31 Oct 2013 11:00am

Handling incomplete multilevel data using multiple imputation and meta-analysis


Multilevel data are often incomplete, and may be missing either at individual level or at cluster level. For example, in an observational meta-analysis of individual participant data exploring the association between carotid intima media thickness and subsequent risk of cardiovascular events, some relevant confounders were recorded in only 3 of 8 studies, and sporadic missingness also occurred. I will describe methods for tackling systematically missing covariates in this study by combining partly adjusted and fully adjusted analyses in a multivariate meta-analysis. I will also describe algorithms for multilevel multiple imputation which exploit two-stage meta-analysis methods.