A1127
Title: Hierarchical imputation of categorical variables in the presence of systematically and sporadically missing data
Authors: Shahab Jolani - Maastricht University (Netherlands) [presenting]
Abstract: In the development of prediction models, data are often combined from different sources, known as individual participants data (IPD) sets. A specific challenge in analysing IPD sets is the presence of systematically missing data when certain variables are not measured in some studies and sporadically missing data when measurements of certain variables are incomplete across different studies. Multiple imputation (MI) is among the better approaches to deal with missing data. However, MI of clustered data, such as IPD meta-analysis, requires advanced imputation routines that preserve the hierarchical structure of data and accommodate both systematically and sporadically missing data. A new class of hierarchical imputation methods have been recently developed within the MICE framework tailored for continuous variables. The extension of these methods to categorical variables is discussed, accommodating the simultaneous presence of systematically and sporadically missing data in nested designs with arbitrary missing data patterns. To address the challenge of the categorical nature of the data, an accept-reject algorithm is proposed during the imputation process. Following theoretical discussions, the performance of the new methodology is evaluated through simulation studies and its application is demonstrated using an IPD set from patients with kidney failure.
