Interpretation of immediate exercise treadmill test: Interreader reliability between cardiologist and noncardiologist in a chest pain evaluation unit

Study objective: To determine whether attending physicians in a chest pain evaluation unit (CPEU) can perform and interpret exercise testing with the same accuracy as cardiologists. Methods: Between January 1996 and November 1998, immediate exercise tests w ere performed and interpreted by internists with additional training in exe rcise testing who serve as attending physicians in a CPEU at a large univer sity medical center. For quality assurance, all tests were overread by a ca rdiologist. Test results were compared for each reader, and ail tests with discrepant readings were reinterpreted by an independent cardiologist who w as blinded to the previous results. Patients' clinical course was monitored for at least 30 days after exercise testing. Results: The study group consisted of 645 patients (347 men, 298 women). Di screpant interpretations were found in 11 (1.7%) patients. The agreement wa s 98.4% (kappa value 0.9618). The majority of discrepancies were insignific ant and were based on subtle differences in the definition of a nondiagnost ic test or the degree of ST-segment shift. Of the 11 discordant readings, t he blinded cardiologist concurred with 5 (45%) of the CPEU interpretations and 4 (36%) of the cardiologist interpretations. in 2 cases, there was disa greement by all 3 interpreters. There was no cardiac morbidity or mortality of any patient with a discrepant reading. Conclusion: Our results suggest that noncardiologists serving as attending physicians in a CPEU can accurately interpret exercise tests and overreadin g by cardiologists far quality assurance is unnecessary.