Audit of a rural hospital's coronary care unit: comparison of two predictive instruments of acute myocardial infarction mortality.

The purpose of the study was to compare observed mortality of a rural hospital coronary care unit with mortality rates estimated by two predictive instruments of mortality. The mortality rates of 86 consecutive patients with confirmed acute myocardial infarction were compared with those predicted by the presence or absence of eight risk factors for mortality identified by the Thrombolysis in Myocardial Infarction (TIMI) trial, and mortality predicted by a logistic regression equation LRE). Seventeen patients (20 per cent) died within 6 weeks of admission; the number of TIMI risk factors present predicted a mortality of 9.8 per cent, and the instrument of Selker's predicted a mortality of 25.9 per cent. Patients with 3 TIMI risk factors had a significantly higher mortality than predicted (46.2 versus 13.0 per cent, p < 0.01). There were no significant differences between the receiver operating characteristic (ROC) curve of either instrument. The predictions of Selker's instrument, however, showed no significant difference from observed mortality, even when the patients were grouped into quintiles, and the predicted mortality rates were corrected for any presumed benefit from thrombolysis. The predictive instrument of Selker more consistently estimates observed mortality than the presence of risk factors identified by the TIMI trial.