SUMMARY To study the effect of coronary care on the hospital mortality rate of acute myocardial infarction, patients treated in a coronary care unit (CCU) were compared with patients treated in general wards. To assess the comparability of cases, the coronary prognostic index of Chapman and Gray (1973) was modified. The modified index was based on serum aspartic aminotransferase (AST) level and cardiogenic shock. Modification was necessary because oliguria, the third variable on which the original index was based, could not always be determined for general ward cases. This modified index assessed very accurately the prognosis of myocardial infarct patients treated in hospital. Patients treated in the coronary care unit had a significantly higher mean modified prognostic index than control patients treated in the general wards, either before or after the introduction of coronary care. They also had a significantly shorter mean delay in admission than either group of general ward cases. These differences would be expected to worsen, not improve, the results of coronary care. By contrast, the two groups of general ward patients were very similar, as shown by mean modified prognostic index, mean admission delay, and results of correlation and regression analysis. In spite of their higher mean modified prognostic index and shorter mean admission delay, the coronary care unit patients had a significant reduction in their mortality rate, which dated from the opening of the unit, to little more than half that for the general ward patients. For general ward patients treated after the introduction of coronary care, on the other hand, the mortality rate was not significantly different from that for ones treated before. Furthermore, the reduction in mortality for coronary care unit cases was approximately the same over the whole range of infarct severity, and did not vary significantly from 15 per cent as the modified index rose from minimum to maximum values. It is suggested that severity of infarction should no longer be a reason for excluding any patients from coronary care.Controversy continues about the effectiveness of coronary care (Donaldson, 1978;Hill et al., 1978). There is little published evidence that the mortality rate of acute myocardial infarction is lower for patients treated in a coronary care unit than for those treated elsewhere, in groups of patients shown otherwise to have a similar probability of dying (Lindholm et al., 1976;Peterson, 1978). This is in spite of the fact that the earliest coronary care units were established over a decade and a half ago (Brown et al., 1963;Day, 1963;Wilburne and Fields, 1963), and notwithstanding numerous reports on the subject. Furthermore, while it seems that most cardiologists believe that coronary care units do improve the overall prognosis of acute myocardial infarction, it is also commonly believed that any reduction in mortality is greatest in patients with mild or only moderately severe