EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

ClinClaim-ARC: Evidence-Constrained Clinical Claim Repair with Multi-Risk Certification

Mingkun Bao, Dong Lu, Yifan Gu, Hao Ji, Yanyan Bu, Haogang Zhu

Information · 03.11.2026 · 17 (11) · 1097

DOI: 10.3390/info17111097 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Information, online erschienen am 03.11.2026. Verfasst von Mingkun Bao und 5 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Science and Technology Major Project, National Natural Science Foundation of China, Zhejiang Provincial Natural Science Foundation.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Medical large language models can produce fluent long-form responses containing contradicted, insufficiently supported, or over-specified clinical claims. We introduce ClinClaim-ARC, an evidence-constrained framework for post-generation clinical claim repair. The framework decomposes responses into structured claims, assigns four evidence states (supported, contradicted, unsupported, and partially supported), and constructs provenance-constrained repair candidates through a Clinical Claim Repair Graph. Unresolved claims trigger claim-targeted evidence acquisition, while Learn-then-Test certifies complete repair policies against Contradiction Retention Risk and Insufficient-Evidence Retention Risk. On a PubMedQA-derived answer-level repair benchmark, ClinClaim-ARC achieves a 76.4% valid repair rate and 87.9% supported-information retention, with held-out empirical contradiction and insufficient-evidence retention risks of 0.032 and 0.074, respectively, against pre-specified calibration targets of 0.05 and 0.10. On K-QA, the framework reduces Hall from 0.214 to 0.103 while maintaining a completeness score of 0.594 and a response rate of 93.5%. Ablation results further show that structured repair, four-state verification, and claim-targeted evidence acquisition each contribute to information-preserving reliability. These results show that evidence-constrained repair reduces reference-evidence-defined factual risk while retaining more supportable clinical information than filtering- only approaches.

Förderung

  • National Science and Technology Major Project
  • National Natural Science Foundation of China
  • Zhejiang Provincial Natural Science Foundation
  • Start-up Funds of Hangzhou International Innovation Institute of Beihang University

Zitieren

APA

Bao, M., Lu, D., Gu, Y., Ji, H., Bu, Y., & Zhu, H. (2026). ClinClaim-ARC: Evidence-Constrained Clinical Claim Repair with Multi-Risk Certification. Information, 17(11), 1097. https://doi.org/10.3390/info17111097

Vancouver

Bao M, Lu D, Gu Y, Ji H, Bu Y, Zhu H. ClinClaim-ARC: Evidence-Constrained Clinical Claim Repair with Multi-Risk Certification. Information. 2026;17(11):1097. doi:10.3390/info17111097

BibTeX
@article{Bao2026,
  author = {Bao, Mingkun and Lu, Dong and Gu, Yifan and Ji, Hao and Bu, Yanyan and Zhu, Haogang},
  title = {ClinClaim-ARC: Evidence-Constrained Clinical Claim Repair with Multi-Risk Certification},
  journal = {Information},
  year = {2026},
  volume = {17},
  number = {11},
  pages = {1097},
  doi = {10.3390/info17111097}
}
Kostenlos starten →