Performance of VF detection parameters in an algorithm design scenario and in a real resuscitation scenario

Detection of Ventricular fibrillation (VF) in automated external defibrillators (AED) is tested following the recommendations of the American Heart Association (AHA). However, nonshockable out-of-hospital cardiac arrest (OHCA) rhythms may be very different from those covered in the AHA recommendations. In this study we compare the performance of four VF detection parameters for a testing database of rhythms covered by the AHA recommendations and a database of OHCA rhythms. Spectral parameters performed better than parameters related to the heart rate or the complexity for the testing database but worse for the OHCA database. The performance of the parameters was very different in an algorithm design scenario compliant with the AHA statement and in a real resuscitation scenario.