Capitation payments based on prior hospitalizations

In many countries the concept of capitating health care insurers is receiving increasing attention. In a competitive environment, capitation should induce insurers to concentrate more on cost containment instead of indulging in risk selection. The necessary premium‐replacing capitation payments should account for predictable variations in annual per‐person health care expenditures as far as these are related to health status. Various studies have shown that crude capitation models based on e.g. age, sex and place of residence, do not reflect expected costs accurately. This implies inefficient pricing possibly leading to risk selection and windfall profits or losses for insurers, thereby undermining the objectives of a capitation system. Using Dutch micro data on some 200,000 individuals, this article simulates various alternative capitation models based on, among others, diagnostic information from previous hospitalizations. Results suggest that the problems of both risk selection and windfall profits/losses may be mitigated substantially by using this type of information together with data on prior costs. These results are not only relevant for situations where competing insurers are capitated, as intended in the Netherlands, but also when providers are capitated, as in the UK, or when HMOs are capitated, as in the US.