Abstract
BACKGROUND: Making an accurate diagnosis is a core skill residents must develop. Assessments of this skill and decisions to grant residents clinical independence often are based on global impressions. A workplace-based assessment of diagnostic accuracy could be a useful part of a competency-based assessment program and could inform decisions about granting residents independence. INNOVATION: We developed a method for measuring diagnostic accuracy that was integrated into the workflow of internal medicine residents and attending physicians. METHODS: Four senior medical residents and 6 attending physicians working in the internal medicine clinical teaching unit of a tertiary hospital participated in this study. To determine their diagnostic accuracy, residents documented a leading diagnosis for each patient they evaluated in the emergency department. After reviewing each case with the resident and after examining the patient, the resident's attending physician documented the diagnosis. Discharge diagnosis was determined by retrospective chart review to allow determination of resident and attending physician diagnostic accuracy. Data were collected for 240 consecutive patients referred for a medicine consultation. RESULTS: Resident diagnostic accuracy was 66% (95% CI 60-72), whereas attending physician accuracy was significantly higher at 79% (95% CI 74-84, P < .001). By logistic regression, the accuracy of the attending physician was found to be influenced by the accuracy of the resident. Participants felt this process motivated them to improve their clinical reasoning. CONCLUSIONS: Measuring resident diagnostic accuracy provides information that could be used in a competency-based assessment program to provide feedback and motivation to stimulate performance improvement.