Incorrectly performed bystander CPR might compromise survival of the cardiac arrest patient. We therefore evaluated the outcome in 3306 out-of-hospital primary cardiac arrests of which 885 received bystander CPR. bystanders performed CPR correctly in 52%, incorrectly in 11%, 31% performed only external chest compressions (ECC) and 6% only mouth-to-mouth ventilation (MMV). The initial ECG in cases without bystander CPR was ventricular fibrillation in 28% (95% confidence interval: 27-30%); 45% (41-50%) and 39% (29-48%), respectively when bystander CPR was performed correctly or incorrectly; 43% (37-49%) when only ECC was applied and 22% (11-33%) when only MMV was practiced. Long term survival, defined as being awake 14 days after CPR, was 16% (13-19%) in patients with correct bystander CPR; 10% (7-14%) and 2% (0-9%), respectively when only ECC or only MMV was performed; 7% (6-8%) when no bystander was involved; 4% (0-8%) when bystander CPR was performed incorrectly. Bystander CPR might have a beneficial effect on survival by maintaining the heart in ventricular fibrillation by ECC. A negative effect of badly performed bystander CPR was not observed compared to cases which had not received bystander CPR.