When should global health actors prioritise more uncertain interventions?

Global health actors use economic evaluations, including cost-effectiveness analyses, to estimate the effect of different interventions they might fund. However, producing reliable cost-effectiveness estimates is difficult, meaning organisations must often choose between funding interventions for which reliable predictions of efficacy exist and those for which they do not. In practice, many organisations appear to be risk-averse, favouring more certain interventions simply because they are more certain. We argue that this practice is not justifiable. Prioritising projects backed by greater evidence might often produce greater health benefits. However, a general tendency to prefer more certain interventions will cause global health actors to overlook opportunities to help less well-studied populations, support promising but complex interventions, address the upstream causes of illness, and conduct the most important impact evaluations. We argue that global health actors should instead adopt nuanced attitudes towards uncertainty and be willing to fund highly uncertain interventions in some cases. We further describe the considerations they should take into account in rendering these judgements.