covariate compatibly with a user-specified substantive model.
We compare the smcfcs command with standard fully conditional
specification imputation using mi impute chained in a simulation study
and illustrative analysis of data from a study investigating time to tumor
recurrence in breast cancer.
View all articles by these authors:
Jonathan W. Bartlett, Tim P. Morris
View all articles with these keywords:
smcfcs, multiple imputation, substantive model compatible, congenial, interactions, nonlinearities
Download citation: BibTeX RIS
Download citation and abstract: BibTeX RIS
|
