Survival rates for out-of-hospital cardiac arrest vary widely among locations. We surveyed the definitions used in published studies of out-of-hospital cardiac arrest. Data from 74 studies involving 36 communities showed survival rates ranging from 2% to 44%. There were five different case definitions and 11 different definitions of survivors. The absence of uniform definitions prevents meaningful intersystem comparisons, prohibits explorations of hypotheses about effective interventions, and interferes with the efforts of quality assurance. The most satisfactory numerator for a survival rate appears to be survival to hospital discharge; the most appropriate denominator appears to be witnessed adult cardiac arrest of presumed heart disease etiology, with ventricular fibrillation as the initial identified rhythm. Proposed definitions for the data emergency medical services systems should report as they examine their cardiac arrest survival rates are presented.