Clinical predictors easily obtained at presentation predict resource utilization in unstable angina

Objective To determine if a risk prediction model for patients with unstable angina would predict resource utilization. Methods and Results Four hundred sixty-five consecutive patients admitted far unstable angina to a tertiary core university-based medical center were prospectively evaluated from June 1, 1992, to June 30, 1995. The proportion of patients receiving coronary angiography, coronary angioplasty, and coronary artery bypass grafting were analyzed according to four risk groups on the basis of a previously published model: Group 1, <2% risk of major complication; Group 2, 2.1% to 5% risk; Group 3, 5.1% to 15% risk; and Group 4, >15.1% risk. Hospital length of stay and estimated cost of hospitalization based on DRG and specific payer ratio of cost-to-charge were also compared between groups. Multiple linear regression analysis was used to determine the influence of estimated risk and procedures on hospital costs. The four groups were well matched for gender, hypertension, tobacco history, and previous percutaneous transluminal coronary angioplasty and myocardial infarction. Group 4 had a higher incidence of previous coronary bypass grafting (35% vs 10%, p = 0.001) and triple vessel or left main coronary artery disease compared with Group 1 (44% vs 13%, p =0.041). Group 4 patients were more likely to be admitted to the coronary care unit compared with Group 2 or Group 1 patients (80% vs Group 1:51% [p = 0.001]; and vs Group 2: 53% [p = 0.001]), more likely to receive heparin (87% vs 71%, p = 0.007), and more likely to receive a beta-blocker or calcium channel blocker (89% vs 74%, p = 0.008) than Group 1. Coronary angioplasty rates were similar for ail groups, but Group 4 patients were more likely to receive coronary bypass grafting than Group 2 or Group 1 (27% vs Group 2: 12%, p = 0.004 and vs Group 1: 8%, p = 0.002). Hospital length of stay was highest in Group 4 and lowest for Group 1. Average hospital costs were significantly less in Group 3 than in Group 4, but higher than in Group 1. Multivariate analysis determined a dependency of costs on risk group with Group 2 having costs 31.4% (95% Cl = 9.8 to 57.2), Group 3 46.7% (24, 3 to 73.1), and Group 4 75% (46.9 to 110.7) higher than Group 1. The use of procedures also significantly increased costs, with PTCA-treated patients having a 44.9% (26.7 to 65.7) increase in costs compared with medically treated patients, and surgically treated patients having a 204.7% increase in costs. Conclusion Resource utilization as assessed by the use of revascularization procedures, length of stay, and hospital costs are influenced by patient acuity estimated from a prediction model on the basis of estimated risk of cardiac complications. The model exerts independent influence on cost even after adjustment for various procedures. The use of revascularization procedures, especially coronary artery surgery, remains a large determinant of hospital cost.