External artifacts by advanced life support providers misleading automated external defibrillators

Abstract

The ECG analysis algorithm of automated external defibrillators (AEDs) shows reduced sensitivity and specificity in the presence of external artifacts. Therefore, ECG analyses are preceded by voice prompts. We investigated if advanced life support (ALS) providers follow these prompts, and the consequences if they do not. In a two-tiered EMS system all 510 ECG analyses from 135 resuscitation attempts with a Laerdal FR2 AED (applied by emergency medical technicians [EMTs] and subsequently used by ALS providers) were prospectively evaluated. The ALS data were compared with data before arrival of ALS providers (EMT data) using Mc Nemar test. In the presence of ALS providers, 286 ECG rhythm analyses were performed. In the 96 analyses with shockable rhythms, artifacts were detected in 35 (36%), leading to a wrongful no shock decision in 19 (20%). Corresponding EMT data were 67 analyses with shockable rhythms, with artifacts in 18 (27%; p<0.001) but a wrongful no shock decision in only 3 (4%; p<0.001). ALS providers also failed to deliver the AED shock in 7 of the 77 analyses with an appropriate shock decision (9%). This was never found in the EMT data. In the 190 analyses of a non-shockable rhythm in the presence of ALS providers, artifacts were detected in 120 (63%) leading to one spurious shock (0.5%). Corresponding EMT data were 157 analyses, with artifacts in 87 (55%; p=0.20) but no spurious shocks. External artifacts were frequently found, sometimes leading to important errors. Consequently, more training is needed, especially for ALS providers.