Michael A Horst1, Brian W Gross, Alan D Cook, Turner M Osler, Eric H Bradburn, Frederick B Rogers. 1. From the Research Institute (M.H.), Lancaster General Health/Penn Medicine; Trauma Services (B.G., E.B., F.R.), Penn Medicine Lancaster General Health, Lancaster, Pennsylvania; Trauma Research Program (A.C.), Chandler Regional Medical Center, Chandler, Arizona; Department of Surgery (T.O.), University of Vermont College of Medicine, Burlington, Vermont.
Abstract
BACKGROUND: Trauma system expansion is a complex process often governed by financial and health care system imperatives. We sought to propose a new, informed approach to trauma system expansion through the use of geospatial mapping. We hypothesized that geospatial mapping set to specific parameters could effectively identify optimal placement of new trauma centers (TC) within an existing trauma system. METHODS: We used Pennsylvania Trauma Systems Foundation registry data of adult (age, ≥ 15 years) trauma for calendar years 2003 to 2015 (n = 408,432), hospital demographics, road networks, and US Census data files. We included TCs and zip codes outside of Pennsylvania to account for edge effects with trauma cases aggregated to the zip code centroid of residence. Our model assumptions included existing Pennsylvania Trauma Systems Foundation Level I and II TCs, a maximum travel time of 60 minutes to the TC, capacity based on mean statewide ratios of trauma cases per hospital bed size, Injury Severity Score, candidate hospitals with 200 or more licensed beds and 30 minutes or longer or 15 minutes or longer from an existing TC in nonurban/urban areas, respectively. We used the Network Analyst Location-Allocation function in ArcGIS Desktop to generate spatial models. RESULTS: Of the 130 candidate sites, only 14 met the bed size and travel time criteria from an existing TC. Approximately 70% of zip codes and 91% of cases were within 60 minutes of an existing TC. Adding one to six new optimally paced TCs increased to a maximum of 82% of zip codes and 96% of cases within 60 minutes of an existing TC. Changes to model assumptions had an impact on which candidate sites were selected. CONCLUSION: Intelligent trauma system design should include an objective process like geospatial to determine the optimum locations for new TCs within existing trauma networks. LEVEL OF EVIDENCE: Epidemiological study, level III.
BACKGROUND:Trauma system expansion is a complex process often governed by financial and health care system imperatives. We sought to propose a new, informed approach to trauma system expansion through the use of geospatial mapping. We hypothesized that geospatial mapping set to specific parameters could effectively identify optimal placement of new trauma centers (TC) within an existing trauma system. METHODS: We used Pennsylvania Trauma Systems Foundation registry data of adult (age, ≥ 15 years) trauma for calendar years 2003 to 2015 (n = 408,432), hospital demographics, road networks, and US Census data files. We included TCs and zip codes outside of Pennsylvania to account for edge effects with trauma cases aggregated to the zip code centroid of residence. Our model assumptions included existing Pennsylvania Trauma Systems Foundation Level I and II TCs, a maximum travel time of 60 minutes to the TC, capacity based on mean statewide ratios of trauma cases per hospital bed size, Injury Severity Score, candidate hospitals with 200 or more licensed beds and 30 minutes or longer or 15 minutes or longer from an existing TC in nonurban/urban areas, respectively. We used the Network Analyst Location-Allocation function in ArcGIS Desktop to generate spatial models. RESULTS: Of the 130 candidate sites, only 14 met the bed size and travel time criteria from an existing TC. Approximately 70% of zip codes and 91% of cases were within 60 minutes of an existing TC. Adding one to six new optimally paced TCs increased to a maximum of 82% of zip codes and 96% of cases within 60 minutes of an existing TC. Changes to model assumptions had an impact on which candidate sites were selected. CONCLUSION: Intelligent trauma system design should include an objective process like geospatial to determine the optimum locations for new TCs within existing trauma networks. LEVEL OF EVIDENCE: Epidemiological study, level III.