Blended versus Traditional Face-to-Face Learning in Medical Interview Training for Students: A Pilot Randomized Controlled Trial

Technical Summary: Blended versus Traditional Face-to-Face Learning in Medical Interview Training

Problem Statement
The COVID-19 pandemic necessitated a rapid shift toward digitally supported educational approaches in medical education, including medical interview training. While blended learning (combining online and face-to-face instruction) has shown benefits in general medical education, its specific value in medical interview training remains unclear. Medical interviews require not only clinical knowledge but also the development of communication skills, rapport building, and real-time interaction. The authors hypothesized that blended learning could serve as a practical educational approach for this specific training modality in the post-pandemic era, yet no prior studies had directly compared blended learning with traditional face-to-face learning in this context.

Methodology
The study was an open-label, parallel-group pilot randomized controlled trial (RCT) conducted at Dokkyo Medical University in Japan between May 2023 and December 2024.

• Participants: Fifth- and sixth-year medical students rotating in the General Internal Medicine department. A total of 26 students were randomized (13 to blended learning, 13 to traditional learning) after screening 197 candidates. One participant withdrew before the intervention, leaving 25 participants who underwent the assigned learning program.
• Intervention: All participants completed three mock medical interviews with simulated patients presenting with common symptoms (chest pain, cough, fatigue).
  • Blended Learning Group: The first two interviews were conducted online via Microsoft Teams; the third was face-to-face.
  • Traditional Learning Group: All three interviews were conducted face-to-face.
  • Protocol: Interviews involved history taking and verbal description of physical examinations within 10 minutes, followed by 15 minutes for patient note documentation. Immediate feedback was provided by an attending physician.
• Outcomes:
  • Primary Outcome: Total score of a 24-item questionnaire derived from two subscales of the Dundee Ready Educational Environment Measure (DREEM): "Perceptions of Learning" and "Perceptions of Atmosphere."
  • Secondary Outcomes: Scores for the individual DREEM subscales and post-intervention self-assessment scores for six clinical competencies (history taking, physical examination, patient care/communication, clinical reasoning, management planning, and patient note documentation).
• Analysis: Analyses utilized a modified intention-to-treat approach. Continuous variables were compared using Student's t-test or analysis of covariance (ANCOVA) with baseline scores as covariates. The study was powered to detect a Cohen's d of 0.80 but was terminated early due to recruitment delays and funding constraints.

Key Results
The trial was terminated early. Of the 25 participants who underwent the assigned program, 23 were included in the primary analysis after excluding two participants from the traditional learning group who had missing questionnaire data preventing the calculation of the total score.

• Primary Outcome: There was no statistically significant difference in the mean total questionnaire scores between the blended learning group (76.7 ± 7.9) and the traditional learning group (75.0 ± 9.4) (P = .63).
• Secondary Outcomes:
  • DREEM Subscales: No significant differences were found in "Perceptions of Learning" (38.6 vs. 37.2, P = .43) or "Perceptions of Atmosphere" (38.0 vs. 37.9, P = .92).
  • Clinical Competencies: Adjusted mean differences in post-intervention self-assessment scores for all six clinical competencies were not statistically significant between groups.
• Adherence: Numerically higher adherence to the scheduled interviews was observed in the blended learning group (75.0%) compared to the traditional group (45.4%), though this was not a primary outcome.

Significance and Claims
The authors position this study as the first RCT to directly compare blended and traditional face-to-face learning specifically for medical interview training. However, they explicitly frame the findings as preliminary evidence rather than proof of equivalence, citing the study's status as an early-terminated pilot with a sample size substantially smaller than planned (23 vs. the target of 52–56).

The paper claims that:

1. Blended learning did not result in significant differences in learner-reported educational outcomes compared to traditional face-to-face learning in this specific context.
2. The lack of significant difference suggests that blended learning may serve as a feasible complementary approach rather than a replacement for traditional face-to-face learning.
3. The higher adherence in the blended group may indicate an implementation advantage in settings with scheduling constraints, potentially offering psychological safety for learners transitioning from classroom to bedside practice.

The authors caution that the study was underpowered to detect small differences and that the modified DREEM instrument used has not been formally validated for medical interview training. They conclude that while blended learning is a viable option, future multicenter studies with larger samples and objective performance measures are required to fully evaluate its efficacy.