OBJECTIVE: To determine whether adverse events occurring after coronary artery bypass surgery in Medicare patients can be predicted from clinical variables representing illness severity at admission.

DESIGN: Retrospective analysis of clinical data abstracted from hospital charts, with development and validation using half-samples of the database. A logistic model was developed using illness severity at admission to predict the occurrence of an adverse event after bypass surgery.

SETTING: Hospitals in seven states.

PATIENTS: Random sample of 2213 Medicare patients 65 years of age or more who underwent bypass surgery between January 1985 and June 1986.

OUTCOME MEASURE: The occurrence of death within 30 days of admission or any of 13 nonfatal postoperative adverse events (for example, myocardial infarction, congestive heart failure, and wound infection).

RESULTS: Thirty-three percent of patients had one or more postoperative adverse events or died within 30 days of admission. Mortality within 30 days of admission was 6.6%; each adverse event was associated with increased mortality (range, 7.5% to 66.7%). Admission predictors of the occurrence of an adverse event included a history of bypass surgery, emergent surgery, a history of chronic obstructive pulmonary disease, the presence of an infiltrate on admission chest radiograph, a pulse of 110 beats/min or more, age, blood urea nitrogen of 10.7 mmol/L (30 mg/dL) or more, acute myocardial infarction at admission, and a history of myocardial infarction; the presence of one- or two-vessel disease was negatively associated with the occurrence of an adverse event. The model c-statistic was 0.64.

CONCLUSIONS: Severity of illness at admission has modest predictive power with respect to adverse-event occurrence in Medicare patients who undergo bypass surgery.