Skip to content


  • Poster presentation
  • Open Access

Challenges of implementing multiple imputation to address missing data in economic evaluations

  • 1, 2,
  • 1, 2,
  • 1,
  • 1, 2,
  • 1,
  • 1 and
  • 1
Trials201516 (Suppl 2) :P26

  • Published:


  • Economic Evaluation
  • School Child
  • Multiple Imputation
  • Case Analysis
  • Imputation Method


Missing cost and outcomes data is a common problem for health economists analysing RCTs. Systematic reviews of the economic evaluation literature suggest that many analysts use crude methods (e.g. complete case analysis or single imputation) which are vulnerable to bias and may underestimate uncertainty. Multiple imputation (MI) is more appropriate if data are missing at random, but raises a number of methodological and practical challenges. Two case study RCTs assessing the cost-effectiveness of new interventions are used to explore these challenges.


The first RCT is a small trial (n=100) of an intervention which aims to improve physical function in patients with a chronic condition. The second is a cluster RCT (n=2221 participants; n=30 clusters) of an intervention which aims to instil healthy practices related to nutrition and exercise in primary school children. We describe challenges at each stage of analysis.


The challenges discussed will include: understanding the prevalence and pattern of missing data; selecting appropriate imputation methods; identifying variables to include in the model and stratifying by random allocation; imputing non-normally distributed costs and bounded EQ-5D values; achieving convergence; accounting for clustering (i.e. multilevel MI); bootstrapping to obtain accelerated bias-corrected confidence intervals; and conducting sensitivity analyses.


We use these case studies to address the challenges of implementing current guidelines for handling missing data in cost-effectiveness analysis as well as to identify gaps in current knowledge and priorities for future research.

Authors’ Affiliations

School of Social and Community Medicine, University of Bristol, Bristol, UK
Bristol Randomised Trials Collaboration, Bristol, UK


© Fernandez et al. 2015

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate. Please note that comments may be removed without notice if they are flagged by another user or do not comply with our community guidelines.