Assessing Housestaff Diagnostic Skills Using a Cardiology Patient Simulator

OBJECTIVE: To assess the cardiovascular physical examination skills of internal medicine housestaff. DESIGN: Cross-sectional assessment of housestaff performance on three valvular abnormality simulations conducted on the cardiology patient simulator, "Harvey." Evaluations were done at the beginning (session I) and end (session II) of the academic year. SETTING: Duke University Medical Center internal medicine training program. SUBJECTS: Sixty-three (59%) of 107 eligible internal medicine housestaff (postgraduate years 1 through 3) agreed to participate and completed session I; 60 (95%) completed session II. MEASUREMENTS: All volunteers were tested on three preprogrammed simulations (mitral regurgitation, mitral stenosis, and aortic regurgitation). RESULTS: The overall correct response rates for all housestaff were 52% for mitral regurgitation, 37% for mitral stenosis, and 54% for aortic regurgitation. No difference was noted in correct response rates between sessions I and II. For mitral regurgitation, correct assessment of the contour of the holosystolic murmur predicted a correct diagnosis (P = 0.002). For mitral stenosis, identification of an opening snap and proper characterization of the mitral area diastolic murmur predicted a correct diagnosis (P < 0.0001). No individual observations were noted for the aortic regurgitation simulation, whose identification by the housestaff was associated with a correct diagnosis. CONCLUSIONS: Housestaff had difficulty establishing a correct diagnosis for simulations of three common valvular heart diseases. Accurate recognition of a few "key" observations was associated with a correct diagnosis in two of the three diseases. Teaching housestaff to elicit and interpret a few critical signs accurately may improve their physical diagnosis abilities.