There is a lack of high-quality information about the effectiveness of resuscitation interventions and international differences in structure, process and outcome after out-of-hospital cardiac arrest and cardiopulmonary resuscitation because data are not collected uniformly. An internet-based international registry could make such evaluations possible, and enable the conduct of large randomized controlled trials of resuscitation therapies. A prospective international cohort study was performed that included 571 infants, children and adults (a) who experienced cardiac arrest requiring chest compressions or external defibrillation, (b) outside the hospital in the study communities and (c) upon whom resuscitation was attempted by EMS personnel. Cardiac arrest was defined as lack of responsiveness, breathing or movement in individuals for whom the EMS system is activated for whom an arrest record is completed. All data were collated via a secure and confidential web-based method by using automated forms processing software with appropriate variable range checks, logic checks and skip rules. Median number of missing responses for each variable was 0 (interquartile range 0, 0). Twenty-seven percent of the patients had a first recorded rhythm of ventricular fibrillation or ventricular tachycardia, 60% had a witnessed arrest, and 34% received bystander CPR. Mean time from call to arrival on scene was 7.1+/-5.1 min. Six percent of the patients survived to hospital discharge. The resuscitation process was highly variable across centers, and survival and neurological outcome were also significantly and independently different across centers. This study shows that it is possible to collect data prospectively describing the structure, process and outcome associated with cardiac arrest in multiple international sites via the internet. Therefore, it is feasible to conduct adequately powered randomized trials of resuscitation therapies in international settings.