Characterization of Regional Poison Center Utilization Through Geospatial Mapping
- Author(s): Olives, Travis D.;
- Westgard, Bjorn C.;
- Steinberg, Lila W.;
- Cole, Jon B.
- et al.
Published Web Locationhttps://doi.org/10.5811/westjem.2020.7.46385
Background: Penetrance is the annual rate of human exposure calls per 1000 persons, a measurethat historically describes poison center (PC) utilization. Penetrance varies by sociodemographiccharacteristics and by geography. Our goal in this study was to characterize the geospatial distributionof PC calls and describe the contribution of geospatial mapping to the understanding of PC utilization.
Methods: This was a single-center, retrospective study of closed, human, non-healthcare facilityexposure calls to a regional PC over a five-year period. Exposure substance, gender, age, andzone improvement plan (ZIP) Code were geocoded to 2010 US Census data (household income,educational attainment, age, primary language) and spatially apportioned to US census tracts, andthen analyzed with linear regression. Penetrance was geospatially mapped and qualitatively analyzed.
Results: From a total of 304,458 exposure calls during the study period, we identified 168,630non-healthcare exposure calls. Of those records, 159,794 included ZIP Codes. After exclusions,we analyzed 156,805 records. Penetrance ranged from 0.081 - 38.47 calls/1000 population/year(median 5.74 calls/1000 persons/year). Regression revealed positive associations between >eighthgradeeducational attainment (β = 5.05, p = 0.008), non-Hispanic Black (β = 1.18, p = 0.032) andAmerican Indian (β = 3.10, p = 0.000) populations, suggesting that regions with higher proportionsof these groups would display greater PC penetrance. Variability explained by regression modellingwas low (R2 = 0.054), as anticipated. Geospatial mapping identified previously undocumentedpenetrance variability that was not evident in regression modeling.
Conclusion: PC calls vary substantially across sociodemographic strata. Higher proportionsof non-Hispanic Black or American Indian residents and >eighth-grade educational attainmentwere associated with higher PC call penetrance. Geospatial mapping identified novel variationsin penetrance that were not identified by regression modelling. Coupled with sociodemographiccorrelates, geospatial mapping may reveal disparities in PC access, identifying communities atwhich PC resources may be appropriately directed. Although the use of penetrance to describe PCutilization has fallen away, it may yet provide an important measure of disparity in healthcare accesswhen coupled with geospatial mapping.