BRIEF RESEARCH REPORT article
Front. Artif. Intell.
Sec. Medicine and Public Health
A Deterministic LLM Framework for Safe, Protocol-Adherent Clinical Decision Support: Application in Hemodialysis Anemia Management (AnemiaCare HD)
Provisionally accepted
Jose Arriola-Montenegro
Charat ThongprayoonBenjamin Bizer
Jing MiaoKarina Ordaya-Gonzales
Iasmina M Craici
Wisit Cheungpasitporn*- Division of Nephrology and Hypertension, Mayo Clinic, Rochester, United States
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Background: Large language models (LLMs) show promise for clinical decision support but often deviate from evidence-based protocols, raising safety and regulatory concerns. Anemia management in hemodialysis patients requires strict adherence to erythropoiesis-stimulating agent (ESA) and intravenous (IV) iron dosing rules, making it a high-risk use case for uncontrolled model behavior. To address this gap, we developed AnemiaCare HD, a deterministic LLM framework engineered to deliver transparent, reproducible, protocol-adherent recommendations. Methods: AnemiaCare HD was evaluated using 600 simulated hemodialysis anemia scenarios derived from a standardized institutional protocol. The model required six fixed clinical inputs (hemoglobin, rate of hemoglobin change, trend direction, transferrin saturation, ferritin, and current ESA dose). Phase 1 tested a loosely structured prompt. Phase 2 implemented deterministic logic incorporating ESA kinetics, iron dosing rules, mandatory timing safeguards, and embedded safety alerts. Two nephrologists assessed protocol adherence. Results: In Phase 1, only 96 of 300 cases (32%) aligned with protocol recommendations, with common errors in ESA titration, iron dosing, and timing violations. Loosely structured prompting produced variable outputs and frequent unsafe recommendations. In contrast, deterministic prompting in Phase 2 resulted in 100% adherence across all 300 cases, eliminating protocol deviations, unsafe iron dosing, and timing violations (p < 0.001). Deterministic encoding generated structured, rationale-based outputs and prevented unsafe recommendations (p < 0.001 vs. Phase 1). Conclusion: Deterministic LLM engineering enables safe, fully protocol-compliant decision support in high-risk domains. AnemiaCare HD demonstrates the viability of transparent, auditable frameworks for clinical use, although real-world integration and prospective validation remain necessary next steps.
Keywords: Anemia management, artificial intelligence, end-stage kidney disease, hemodialysis, Large language models
Received: 19 Oct 2025; Accepted: 02 Dec 2025.
Copyright: © 2025 Arriola-Montenegro, Thongprayoon, Bizer, Miao, Ordaya-Gonzales, Craici and Cheungpasitporn. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Wisit Cheungpasitporn
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.