Sudden cardiac arrest represents a critical medical emergency where the quality of resuscitation measures is crucial for favorable outcomes[1,2]. The quality of cardiopulmonary resuscitation (CPR) can be assessed through the analysis of data recorded by the defibrillator during CPR. The data is routinely recorded at the base of the emergency medical vehicles at the University Hospital of Graz.
 To assess CPR quality, algorithms were implemented to identify chest compressions based on the chest impedance and ventilations using the capnography signals. With these algorithms, parameters for CPR quality can be determined. A dataset with 522 cases was used, of which 97 cases were automatically discarded as they contained no or no relevant impedance signals.
 The mean chest compression rate of the cases is 116.3 ± 9.1 compressions per minute, which is within the target range of 90 – 120. However, in 31% of cases, the rate was found to be too high. The proportion of minutes without an adequate mean chest compression rate was 29.2 ([6.2; 60.4]) %.
 The mean ventilation frequency of all cases with capnography signal (150 cases) was 13.5 ± 3.6 ventilations per minute and, therefore, within the target range of a maximum of 15 ventilations. However, in 34% of cases, the ventilation frequency was too high.
 This work shows that a systematic and automated evaluation of defibrillator recordings is suitable for assessing the quality of resuscitation measures. The determined parameters produced plausible results and are consistent with the results of preliminary studies at the Medical University of Graz, which were carried out manually on significantly smaller datasets.