Can we afford to ignore missing data in cost-effectiveness analyses?

Excerpt

In order to inform resource allocation and health technology coverage decisions, cost-effectiveness analyses (CEAs) are increasingly being undertaken alongside clinical trials, and reported in both specialist and general medical journals, e.g. [1‐3]. There appears to be widespread acceptance of the view that such coverage decisions must take into account both the cost of an intervention and its effectiveness and tolerability. However, collecting resource use, and hence cost data on a patient-by-patient basis in clinical trials, often results in missing data, both in retrospective and prospective cost studies, e.g. [3‐5]. Failing to account for the missing cost data properly can produce biased results due to the potential unrepresentativeness of the patients with complete data. This article outlines the importance of not ignoring, and dealing appropriately with, the missing data problem when conducting a trial-based CEA. …