Expert elicitation is often used in economic analyses to estimate uncertain or unobserved parameters for decision models. However, it has rarely been used in the context of local decision-making. A pragmatic elicitation process was used during a local economic evaluation to prompt local experts to assess the relevance of the published evidence to their setting, and to adjust the published effect estimates to better reflect the intervention effect expected in their setting.

Elicitation was undertaken for two interventions that targeted the prevention of hospital-acquired hypoglycemia. Six clinical experts from within the Southern Adelaide Local Health Network (SALHN) were systematically presented with information on the setting of the published evaluation and their local setting. This included information on the hospital and quality of care, patient characteristics, and the research context. After comparing the settings, the experts were asked to estimate the most realistic, most pessimistic, and most optimistic intervention effects for their local context.

The local intervention effect was estimated to be smaller than the published estimate for both interventions. For one intervention, this was driven by the lower complexity of the local patient cohort. For the other intervention, it was driven by differences in the scope of implementation, with hospital-wide local implementation expected to reduce staff buy-in relative to the targeted implementation used in the published evaluation. The elicited local intervention effects were used in a cost-consequence analysis to estimate the likely costs and effects of the interventions if they were implemented locally.

The pragmatic elicitation process provides a feasible and acceptable way to assess and transparently adjust the published effect estimates to better reflect the expected intervention effect in the local setting.

Including this step in local economic evaluations can increase the relevance of these evaluations to local decision makers. Further development and application of these methods may facilitate greater use of economic evaluation in local settings.