Source record

Published by arXiv cs.MA (T1 source). The original is at https://arxiv.org/abs/2609.10939.

Pipeline notes

The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.

RCT (N=100 medical students) evaluating a scaffolding-oriented multi-agent LLM platform for clinical interview training, comprising a patient agent, a Socratic tutor agent, and a turn-level evaluator. The multi-agent condition improved OSCE scores in communication, empathy, and history-taking versus a control with progressive information disclosure, though final diagnostic accuracy did not differ. A multi-expert annotated dataset is released.

Provides controlled-trial evidence that role-separated agent scaffolding can raise process quality in simulated training without inflating outcome scores, a useful signal for multi-agent tutoring system design.

Cited by

No citations on record.