AIM: To describe possible factors modifying the effect of bystander cardiopulmonary resuscitation on survival among patients suffering an out-of-hospital cardiac arrest. PATIENTS: A national survey in Sweden among patients suffering out-of-hospital cardiac arrest and in whom resuscitative efforts were attempted. Sixty per cent of ambulance organizations were included. DESIGN: Prospective evaluation. Survival was defined as survival 1 month after cardiac arrest. RESULTS: In all, 14065 reports were included in the evaluation. Of these, resuscitation efforts were attempted in 10966 cases, of which 1089 were witnessed by ambulance crews. The report deals with the remaining 9877 patients, of whom bystander cardiopulmonary resuscitation was attempted in 36%. Survival to 1 month was 8.2% among patients who received bystander cardiopulmonary resuscitation vs 2.5% among patients who did not receive it (odds ratio 3.5, 95% confidence interval 2.9-4.3). The effect of bystander cardiopulmonary resuscitation on survival was related to: (1) the interval between collapse and the start of bystander cardiopulmonary resuscitation (effect more marked in patients who experienced a short delay); (2) the quality of bystander cardiopulmonary resuscitation (effect more marked if both chest compressions and ventilation were performed than if either of them was performed alone); (3) the category of bystander (effect more marked if bystander cardiopulmonary resuscitation was performed by a non-layperson); (4) interval between collapse and arrival of the ambulance (effect more marked if this interval was prolonged); (5) age (effect more marked in bystander cardiopulmonary resuscitation among the elderly); and (6) the location of the arrest (effect more marked if the arrest took place outside the home). CONCLUSION: The effect of bystander cardiopulmonary resuscitation on survival after an out-of-hospital cardiac arrest can be modified by various factors. Factors that were associated with the effect of bystander cardiopulmonary resuscitation were the interval between the collapse and the start of bystander cardiopulmonary resuscitation, the quality of bystander cardiopulmonary resuscitation, whether or not the bystander was a layperson, the interval between collapse and the arrival of the ambulance, age and the place of arrest.