American Journal of Respiratory and Critical Care Medicine, 2023/05, 207(1):. doi: 10.1164/ajrccm-conference.2023.B102, type of study: ,css,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

Full text    (https://www.atsjournals.org/doi/abs/10.1164/ajrccm-conference.2023.207.1_MeetingAbstracts.A4298)

Keywords:

clinical reasoning [85]
congress [77]
cross sectional study [894]
diagnosis [412]
medical students [693]
osteopathic medicine [2108]

Abstract:

Rationale: Development of clinical reasoning requires deliberate practice: continuous structured practice with feedback (Abdulnour, NEJM, 2022). Workplace-based deliberate practice is the cornerstone of medical education. Still, it is resource intensive, and students may have gaps in exposure to the corpus of illnesses considered essential to training. NEJM Healer is a virtual patient simulator based on illness script theory. Learners explicitly apply clinical reasoning skills as they advance through the encounter and are provided feedback. This study examined whether NEJM Healer performance correlated with learner training level and medical knowledge. Methods: NEJM Healer cases were assigned to 632 pre-clerkship and 481 post-clerkship medical students at Lake Erie College of Osteopathic Medicine (LECOM). During an encounter, students identify relevant clinical data, write problem representations, and develop a differential diagnosis at four data stages (triage, history, physical examination, and pre-existing diagnostic tests). Students also build illness scripts for their diagnostic hypotheses, identify their diagnostic confidence level, and define a management plan. Student diagnostic accuracy-percent of expert hypotheses captured by the student-was compared to a consensus differential diagnosis developed by a small group of experienced clinicians, which included generalists and subspecialists in their respective content areas. Diagnostic accuracy was compared to national standardized medical knowledge test scores that occurred shortly after clerkships. Comparisons between groups were performed using t-tests, Wilcoxon-rank sum, or Pearson's chi-square tests, depending on the distribution of the variable. Results: Each student completed five or more virtual encounters. On average, pre-clerkship students spent more time per NEJM Healer encounter than post-clerkship students (40 vs. 31 minutes, p < 0.0001). Post-clerkship students had higher diagnostic accuracy throughout the case (50% vs. 44%, p < 0.0001) and were more likely to identify the lead diagnosis (72% vs. 65%, p <0.0001). In a representative encounter, stage-specific diagnostic accuracy improved for all students as the encounter progressed and was higher for post-clerkship students at each stage (Figure 1). Post-clerkship students were more confident in their diagnosis (54% vs. 27%, p < 0.001). Diagnostic accuracy had a weak but significant correlation with standardized scores on a knowledge test (R2 = 0.05, p < 0.0001). Conclusions: Performance in NEJM Healer correlated with students' clinical experience level and allowed identification of students who failed to improve in specific knowledge areas over a clerkship. NEJM Healer may be helpful for educators to identify students requiring additional deliberate practice. (Figure Presented).