Overview of topic: Whenever individual level data are used as the basis of a statistical modelling, partially or fully unobserved (often referred to as “missing”) data is likely to limit the amount of information available from the evidence which might lead to biased inferences. While health economics and outcome research (HEOR) is no exception, there are also other context-specific issues and complexities that often make the modellers work harder, with an even higher potential for biased inferences made from the data and impacting the whole decision making process that characterises HEOR.

Building on previous work developed by the ISPOR Missing Data in HEOR Working Group, we have updated a comprehensive literature review to understand methodological approaches used to account for missing data in cost-effectiveness analyses. Studies were eligible only if they focused on addressing missing data for costs, Health Related Quality of Life (HRQoL) measures, or utility. We have also collected and critically reviewed current guidelines and suggestions on how to deal with missing data, specifically in HEOR.

Leveraging the findings of the review we have produced a characterization of different methods available for handling missing data in HEOR and have provided key strengths and limitations of different classes of methods.

Finally, we have consolidated this information into a generalized practical framework for implementing best practices focused on the reporting of missing data and methodological considerations/justification for accounting for missing data. The objective and goal of this forum will be to focus on presenting the key findings of this project, potential recommendations, and have a discussion to gain feedback on a potential framework for dealing with missing data.