Western Journal of Emergency Medicine (Sep 2016)

A Comparison of Chest Compression Quality Delivered During On-Scene and Ground Transport Cardiopulmonary Resuscitation

  • Christopher S. Russi,
  • Logan J. Kolb,
  • Erik P. Hess,
  • Lucas A. Myers,
  • Christine M. Lohse,
  • Roger D. White

DOI
https://doi.org/10.5811/westjem.2016.6.29949
Journal volume & issue
Vol. 17, no. 5

Abstract

Read online

Introduction: American Heart Association (AHA) guidelines recommend cardiopulmonary resuscitation (CPR) chest compressions 1.5 to 2 inches (3.75-5 cm) deep at 100 to 120 per minute. Recent studies demonstrated that manual CPR by emergency medical services (EMS) personnel is substandard. We hypothesized that transport CPR quality is significantly worse than on-scene CPR quality. Methods: We analyzed adult patients receiving on-scene and transport chest compressions from nine EMS sites across Minnesota and Wisconsin from May 2008 to July 2010. Two periods were analyzed: before and after visual feedback. CPR data were collected and exported with the Zoll M series monitor and a sternally placed accelerometer measuring chest compression rate and depth. We compared compression data with 2010 AHA guidelines and Zoll RescueNet Code Review software. CPR depth and rate were “above (deep),” “in,” or “below (shallow)” the target range according to AHA guidelines. We paired on-scene and transport data for each patient; paired proportions were compared with the nonparametric Wilcoxon signed rank test. Results: In the pre-feedback period, we analyzed 105 of 140 paired cases (75.0%); in the post-feedback period, 35 of 140 paired cases (25.0%) were analyzed. The proportion of correct depths during on-scene compressions (median, 41.9%; interquartile range [IQR], 16.1-73.1) was higher compared to the paired transport proportion (median, 8.7%; IQR, 2.7-48.9). Proportions of on-scene median correct rates and transport median correct depths did not improve in the post-feedback period. Conclusion: Transport chest compressions are significantly worse than on-scene compressions. Implementation of visual real-time feedback did not affect performance.

Keywords