Do clinical interview transcripts generated by speech recognition software improve clinical reasoning performance in...

Area of Science:

• Medical Education Technology
• Clinical Skills Assessment
• Speech Recognition Applications

Background:

• Evaluating medical interviews is crucial for training competent physicians.
• Traditional feedback methods may lack specificity and precision.
• Speech recognition software offers a potential tool for enhanced feedback.

Purpose of the Study:

• To determine if speech recognition software (SRS) provides more specific and precise feedback for medical interview evaluation.
• To compare the impact of SRS feedback versus voice-recording feedback on medical students' performance.

Main Methods:

• A prospective observational trial involving 79 medical students.
• Students were assigned to either SRS feedback (n=39) or IC recorder feedback (n=40).
• Performance was assessed using diagnostic accuracy, mini-clinical evaluation exercise (mini-CEX), and a checklist after mock patient encounters.

Main Results:

• The SRS feedback group showed significantly higher diagnostic accuracy rates (89.7% vs. 67.5%).
• SRS feedback group achieved higher mini-CEX scores for clinical competence (7.4 vs. 6.1).
• Checklist scores for clinical performance were also significantly higher in the SRS group (16.1 vs. 13.8).

Conclusions:

• Speech-recognition-based feedback enhances diagnostic accuracy in medical interviews.
• SRS feedback leads to improved overall clinical competence and performance as measured by mini-CEX and checklist scores.
• This technology represents a valuable tool for medical education and skills assessment.