Short-term efficacy analysis of applying DeepSeek R1 large model software in clinical teaching for medical undergraduates

Objective To evaluate the application and teaching effectiveness of artificial intelligence large model software DeepSeek R1 in undergraduate clinical teaching of thyroid and breast surgery, and explore its promoting effect and limitation on the cultivation of students′ clinical ability.

Methods A randomized controlled trial was conducted with 80 medical undergraduates divided into an observation group (DeepSeek combined with traditional teaching) and a control group (traditional teaching), with 40 students in each group.The observation group received personalized case recommendation, 3D anatomical simulation, and clinical decision feedback through DeepSeek, while the control group underwent theoretical lecture, bedside teaching, and case discussion.After a 4-week teaching period, the theoretical knowledge, clinical skill, case analysis, and doctor-patient communication ability in the two groups were evaluated.

Results The case analysis, clinical skill, and diagnostic accuracy and treatment plan rationality in the observation group were better than those in the control group (P < 0.05 to P < 0.01), while there was no statistically significant difference in doctor-patient communication ability between the two groups (P>0.05).

Conclusions DeepSeek effectively enhances the clinical teaching effectiveness in thyroid and breast surgery, but its limitations in emotional interaction and humanistic care needs to be integrated with traditional teaching and humanistic education to cultivate comprehensive medical professionals.