This article demonstrates that the variation among communities in hospital use, measured by the total patient-day rate, is overwhelmingly associated with the variation in discharge rates. In particular, this variation is primarily attributable to the variation in nonsurgical discharge rates. While there is residual variance associated with variance in length of stay and interaction effects, more than two-thirds of the variance in the patient-day rates is attributable to variance in discharge rates. Further, little variation is demonstrated across communities in total average length of stay. High use communities have high discharge rates which are not explicable in terms of several need-determining characteristics of the community populations. Discharge-rate variation is demonstrated to be strongly associated with differences in the supply of medical care resources--in particular, acute care beds, surgeons, and nonsurgical specialists.