O3019: Geocoding Pediatric Sepsis “hot-spots” Using Electronic Clinical Data

Background: Pediatric sepsis is consistently a leading cause of morbidity and mortality in children. Causes of sepsis are multifactorial and can be exacerbated by environmental factors. Using sepsis encounters identified by an electronic algorithm, we aimed to describe the spatial distribution of sepsis patients admitted to our center and identify geographic “hot-spots."

Methods: This was a retrospective study including patients with clinically-defined sepsis at our large, quaternary academic center from January 20, 2011 to May 20, 2021. Sepsis patients were identified by previously validated algorithm which reflects presence of infection with organ dysfunction based on data in the electronic health record. We included hospital-level (setting where sepsis occurred) as well as patient-level characteristics (age, sex, presence of complex chronic condition, and insurance type). Using patient ZIP code, we mapped patients with sepsis recognized in the emergency department to better assess the catchment area of community sepsis (versus sepsis which occurred in the hospital or was referred). We additionally mapped all encounters (regardless of setting) for Philadelphia. The encounter rate was reflective of episodes during entire study period per 10,000 children under 18 years of age, based on 2019 American Community Survey data. Geospatial analyses and visualizations were performed using ArcGIS Pro 2.8 (Esri, Redlands, California).

Results: Among 7,288 sepsis episodes, there were 4,532 unique patients. Using patients’ most recent sepsis episode as the index encounter, the mortality rate was 9.7%. The median age was 7 (IQR 2-14) years with 44% identified as Non-Hispanic White, 28% Non-Hispanic Black, 14% Hispanic/Latino and 15% as Other. Roughly half (48%) were privately insured, with the remainder publicly insured ( < 1% self-pay/charity care). A majority of patients (76%) had presence of at least one complex chronic condition. Nearly half of index sepsis episodes occurred in the emergency department (45%) with the remainder occurring in the ICU (16%), inpatient floors (16%), outside hospital transfers (19%) and perioperative settings (3.7%). Figure 1 shows the encounter rates for patients seen in the emergency department, demonstrating the large catchment area of patients presenting to the ED. Figure 2 shows visualization of encounter rates in Philadelphia, illustrating two ZIP codes with >120 encounters per 10,000 children.

Conclusion: Geospatial mapping of patients with sepsis using electronic clinical data shows areas of high sepsis prevalence and illustrates that many patients travel long distances before presenting to ED, which has implications for potential delays in recognition/treatment. Further research is needed to assess what, if any, specific environmental factors/social determinants may contribute to geographic hot-spots, as well as to assess outcomes of patients who travel further and/or bypass local hospitals in order to seek care in our ED. Investigating these factors has implications for local sepsis policy and practice.