When faced with a traumatic situation that requires medical care—time is crucial. Dr. Michael A. Horst, Director of Research Biostatistics at Lancaster General Research Institute, recently purchased statewide hospitalization data from the Pennsylvania Health Care Cost Containment Council (PHC4) to study the optimal placement of new trauma centers within Pennsylvania’s established trauma system. When his study is complete, the geospatial mapping techniques and models may help direct the future advancement and development of a statewide trauma system that will provide optimal access to quality definitive care to trauma patients in Pennsylvania.
In explaining his study, Dr. Horst notes that there are many areas in Pennsylvania where there is limited access to a trauma center. Combining data from the Pennsylvania Trauma Systems Foundation (data that includes trauma cases transported to designated trauma centers, where the patient tends to be in close proximity to existing trauma centers) with PHC4 data (data that includes statewide trauma cases) provides insight into areas of the state where there are significant numbers of trauma cases that do not end up in a trauma center. By overlaying both datasets, Dr. Horst is aggregating data to the zip code level and using geospatial mapping to assess differences. Additional geospatial data being used includes hospital demographics, road networks with speed limits to calculate travel time and US Census data. Recognizing that patients living in border regions of the state might seek care in facilities outside of Pennsylvania, the project includes trauma centers/hospitals across state lines.
Dr. Horst’s approach centers on two perspectives: 1) optimally, where would new centers be placed given the existing trauma network and 2) if starting with a “clean slate”, where would trauma centers be located to provide optimal access across the state and how would that compare with the existing trauma network? Dr. Horst explains, “To achieve this, we are investigating travel time from zip code areas to trauma centers in various ranges of time, based upon the locations of potential candidate hospitals and capacity of the candidate hospital.” A location-allocation algorithm (a geographic information approach for working with maps and geographic data) is used to iteratively run models to optimally locate trauma centers—either adding to the existing trauma network or assuming a “clean slate”.
Dr. Horst adds, “The algorithm tries to find the optimal location based upon minimizing travel time to the volume of trauma patients in their home zip code and the capacity of the candidate site. We calculated the ratio of adult trauma to adult population to come up with a trauma rate per 1,000 population and have mapped that across Pennsylvania per zip code of patient residence, but are using the actual trauma volume from each zip code as the parameter in our models.”
Early findings of Dr. Horst’s work suggest that there are geographic areas with a greater need for trauma centers—particularly in Pennsylvania’s rural communities where travel times to existing trauma centers and time to definitive care is significantly longer as well as the high volume of severe trauma cases not going to a trauma center in those areas.
Dr. Horst will be presenting two papers using the PHC4 data at the September 2017 annual meeting of the American Association for the Surgery of Trauma and subsequently will prepare a manuscript for publication. For now, Dr. Horst will complete the analyses for this study to see what questions emerge for future research—potentially using PHC4 data to examine other topics of interest relative to trauma.