Physician-level variation in the diagnosis of myocardial infarction and the use of angiography among Veterans with elevated troponin

David E. Winchester, Nayan Agarwal, Lucas Burke, Steven Bradley, Tatiana Schember, Carsten Schmalfuss

Cardiac troponin assays have improved the ability to detect myocardial damage. However, ascertaining whether troponin elevation is due to myocardial infarction (MI) or secondary to another process can be challenging. Our aim is to evaluate provider-level variation in the diagnosis of MI and the use of invasive coronary angiography (ICA) among patients with undifferentiated elevations in cardiac troponin. We analyzed data from all patients with elevated troponin levels in a single Veterans Affairs (VA) Medical Center between 2006 and 2007. One of several cardiologists prospectively evaluated each patient's presentation and course of care. We compared the frequency of MI diagnosis and ICA use between physicians using univariate odds ratios (OR). Among 761 patients, 34.0 % were diagnosed with MI and 25.9 % underwent ICA. The unadjusted rates of MI (23.9 to 56.7 %, P = 0.02) and ICA (17.3 to 73.3 %, P < 0.001) differed between physicians. Comparing the patient cohorts for each physician, baseline characteristics were similar except for chest pain. In multivariate regression, factors associated with the use of cardiac ICA included an abnormal electrocardiograph (ECG) (OR = 1.89, P = 0.014), level of troponin (OR = 1.71, P = 0.004), chest pain (OR = 8.60, P < 0.001), and care by non-VA physicians (OR = 4.45, P = 0.006). One physician had a lower ICA use (OR = 0.56, P = 0.017). In multivariate regression of MI, no physician-level variation was observed. Among patients with elevated troponin, the likelihood of being diagnosed with MI and undergoing ICA is dependent on their clinical presentation. After adjustment, physician-level variation in care was observed for the use of ICA, but not for the diagnosis of MI.