Date Presented 4/20/2018
Study findings revealed significant differences in self-care and mobility outcomes for inpatient and skilled nursing stroke rehabilitation in the United States. Clinicians and researchers need to investigate patient and facility factors associated with skilled nursing stroke rehabilitation outcomes
Primary Author and Speaker: Timothy Reistetter
Additional Authors and Speakers: Ickpyo Hong
Contributing Authors: Karl Eschbach, Daniel Jupiter, John Prohaska, Amol Karmarkar, Kenneth Ottenbacher
PURPOSE: Health policymakers have expressed concern about the presence of geographic variation in health care use and cost. Studies by the Institute of Medicine (2012, 2013) suggest that postacute care accounts for 73% of the variation in health care costs. Policymakers have suggested that minimizing geographic variation is a viable method for improving quality of care while decreasing cost. Studies of geographic variation have predominantly used regions or boundaries based on acute care hospital utilization. There are no specific geographic boundaries based on delivery of rehabilitation. Therefore, the purpose of this project was to develop and test a method for defining rehabilitation service areas (RSAs) in the United States that would allow for accurate studies of variation in rehabilitation to facilitate efforts to improve rehabilitation outcomes and quality of care.
METHOD: This cross-sectional design study used 100% 2010 Texas Medicare acute care admission data for 893,028 patients across 335 hospitals. We linked Medicare claims data from the MedPAR, Master Beneficiary, and Provider of Service files with zip code census data. Our design used a proof of concept approach. We chose to test our algorithm for developing service areas on acute care data because the delivery pattern for hospitals was well established by the Dartmouth Atlas Group (1996). Our hypothesis was that our clustering algorithm based on acute care data would approximate the Dartmouth Atlas hospital referral regions, which would suggest that our analytical approach can be used to develop RSAs.
For each patient, we obtained the home zip code and the hospital zip code where they received care. We used the ZIP Code Tabulation Area (ZCTA) file to obtain accurate geographic positions and to calculate the internal geographic centroid for each ZCTA using ArcGIS Version 10.3 (Esri, Redlands, CA). These zip codes were then clustered using Ward’s algorithm. Finally, the clustered zip codes were mapped along with the Dartmouth Atlas Hospital Referral Regions (HRR) using ArcGIS.
RESULTS: Service areas identified by our algorithm yielded 20 regions. Compared with the 22 Dartmouth Atlas regions, our clustering algorithm closely matched established HRR boundaries.
CONCLUSION: Study findings support the use of our clustering algorithm for developing RSAs. Health care service areas are useful tools for the study of small area variations in care, but none have been identified to reflect the unique service patterns for postacute care rehabilitation. A next step is to evaluate the utility of RSAs in describing patterns of community discharge, length of stay, and readmissions after adjusting for individual, community, and facility characteristics. Further research is needed comparing RSAs to other geographic boundaries used by policymakers to better understand and document the value of rehabilitation services for reaching optimal outcomes across the country.
IMPACT STATEMENT: The number of people requiring postacute care rehabilitation is expected to grow as the population ages. As policymakers move forward with reform efforts for improving transitions through postacute services, it is critical to have a tool (RSAs) for understanding variation in the use, outcomes, and quality of care. This study aligns with the American Occupational Therapy Association’s health services research priority and highlights the need for effectiveness studies examining the delivery and outcomes of rehabilitation services provided across the United States.
References
Dartmouth Atlas Group. (1996). The Dartmouth atlas of health care. Chicago: American Hospital Publishing
Institute of Medicine. (2012). Geographic adjustment in Medicare payment: Phase II. Implications for access, quality, and efficiency. Washington, DC: National Academies Press.
Institute of Medicine. (2013). Variation in health care spending: Target decision making, not geography. Washington, DC: National Academies Press.
