Predictive modelling supports policy at two levels. The first is the role of predictive modelling to help determine priorities for health policy, such as risk adjustment models for resource allocation, models of future demand for health, models of future health status, and labour market models. The second is how modelling is used to inform decision-making about health service delivery itself, either to groups of patients or to individuals. For instance, current health policy tends to prioritise the identification of the very high intensity users of unplanned secondary care.
Models therefore augment decisions through better use of information. Modelling the future, though, is always uncertain, so predictions are only useful when they provide guidance to decision-makers to some acceptable level of meaning.
In policy, the use of models depends on understanding what decisions the models are to provide help with. — how can models help optimise the decisions we need to make?