Implementing multiple imputation with sensitivity analysis in large-scale longitudinal studies

Principal Investigator: 
Assoc. Prof. Katherine Lee

Missing data are ubiquitous in longitudinal studies. In particular, as studies become more complex, researchers are faced with ever-increasing problems of drop-out and missing data. It is imperative that missing data be handled appropriately in the statistical analysis to minimise the risk of biased findings and reduced precision.

Multiple imputation (MI) has become a popular approach for dealing with missing data and is now commonly requested by editors and journal reviewers. Under this approach missing values are imputed multiple times using a regression (imputation) model based on the available data and resulting inference combined across the completed datasets, producing results that recover information from incomplete records while incorporating the uncertainty created by the missing data.  However, currently available techniques for conducting MI are severely limited in their applicability to large-scale longitudinal as 1) they often cannot handle the large number of variables that need to be included in the imputation model and do not exploit potentially informative trends over time within variables, and 2) the standard application of MI relies on the assumption that data are missing at random (MAR) i.e. that the likelihood of a value being missing does not depend on unobserved data. This assumption that is often questionable in practice. This research aims to address these gaps in the current literature by:

1)      developing and evaluating novel algorithms for imputing complex, longitudinal data, and

2)      developing approaches for conducting sensitivity analyses within the MI framework when data are suspected to be missing not at random (MNAR).



  • Prof Melissa Wake (Centre for Community Child Health, MCRI)
  • Prof George Patton (Adolescent Health Research, MCRI)
  • Prof Dallas English (Melbourne School of Population and Global Health)
  • Prof James Carpenter (London School of Hygiene & Tropical Medicine)
  • Missing Data, Imputation & Analysis (MiDIA) group, U.K. (including researchers at the London School of Hygiene & Tropical Medicine, MRC Biostatistics Unit Cambridge, and University of Bristol)