Geospatial Analysis of Pediatric EMS Run Density and Endotracheal Intubation

Introduction: The association between geographic factors, including transport distance, and pediatric emergency medical services (EMS) run clustering on out-of-hospital pediatric endotracheal intubation is unclear. The objective of this study was to determine if endotracheal intubation procedures are more likely to occur at greater distances from the hospital and near clusters of pediatric calls. Methods: This was a retrospective observational study including all EMS runs for patients less than 18 years of age from 2008 to 2014 in a geographically large and diverse Oregon county that includes densely populated urban areas near Portland and remote rural areas. We geocoded scene addresses using the automated address locator created in the cloud-based mapping platform ArcGIS, supplemented with manual address geocoding for remaining cases. We then use the Getis-Ord Gi spatial statistic feature in ArcGIS to map statistically significant spatial clusters (hot spots) of pediatric EMS runs throughout the county. We then superimposed all intubation procedures performed during the study period on maps of pediatric EMS-run hot spots, pediatric population density, fire stations, and hospitals. We also performed multivariable logistic regression to determine if distance traveled to the hospital was associated with intubation after controlling for several confounding variables. Results: We identified a total of 7,797 pediatric EMS runs during the study period and 38 endotracheal intubations. In univariate analysis we found that patients who were intubated were similar to those who were not in gender and whether or not they were transported to a children's hospital. Intubated patients tended to be transported shorter distances and were older than non-intubated patients. Increased distance from the hospital was associated with reduced odds of intubation after controlling for age, sex, scene location, and trauma system entry status in a multivariate logistic regression. The locations of intubations were superimposed on hot spots of all pediatric EMS runs. This map demonstrates that most of the intubations occurred within areas where pediatric EMS calls were highly clustered. By mapping the intubation procedures and pediatric population density, we found that intubation procedures were not clustered in a similar distribution to the pediatric population in the county. Conclusion: In this geographically diverse county the location of intubation procedures was similar to the clustering of pediatric EMS calls, and increased distance from the hospital was associated with reduced odds of intubation after controlling for several potential confounding variables.