Introduction: PulsePoint is a smartphone-based system that alerts layperson responders to out-of-hospital cardiac arrest (OHCA) events in non-residential locations within the United States. These locations can be classified to inform public access defibrillator (PAD) placement and focus responder training and recruitment efforts in high-frequency areas. Our objective was to manually classify the locations of PulsePoint alerts to better understand how PulsePoint performs across a variety of location types. Methods: Records were obtained for each PulsePoint alert in Allegheny County, PA, USA (pop. 1.2M) from 06/24/2016 to 10/28/2020. PulsePoint generates an alert when a potential OHCA occurs in public and is within 400m of at least one responder. Alert data contains address and location name, and the number and locations of nearby responders and PADs. We placed each location into one of the following categories: “education”, “consumer goods/services”, “restaurants/bars”, “transit hubs”, “hotels”, “entertainment venues”, “recreational facilities”, “government/municipal”, “offices”, “industrial/manufacturing”, “religious institutions”, “parks/outdoor spaces”, “sidewalk/street”, and “shelters/supportive housing”. “Unspecified/unknown” was used for locations that could not be classified due to insufficient information. We excluded all outpatient health, assisted-living, daycare, and correctional facilities. For each location category, we calculated the event count and the median number of nearby responders and PADs. Results: There were 994 included events during the capture period. The most frequent location category was “sidewalk/street”, which represented 24 percent of all alerts; the least frequent was “transit hubs”, which represented 0.7 percent of all alerts. “Shelters/supportive housing” had the highest median number of responders (7, IQR 1-14), while “industrial/manufacturing” had the lowest (1, IQR 1-3.5). “Transit hubs” had the highest median number of PADs (31, IQR 31-41), while “parks/outdoor spaces” had the lowest (0.5, IQR 0-1). Conclusion: PulsePoint alert-generating events occur across a wide variety of location types. Further research is needed to understand the accuracy of the system’s location data versus EMS dispatch records. Nonetheless, the data may have implications for system improvement by highlighting which location types have low numbers of nearby responders and PADs and may benefit from strategic PAD placements and training efforts.
Read full abstract