We model the locational determinants of nine categories of healthcare services in the contiguous United States using restricted access federal establishment data. These data enable close examination of rural health services, which are subject to suppression in publicly published data sources. After reviewing differences in public and unsuppressed restricted data and testing underlying data generation processes for each healthcare industry, including the Poisson, negative binomial, and their zero-inflated counterparts, we estimate marginal effects for four categories of independent variables: place-based factors, financial access, characteristics of population, and industry interdependencies. Findings show establishments are less likely to be found with high concentrations of Medicare and Medicaid recipients, while agglomerations are associated with more establishments. Nonemployer establishments serve a broader spectrum of people, but the rural poor still experience less access to health care.

中文翻译：

---

用受限的联邦行政数据估计医疗机构位置的决定因素

我们使用受限访问联邦机构数据对美国本土九类医疗保健服务的位置决定因素进行建模。这些数据有助于仔细检查农村卫生服务，这些服务在公开发布的数据源中受到压制。在审查了公共和未抑制受限数据的差异并测试了每个医疗保健行业的基础数据生成过程（包括泊松、负二项式及其零膨胀对应项）后，我们估计了四类自变量的边际效应：基于地点的因素、金融渠道、人口特征和行业相互依存关系。调查结果显示，医疗保险和医疗补助接受者集中的场所不太可能出现，而聚集与更多的机构相关联。非雇主机构为更广泛的人群提供服务，但农村贫困人口获得医疗保健的机会仍然较少。