|
Citation
|
Abraham PJ, Crowley BM, Moore D, Stephens S, Minor M, Griffin RL, Hashmi Z, Holcomb JB, Kerby JD, Jansen JO. J. Trauma Acute Care Surg. 2021; ePub(ePub): ePub. |
|
Abstract
|
BACKGROUND: Understanding geographic patterns of injury is essential to operating an effective trauma system and targeting injury prevention. Choropleth maps are helpful in showing spatial relationships, but are unable to provide estimates of spread or degrees of confidence. Funnel plots overcome this issue and are a recommended graphical aid for comparisons that allow quantification of precision. The purpose of this project was to demonstrate the complementary roles of choropleth maps and funnel plots in providing a thorough representation of geographic trauma data.
METHODS: This is a retrospective analysis of EMS transport data of adult patients in Alabama from July 2015 through June 2020. Choropleth maps of case volume and observed-to-expected ratios of incidence were created utilizing U.S. Census Bureau data. Funnel plots were created to relate incidence rate to county population. Subgroup analyses included patients with critical physiology, penetrating, blunt, and burn injuries.
RESULTS: We identified 65,247 trauma incidents during the study period. The overall statewide incidence rate was 133 per 10,000 persons. The highest number of incidents occurred in the most populous counties (Jefferson - 10,768, Mobile - 5,642). Choropleth maps for overall incidence and subgroups highlighted that spatial distribution of overall case volume and observed-to-expected ratios are not always congruent. Funnel plots identified possible and probable outliers, and revealed skewed or otherwise unique patterns among injury subgroups.
CONCLUSIONS: This study demonstrates the complementarity of choropleth maps and funnel plots in describing trauma patterns. Comprehensive geospatial analyses may help guide a data-driven approach to trauma system optimization and injury prevention. Combining maps of case counts, incidence, and funnel plots helps to not only identify geographic trends in data, but also quantifies outliers and displays how far results fall outside the expected range. The combination of these tools provides a more comprehensive geospatial analysis than either tool could provide on its own. LEVEL OF EVIDENCE: IV Epidemiological.
Language: en |