Virtual patients to explore and develop clinical case summary statement skills amongst Japanese resident physicians: a mixed methods study

Abstract

Background: In Western clinical training, formulation of a summary statement (SS) is a core exercise for articulation, evaluation, and improvement of clinical reasoning (CR). In Japanese clinical training, structured guidance in developing CR, including opportunity for SS practice, is uncommon, and the present status of case summarization skills is unclear. We used Virtual Patients (VPs) to explore Japanese junior residents' SS styles and the effectiveness of VPs on improving SS quality.

Methods: All first-year junior resident physicians at 4 residency programs (n = 54) were assigned randomized sequences of 5 VP modules, rolled out at 6 day intervals. During each module, participants free-texted a case summary and then reviewed a model summary. Thematic analysis was used to identify SS styles and each SS was categorized accordingly. Frequency of SS styles, and SS CR quality determined by 1) an internally developed Key Feature rubric and 2) demonstration of semantic qualification, were compared across modules.

Results: Four SS styles were identified: numbered features matched to differential diagnoses, differential diagnoses with supportive comments, feature listing, and narrative summarization. From module #1 to #5, significant increases in the narrative summarization SS style (p = 0.016), SS CR quality score (p = 0.021) and percentage of semantically driven SS (p = 0.003) were observed.

Conclusions: Our study of Japanese junior residents identified distinct clinical case summary statement styles, and observed adoption of the narrative summarization style and improvement in the CR quality of summary statements during a series of VP cases.

Fig. 1

Change in percentage of summary assessment style over course of 5 VPs modules

Fig. 2

Clinical reasoning quality of summary assessments completed in VP modules. Legend: Scores determined by Key Feature rubric. *Compared to Module 1, improvement in performance achieved statistical significance on Module 4 (p = 0.040) and Module 5 (p = 0.021)