Supply-induced demand for hospital care.

Preliminary analysis of hospital utilization data indicates that Roemer's Law may still be operative in rural Iowa: counties with more hospital beds per capita have more hospital utilization per capita. However, when patient origin data are analyzed findings are entirely different. Bed supply is not related to utilization rates in Iowa counties. Instead, the number of unique hospital services is associated with higher utilization rates. The impact of this characteristic of hospital supply, however, is much weaker than the original Roemer effect. The contradiction of these findings with Roemer's Law is apparently a result of a methods effect: use of hospital utilization data which are not derived from actual population experience reveals a relationship which is a statistical artifact. The data also reveal that suburban counties have higher utilization rates than either rural or urban (MSA core) counties.