Quality of bystander cardiopulmonary resuscitation (CPR) affect patient survival. Quality of professional CPR on patients has not been studied in detail, but it is regularly reported that the quality when tested on manikins deteriorates dramatically within months after training. Automated direct feedback on CPR quality from manikins brings quality back within a couple of minutes. Similar feedback has been incorporated into a defibrillator which also monitors quality of CPR. We hypothesise that quality of professional clinical CPR improves with such feedback

Defibrillators which monitor quality of CPR via changes in thoracic impedance (for ventilation) and movement of the sternum employing an accelerometer (for chest compressions) will be employed in ambulances in Akershus county (Norway), Stockholm (Sweden) and London (UK). During phase 1 quality of CPR will be monitored without feedback from the defibrillator. During phase 2 the ambulance personnel will receive feedback via the defibrillator. During phase 3 the ambulance personnel will be retrained with particular attention to the quality problems that became apparent in phase 3.

Quality of CPR will be continuously recorded by the defibrillators and the data collected and sent via internet to Laerdal Medical. All other cardiac arrest data including survival will be recorded using standard datasets for cardiac arrest research as developed by a task force with members from the organisations in International Liaison Committee on resuscitation (Utstein guidelines). The data will be annotated and analyzed in detail by researchers at University of Oslo

Inclusion Criteria:

• Cardiac arrest out-of-hospital

Exclusion Criteria:

• < 18 years old