Sensitivity and specificity of an automated external defibrillator algorithm designed for pediatric patients

Objective: Electrocardiographic (ECG) rhythm analysis algorithms for cardiac rhythm analysis in automated external defibrillators (AEDs) have been tested against pediatric patient rhythms (patients ≤8 years old) using adult ECG algorithm criteria. However these adult algorithms may fail to detect non-shockable pediatric tachycardias because they do not account for the difference in the rates of normal sinus rhythm and typical tachyarrhythmias in childhood. Methods: This study was designed to define shockable and non-shockable rhythm detection criteria specific to pediatric patients to create a pediatric rhythm database of annotated rhythms, to develop a pediatric-based AED rhythm analysis algorithm, and to test the algorithm's accuracy. Pediatric rhythm detection criteria were defined for coarse ventricular fibrillation, rapid ventricular tachycardia, and non-shockable rhythms, including pediatric supraventricular tachycardia. Pediatric rhythms were collected as sustained, classifiable, rhythms ≥9 s in length, and were annotated by pediatric electrophysiologists as clinically shockable or non-shockable based on pediatric criteria. Rhythms were placed into a pediatric rhythm database; each rhythm was converted to digitally accessible, public-domain, MIT rhythm data format. The database was used to evaluate a pediatric-based AED rhythm analysis algorithm. Results: Electrocardiographic rhythms from 198 children were recorded. There were 120 shockable rhythms from 49 patients (sensitivity; coarse ventricular fibrillation: 42 rhythms, 100%; rapid ventricular tachycardia: 78 rhythms, 94%), for combined sensitivity of 96.0% (115/120). There were 585 non-shockable rhythms from 155 patients (specificity normal sinus: 208 rhythms, 100%; asystole: 29 rhythms, 100%; supraventricular tachycardia: 161 rhythms, 99%; other arrhythmias: 187 rhythms, 100%), for combined specificity of 99.7% (583/585). Overall accuracy for shockable and non-shockable rhythms was 99.0% (702/709). Conclusions: New pediatric rhythm detection criteria were defined and analysis based on these criteria demonstrated both high sensitivity (coarse ventricular fibrillation, rapid ventricular tachycardia) and high specificity (non-shockable rhythms, including supraventricular tachycardia). A pediatric-based AED can detect shockable rhythms correctly, making it safe and exceptionally effective for children.