Does Non-Targeted Community CPR Training Increase Bystander CPR Frequency?

Abstract

Objective: Only 37% of out-of-hospital cardiac arrests (OHCA) receive bystander Cardiopulmonary resuscitation (CPR) in Kent County, MI. In May 2014, prehospital providers offered one-time, point-of-contact compression-only CPR training to 2,253 passersby at 7 public locations in Grand Rapids, Michigan. To assess the impact of this intervention, we compared bystander CPR frequency and clinical outcomes in regions surrounding training sites before and after the intervention, adjusting for prehospital covariates. We aimed to assess the effect of this broad, non-targeted intervention on bystander CPR frequency, type of CPR utilized, and clinical outcomes. We also tested for differences in geospatial variation of bystander CPR and clinical outcomes clustered around training sites. Methods: Retrospective, observational, before-after study of adult, EMS-treated OHCA in Kent County from January 1, 2010 to December 31, 2015. We generated a 5-kilometer radius surrounding each training site to estimate any geospatial influence that training sites might have on bystander CPR frequency in nearby OHCA cases. Chi-squared, Fisher's exact, and t-tests assessed differences in subject features. Difference-in-differences analysis with generalized estimating equation (GEE) modeling assessed bystander CPR frequency, adjusting for training site, covariates (age, sex, witnessed, shockable rhythm, public location), and clustering around training sites. Similar modeling tested for changes in bystander CPR type, return of spontaneous circulation (ROSC), survival to hospital discharge, and cerebral performance category (CPC) of 1–2 at hospital discharge. Results: We included 899 cases before and 587 cases post-intervention. Overall, we observed no increase in the frequency of bystander CPR or favorable clinical outcomes. We did observe an increase in compression-only CPR, but this was paradoxically restricted to OHCA cases falling outside radii around training sites. In adjusted modeling, the bystander CPR training intervention was not associated with bystander CPR frequency (β −0.002; 95% CI −0.16, 0.15), compression-only CPR (β −0.06; 95% CI −0.15, 0.02), ROSC (β −0.06; 95% CI −0.21, 0.25), survival (β −0.02; 95% CI −0.11, 0.06), or favorable neurologic outcome (β −0.01; 95% CI −0.07, 0.09). Conclusions: We observed no impact in bystander CPR performance or outcomes from a blanket, non-targeted approach to community CPR education. The effect of targeted CPR education in locales with known low bystander CPR rates should be tested in this region.