Evaluation of cardiopulmonary resuscitation skills of general practitioners using different scoring methods

In this study we evaluated the practical performance of 70 general practitioners in cardiopulmonary resuscitation (CPR) before and after instruction and compared checklist-based scores to mechanical recording scores in order to investigate which scoring method is preferable. Both checklist and recording strip-based scores showed significant improvement after instruction, but only 37% were judged proficient according to the American Heart Association standards (checklist scoring), and 47% according to the recording print-based scoring system, while raters judged 97% as satisfactory by general impression. Interrater reliability was highest for the recording print (0.97) and lower for the checklist (0.79), especially for CPR performance (0.56). Comparison of checklist and recording print showed that the checklist was specific but not very sensitive in identifying poor performance for cardiac compression rate, since observers overestimated performance. The correlation for CPR performance between checklist score and recording strip score was low (0.45), indicating that candidates were ranked differently. The correlation between diagnosis and performance score was low for checklist as well as recording print (0.22), indicating that the score on diagnosis was a poor predictor for the score on performance of CPR. These results support the use of the recording manikin as compared with the use of a checklist for formative evaluation of basic life support skills. However, as proficiency in diagnosis and performance in CPR are poorly correlated, assessment of diagnosis using a checklist must be included. Therefore we strongly recommend the combination of assessment by observers using a checklist for diagnostic procedures and the recording strip of the manikin for performance of CPR, as employed in most evaluation schemes. (C) 1997 Elsevier Science Ireland Ltd.