VT - Vermont Rational Service Areas

Data Layer Name: Vermont Rational Service Areas (RSAs)Alternate Name: Vermont RSAsOverview:Rational Service Areas (RSAs), originally developed in 2001 and revised in 2011, are generalized catchment areas relating to the delivery of primary health care services. In Vermont, RSA area delineations rely primarily on utilization data. The methods used are similar to those used by David Goodman to define primary care service areas based on Medicare data, but include additional sources of utilization data. Using these methods, towns were assigned based on where residents are going for their primary care. The process used to delineate Vermont RSAs was iterative. It began by examining utilization patterns based on: (1) the primary care service areas that Goodman had defined for Vermont from Medicare data; (2) Vermont Medicaid assignments of clients to primary care providers; and, (3) responses to the “town of residence”/”town of primary care” questions in the Vermont Behavioral Risk Factor survey. Taking into account the limitations of each of these sources of data, VDH statisticians defined preliminary town centers and were able to assign approximately two/thirds of the towns to a town center. For towns with no clear utilization patterns, they examined mileage from these preliminary centers, and mileage from towns that had primary care physicians. Contiguity of areas was also examined. A few centers were added and others were deleted. After all towns were assigned to a center and mapped, outliers were identified and reviewed by referring to both mileage maps and utilization patterns. Drive time information was not available. In some cases where the mileage map seemed to indicate one center, but the utilization patterns were strongly supportive of another center, utilization was used as a proxy for drive time.Preliminary RSAs were presented to the Vermont Primary Care Collaborative, the Vermont Coalition of Clinics for the Uninsured and other community members for their feedback. Department of Health District Directors from the Division of Community Public Health were also consulted. These groups suggested modifications to the areas based on their experience working in the areas in question. As a result of this review a few centers were added, deleted and combined, and several towns were reassigned. The Vermont Primary Care Collaborative reviewed the final version of RSAs. The result of this process is 38 Rational Service Areas.Given the limitations of the information available for this purpose, the delineation approach was deemed reasonable and has resulted in a set of RSAs that have been widely reviewed and accepted. Because of the iterative process, it is recognized that this is not a "pure" methodology in the sense that someone else attempting to replicate this process would probably not produce exactly the same results. RSAs have been reviewed periodically to keep up with changes in demographics and provider practice locations. One revision occurred in 2011. This 2011 revision took towns that had originally been assigned as using out-of-state providers and reassigned them to Vermont RSAs. Technical Details:Vermont RSAs were defined using 3 sources of primary care utilization data and mileage maps. Each of the data sources had limitations, and these limitations had to be considered as towns were assigned to a RSA. A description of each of these data sources is provided. Medicare utilization data was obtained from the Primary Care Service Areas developed by David Goodman using 1996 and 1997 Medicare Part B and Outpatient files. Thirty-eight primary care service areas were defined for Vermont. The major limitation of these assignments was that they were based on zip codes rather than town boundaries. Many small towns do not have their own zip code, or the town may be divided into multiple zip codes shared with multiple other towns. As the utilization data was reviewed consideration was given to whether the zip code in question represented the town, or whether utilization from that town may have been masked by a larger town's utilization patterns. A second consideration was that the Medicare data used 1996 & 1997 utilization. In areas where there were new practices established after 1997, the Medicare data would not be able to reflect their utilization.Medicaid claims data only included children age 17 and under. The file contained Medicaid clients in 2000 with the town of residence of the client and the town of the primary care provider. The limitation in this file was that although the Medicaid database included a field for the geographic location of the provider separate from the mailing address, after examining the file it was determined that in many cases the mailing address was also being entered into the geographic location. In areas where practices were owned by a larger organization, the utilization patterns could not be determined. For example, in the St. Johnsbury RSA there were practices owned by an out-of-state medical center. Although it is known that there are medicaid providers in some of the towns in that area, all of the utilization was coded to out of state. Therefore the Medicaid data had to be disregarded in this area. The St. Johnsbury RSA was subsequently defined around three town centers (St. Johnsbury, Lyndon, and Danville) because more precise utilization patterns could not be distinguished.The BRFSS data was obtained from the 1998-2000 surveys. Respondents were asked for the town of their primary care provider. The town of residence of the respondent is also collected. These responses represented all Vermonters age 18-64 years old, regardless of type of insurance. The limitation of this data was small number of respondents in the smaller towns. Mileage information was obtained from the Vermont Medicaid program. This mileage information was derived using GIS mapping software to assess all statewide roads. However, drive-time data could not be determined at that time because there was no distinction between primary and secondary roads. The Medicaid program applied GIS mapping software to assign clients to primary care providers using 15 miles as a proxy for 30-minute drive time. This standard was also used in 2001 when the original RSAs were developed.The VDH Public Health Statistics program periodically updates RSA GIS data. (last updated in 2011)

Data and Resources

Additional Info

Field Value
Maintainer Admin_VCGI VCGI
Last Updated July 30, 2019, 21:31 (CDT)
Created July 30, 2019, 21:31 (CDT)