Disaggregation of the demand for hospital care

In this article we want to confront some of the results of the theoretical literature on aggregation with the empirical consequences of aggregation in the context of the analysis of demand for hospital care. There has been an evolution in the estimation of hospital demand functions from macro to micro studies which has resulted in diverging findings. Most studies of aggregate demand at the regional level estimated extremely large effects of hospital bed supply whereas studies based on individual data typically found health status to have a dominant influence on hospital utilization. Very often, comparison of these studies is difficult because different data sources and model specifications have been used. In this article, we will estimate the same demand functions on both the macro and micro level using one data base in order to facilitate comparison. <BR> For this purpose, we use data on some 230 000 individuals who are insured with a private health insurance company in the Netherlands. The empirical analysis presented in Section IV is preceded by a discussion of the peculiarities of hospital demand functions (Section II), and by a theoretical examination of the consequences of aggregation (Section III).