Willan and Eckermann describe a method for dealing with heterogeneity in value of information (VOI) calculations for prioritising and designing new research. Their article raises three fundamental (inter-related) issues for VOI methods: (1) how to make sense of the concept of uncertainty in a cost-effectiveness analysis (CEA) model, (2) the interpretation of heterogeneity in CEA, and (3) the relationship between data from a new study and the CEA model when there is heterogeneity. We discuss these three issues using an illustrative example meta-analysis of magnesium for myocardial infarction. Careful consideration of the relationship between existing (and future) evidence and the CEA model is required to provide practical VOI methods that can help research funders prioritise new research in the face of heterogeneity. Copyright © 2011 John Wiley & Sons, Ltd.