Hospital use varies among populations due to access, socio-demographic characteristics, and overall health care needs. Further, the interaction between populations and health care providers is often mediated by governing bodies, such as Certificate of Need programs, that regulate the supply of health care resources. The intersection of the spatial and aspatial components of access and utilization within a regulated health care market provide the backdrop for this research. The outcomes provide insights that inform future health services research and offer guidance for public policy initiatives. The research approach adopted in this dissertation addresses both methodological and theoretical issues related to the stud} of access and utilization and the nature of the relationship between them. The project is comprised by three sequential studies tied together within the framework of assessing access and utilization in a regulated health system.The first study examines methods used to measure distance among locations. Specifically, the work addresses the theoretical and applied implications of using raster and network data models for identifying areas with limited geographic accessibility. The findings suggest that the network data model provides a more accurate framework for estimating vehicular travel time along roadways, while the raster data model offers advantages in scenarios where roadways are not the primary route of travel. The second study offers a methodology for clustering spatial observations having multiple attribute values. The specific focus of the work is the formation of Hospital Groups, the allocation units used in a state-level methodology for predicting future hospital bed demand. The main outcome of the research is the methodology itself, which provides a substantial advance over the previous methodologies used in health services research by way of its ability to cluster observations based on overall patterns of health care utilization and geographic location, simultaneously. Using knowledge gained from the first two studies, the final portion of the dissertation explores the relationship between the availability of hospital beds and the utilization of hospital services. The focus of the study is Roemer's Law, which states that a hospital bed built is a bed filled. The findings of this study provide strong support for the concept that greater levels of hospital bed availability lead to higher hospital utilization rates. This relationship is confirmed at various levels of data aggregation, demonstrating that the observed impact of availability on utilization is stable across geographic scales of analysis.The main outcomes of this research can be separated into those relating to advancement in health services research and those relating to public policy. From a public policy perspective, this dissertation offers updated methodologies for identifying areas with limited geographic accessibility and grouping health-based observations. In addition, the final study finds strong evidence of the effects of Roemer's Law, thus providing support for the continued regulation of hospital bed availability. This dissertation also contributes significant new knowledge to the field of health services research. The specific salient outcomes include: detailing both the theoretical and applied differences between the raster and network data models for estimating travel time among locations, offering a methodology that simultaneously clusters observations based on comprehensive patterns of utilization and geographic location, and producing compelling, robust evidence that hospital availability has a positive, significant relationship with hospital utilization rates.
Read
