"A built bed is a filled bed?" An empirical re-examination.

This article provides an empirical re-examination of the relationship between regional hospital bed supply and the utilization of hospital care. It tests the hypothesis that the divergence of findings between studies based on micro-data (at the individual level) and those based on macro-data (at the regional level) is due to aggregation and specification bias. The main conclusion is that neither source of bias can account for the observed differences. Some other possible explanations are put forward. Regardless of the level of aggregation, a positive effect is found of bed supply on length of hospital stay but not on admission rates. This may be the result of major changes which have taken place in the financing of hospital services in the Netherlands during the last decade.