Consumer AI health assistants increasingly connect medical records, patient reports, and wearable data across repeated interactions, but evaluation still focuses on isolated prompts and answers. This Perspective has two objectives: to define a journey-level evaluation and governance framework for record-linked patien...
Zonghai Yao, Hong Yu· npj Digital Medicine· 0 citations
Large language model systems can retrieve and cite medical evidence, yet their outputs may remain difficult to review during care. We define reviewability through evidence adequacy for action-changing claims, exact source verification, and presentation matched to a specified reviewer and decision point. We propose a...
Zonghai Yao, Hong Yu· npj Digital Medicine· 0 citations
Next-encounter ICD forecasting predicts which standardized diagnosis codes will be documented at a future visit from the longitudinal record available beforehand. The task is prospective and multi-label: the target note does not yet exist, and several codes may be correct. Structured EHR foundation models capture recur...
Junda Wang, Meysam Ghaffari, Akshat Choube et al.· 0 citations
MultiViewDx, a partly physician-validated multimodal instruction dataset for evidence-linked multi-view medical imaging diagnosis, is introduced and Controlled ablations and clinician evaluation show that both case-level multi-view organization and evidence-linked reasoning targets contribute to the gain.
Jun-Da Wang, Zonghai Yao, Yujan Ting et al.· Transactions of the Associat...· 2 citations· ⚡1
This work introduces MUSES, a million-instance benchmark for prospective intellectual-roots retrieval over a fixed 2.33M-paper corpus, and releases CiteRoots, both CiteRoots layers, and a distilled open companion judge for future work on prospective retrieval and intellectual roots.
This study evaluated both closed-source and open-source large language models for extracting and prioritizing medical jargon from EHR notes relevant to individual patients, leveraging prompting techniques, fine-tuning, and data augmentation and found that model performance could deviate largely based on prompting style...
W. Jang, Sharmin Sultana, Zonghai Yao et al.· JMIR AI· 1 citation
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
