Effects of training in cardiopulmonary resuscitation on competence and patient outcome

Between 1981 and 1985 we carried out a study in two medium-sized nonteaching community hospitals to determine the rate of deterioration of knowledge and skills in cardiopulmonary resuscitation (CPR) among physicians and nurses, the accuracy of their perceptions of their knowledge and skills, the effects of practice on retention and the effect of CPR training on mortality. The participants' knowledge and skills were measured before training and immediately after, 6 months after and 12 months after training. Information on all attempts at CPR involving hospital staff was collected from medical records and from interviews with the participants. A total of 31 physicians and 54 nurses were followed during the study. Six months after training there was no difference in CPR knowledge or skills between the physicians and the nurses. In both groups CPR skills had deteriorated to near pretraining levels. By 6 months the physicians' knowledge had deteriorated to a level not significantly different from that before training. The nurses maintained a significant improvement in knowledge test scores at 12 months over those before training (p = 0.037). The physicians had an accurate perception of their knowledge but not their skills 6 months and 12 months after training, whereas the nurses did not accurately perceive either their knowledge or their skills after training. Experience with CPR did not contribute to post-training knowledge or skills in either group. There was no evidence that death rates were lower when basic life support (BLS) was begun by trained staff than when it was begun by untrained staff. The probability of survival was greater when BLS was begun within 4 minutes of arrest than when it was begun after 4 minutes, regardless of whether advanced cardiac life support was begun within 10 minutes.