RESUMEN
BACKGROUND: Teaching diagnostic reasoning and giving feedback has an important role in medical education. Clinicians who teach may recognise errors, but be unfamiliar with the terminology used to describe them, leading to a lack of consistent and useful student feedback. OBJECTIVE: This prospective project evaluation study aimed to develop an examiner training package regarding errors in diagnostic reasoning, utilising consistent language and feedback tool, and report on diagnostic reasoning errors in second year medical students over the transition from preclinical to early clinical training at objective structured clinical exams (OSCEs). RESULTS: Likert questionnaire regarding examining, assessment and feedback pre- and post-training showed improvement in all measures, including examiner feedback confidence post training (p < .001). Students (n = 235) within the cohort were examined at the first preclinical OSCE 12 weeks into the teaching year and 236 students at the end of year OSCE. A range of 0-6 diagnostic reasoning errors were reported for individual students. When comparing mean history station scores at the preclinical OSCE for students who were observed to have diagnostic reasoning errors, students with 'poor pattern recognition' had a 4.2% lower score than those without this error type (p = .04, 95% CI of difference .14, 8.32), while those with 'unfocused data collection' error had a station score 7.7% lower than those without this error (p < .001, 95% CI of difference 3.50, 11.99). At the end of teaching year clinical OSCE, all common error types were associated with poorer performance. Error pattern shifted through the two longitudinal assessments, resulting in 'poor pattern recognition' having reduced and 'too narrow' and 'premature closure' increased rates. CONCLUSIONS: Incorporating the identification and feedback of common diagnostic reasoning errors into existing clinical assessments was feasible and easy to implement. Understanding, identifying and providing consistent feedback on common errors assists educators and could guide curriculum design.