Article Text

Download PDFPDF
Insufficient quality of public automated external defibrillator recordings in the greater Paris area, a descriptive study
  1. Clément Derkenne1,
  2. Daniel Jost1,
  3. Pierre Alexandre Haruel1,
  4. Romain Kedzierewicz1,
  5. Benoit Frattini1,
  6. Lemoine Frédéric1,
  7. Pascal Diegelmann1,
  8. Jeremy Gouze1,
  9. Florian Roquet1,
  10. Stéphane Travers1,
  11. Paris Fire Brigade Cardiac Arrest Task Force1,
  12. Rene Bihannic1,
  13. Bertrand Prunet1,2
  1. 1 Emergency Medical Department, Paris Fire Brigade, Clamart, France
  2. 2 French Military Health Service, Val de Grâce Military Academy, Paris, France
  1. Correspondence to Dr Clément Derkenne, Emergency Medical Department, Paris Fire Brigade, Clamart, France; clement.derkenne{at}


Background Western countries report a significant increase in the proportion of patients who experience out-of-hospital cardiac arrests (OHCAs) and benefit from a public automated external defibrillator (pAED) before the arrival of rescue teams. However, recordings of devices recovered after resuscitation are of variable quality. Analysis of these data may inform decisions of whether to implement an internal defibrillator for survivors, and provide useful information about the performance of pAED algorithms and the actions of bystanders.

Objective To investigate the quality of the information recorded by pAEDs during OHCAs in the Paris area.

Methods pAED files used for some of the 8629 OHCAs that occurred in the greater Paris area between 1 January 2017 and 31 April 2019 on the day of the arrest were collected. The presence and accuracy of 23 factors required to interpret the recording was noted, including readability of the ECG, the presence of an impedance curve and the accuracy of the date and time. The recordings were analysed to assess the diagnostic and therapeutic performance of the pAEDs used.

Results A total of 258 patients with an OHCA received assistance from a pAED, and 182 recordings were recovered. The pAEDs were made by 12 different manufacturers. Data extraction required eight different transmission modes and 16 software programmes; recordings were of highly heterogeneous quality. Two per cent of the recordings were of such poor quality that they were not interpretable. Among the 98% remaining, only 43% included a thoracic impedance curve, 34% the intensity of the shocks delivered and 8% the patient name. The date and time were accurate in 68% and 48% of recordings, respectively. The pAEDs had 87.6% (95% CI 83.7% to 91.0%) sensitivity and 99.5% (99.5% to 99.5%) specificity for defibrillating shockable rhythms (positive predictive value 98.2% (96.4% to 99.0%), negative predictive value 96.4% (95.3% to 96.8%)). The absence of important variables prevented the analysis of approximately half of the inappropriate decisions made by pAEDs.

Conclusion Collection of pAED recordings is a major challenge. Their analysis is compromised by heterogeneity and poor quality (incomplete maintenance records, patient details and logs). AED recordings are currently the most relevant resource to track pAED performance and bystander practices. The quality of these recordings needs to improve.

  • cardiac arrest
  • ECG
  • interpretation
  • equipment evaluation
  • prehospital care
  • first responders

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Handling editor Richard Body

  • CD and DJ contributed equally.

  • Contributors CD, DJ, ST and PAH conceived the study and designed the trial. CD, PD, JG, LF, BF and RB supervised data collection. CD and FR provided statistical advice on the study design and analysed the data. CD, DJ, RK and BP drafted the manuscript, and all authors contributed substantially to its revision.

  • Funding This study was supported by institutional funding from the Emergency Medical Department of the Paris Fire Brigade, Paris, France.

  • Competing interests None declared.

  • Patient consent for publication Not required.

  • Ethics approval The study was approved by the ethics committee of the French Anaesthesia and Intensive Care Society (IRB 00010254-2017-029).

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data availability statement Data are available in a public, open access repository. Data are available upon reasonable request. all recordings, with the exception of nominative recordings, will be available on request by email at