Effect of exercise treadmill testing and stress imaging on the triage of patients with chest pain: CHEER substudy

OBJECTIVE: To determine whether stress imaging for patients who are unsuitable for exercise treadmill testing (ETT) as part of a chest pain unit (CPU) triage strategy resulted in incremental benefit in clinical outcomes and relative costs compared with patients randomized to routine hospital admission. PATIENTS AND METHODS: Clinical outcomes and medical resource utilization were examined at the Mayo Clinic in Rochester, Minn, for 212 intermediate-risk patients with unstable angina randomized to a CPU and compared with 212 patients randomized to routine admission from November 21, 1995, to March 18, 1997. Patients in stable condition in the CPU underwent ETT; if patients were unsuitable for ETT, stress imaging was performed. Costs for CPU evaluation and outcomes were assessed during a 6-month follow-up. RESULTS: During the observation period, 60 patients (33%) were admitted to the hospital. Of the 152 remaining patients, 125 (82%) underwent ETT (91 had normal results), and 27 (18%) underwent stress imaging (3 had normal results). Patients with normal ETT or stress imaging results had no primary events at 6-month follow-up. Patients admitted to the hospital who underwent stress imaging had an insignificantly higher 6-month event rate compared with patients who underwent ETT (16.7% vs 3.1%; P=.38). The standardized resource-based relative-value units (RBRVUs) for patients who underwent ETT and stress imaging during follow-up were 19.4 and 56.4 RBRVUs, respectively, compared with 51.4 (ETT) and 52.1 (stress imaging) RBRVUs for similar numbers of patients randomized to routine admission. CONCLUSIONS: Exercise treadmill testing safely stratified most intermediate-risk patients with unstable angina and was less costly than routine admission. Patients not suitable for ETT are likely to have abnormal stress imaging results. They represent a higher-risk cohort that could be routinely admitted to the hospital without reducing the effectiveness of the CPU strategy.