OBJECTIVES:

This study aimed to estimate a valuation function that represents public preferences for the allocation of health care resources and to demonstrate how this function could inform uncertain funding decisions using the decision-making plane.

METHODS:

A choice experiment was conducted in a representative sample of the English population (N = 1,021), to elicit public preferences for changes in outcomes that result from implementing an intervention. To account for opportunity costs, outcomes were defined as the net change in population health and the net change in medical expenditure. Different linear and non-linear specifications of the valuation function were estimated. The use of the best function was illustrated in a simulation that evaluated a hypothetical intervention.

RESULTS:

Public preferences for resource allocation were best represented by a valuation function with non-constant marginal sensitivities to changes in population health and medical expenditures. This means that compensating the risk of negative outcomes with a chance of positive outcomes, becomes more challenging as the risk increases. For example, a loss of £10,000 due to high medical expenditure could be compensated by a chance of gaining four additional years in good health, while a loss of £20,000 requires nine additional years in good health. Comparing the non-linear specification of the valuation function to a linear specification in a simulation suggests that attaching equal weights to all outcomes can lead to sub-optimal funding decisions at the margin. Specifically, the evaluated intervention was found to provide no additional benefit using a linear valuation function, but was valued positively using the non-linear specification.

CONCLUSIONS:

The findings of this study suggest that preferences for the allocation of health care resources are affected by both the direction of an uncertain outcome and the size of the outcome. This raises concern about current approaches that attach equal weights to stochastic outcomes.