|
Abstract
|
BACKGROUND: Geographic information systems (GIS) are often used to analyze trauma systems. GIS-based approaches can model access to a trauma center (TC), including estimates of transport time and population coverage, when accurate trauma registry and EMS data are not available. We hypothesized that estimates of trauma system performance calculated using a standard GIS method with public data would be comparable to trauma registry data.
METHODS: A standardized GIS-based method was used to estimate metrics of TC access in a regional trauma system in which the number of TC increased from one to three over a 3-year period. Registry data from the index TC in the system was evaluated for different time periods during this evolution. The number of admissions to the TC in different time periods was compared to changes predicted by the GIS-based model, and the distribution of observed ground-based transportation times was compared to the predicted distribution.
RESULTS: With the addition of two TC to the system, the volume of patients transported by ground to the index TC decreased by 30%. However, the model predicted a 68% decrease in population having the shortest predicted transport time to the index TC. The model predicted the geographic trend seen in the registry data, but many patients were transported to the index TC even though it was not the closest center. Observed transport times were uniformly shorter than predicted times.
CONCLUSIONS: The GIS-based model qualitatively predicted changes in distribution of trauma patients, but registry data highlight that field triage decisions are more complex than model assumptions. Similarly, transport times were systematically overestimated. This suggests that model assumptions, such as vehicle speed, based on normal traffic may not fully reflect EMS operations. There remains great need for metrics to guide policy based on widely available data.
LEVEL OF EVIDENCE: Level III, retrospective observational study.
Language: en |