Inaccurate treatment decisions of automated external defibrillators used by emergency medical services personnel: Incidence, cause and impact on outcome

BackgroundThe rhythm analysis algorithm (RAA) of automated external defibrillators (AEDs) may be deceived by many factors. In this observational study we assessed RAA accuracy in prehospital interventions. For every rhythm analysis judged to be inaccurate, we looked for causal factors and estimated the impact on outcome.MethodsIn 135 consecutive patients, two physicians reviewed 837 rhythm analyses independently. When they disagreed, a third physician made the final decision.ResultsAmong 148 shockable episodes, 23 (16%) were not recognized by the RAA due to external artifacts (n = 7), fine ventricular fibrillation (VF; n = 7), RAA error without external artifacts (n = 4) or a combination of factors (n = 5). In six cases the omitted/delayed shock was judged to be of clinical relevance: survival with some neurological deficit (n = 4), death without regaining consciousness (n = 1) and no restoration of spontaneous circulation (n = 1).In 689 non-shockable episodes, the RAA decided “shockable” 25 times (4%). This wrongful decision was due to external artifacts (n = 9), a concurrent shock of an internal cardioverter defibrillator (n = 1), RAA error without external artifacts (n = 13) or a combination of factors (n = 2). Fifteen spurious shocks were delivered. As these non-shockable rhythms did not deteriorate after the shock, we assumed that no significant harm was done.ConclusionsUp to 16% of shockable rhythms were not detected and 4% of non-shockable rhythms were interpreted as shockable. Therefore, all AED interventions should be reviewed. Feedback to caregivers may avoid future deleterious interactions with the AED, whereas AED manufacturers may use this information to improve RAA accuracy. This approach may improve the outcome of some VF patients.