Can you get there from here? An analysis of walkability among PulsePoint CPR alert dispatches

Bystander CPR before the arrival of EMS is a major factor in out of hospital cardiac arrest (OHCA) survival. To recruit trained bystanders, mobile phone-based alert systems have been developed, but their limitations are not well understood.

We investigated the effects of landscape features on the capabilities of the PulsePoint CPR dispatch platform in Allegheny County, Pennsylvania. We hypothesized that landscape features would reduce walkable area within dispatch zones and that larger alert radii could mitigate these effects.

CPR alert location data were obtained from the Allegheny County 911 PulsePoint deployment from July 2016-2019 (n = 1100). PulsePoint, a smartphone-based citizen CPR dispatch platform, alerts volunteers to public OHCAs within 400 m. Digital maps of alerts were generated for walkability analysis using the image Labeler MATLAB polygon tool. Unwalkable areas were labeled, classified into five categories, and quantified.

Of the 1100 events analyzed, encompassing 212 mi², 357 (32.45%) had no impediments to walkability. Within a 400 m radius of partially impeded events, the median proportion of obstructed area was 0.2250 (Min: 0.0005, Max: 0.8338, IQR: 0.3004, S_x: 0.1923). When the alert radius was expanded to 468 m, the median increased to 0.2336 (Min: 0.0016, Max: 0.8597, IQR: 0.3064, S_x: 0.1950). The percentages of total unwalkable area by each category were: terrain (54.286%), water (24.674%), road w/o crosswalk (11.3197%), railroad (7.4100%), private property (2.3102%).

In this region, most PulsePoint alerts had unwalkable areas, mostly from terrain and water. Contrary to our initial hypotheses, moderately increasing dispatch radius did not improve walkability.