Dr Coskey's letter raises two important points. First, we agree that the effectiveness of CPR will and should improve from year to year in units that have a stable staffing and a good training program. This will cause some decrease in mortality, as is illustrated in comparing our groups A and B. As discussed in the article, a fall in mortality occurred but was not decisive until monitoring facilities were added, even though the training standards were higher.
The more important question is whether the high mortality in the first two days of admission in group C actually caused the lower mortality of these patients in days 3 to 14 when they entered the study of intermediate coronary care. There were 159 deaths in the series, 50 (7.6%; incorrectly shown as 6.6% in the article) in group A, 57 (8.1%) in group B, and 52 (7.1%) in group C. The