Skip to main content
Search

# Clinical Decision Support

10 related articles

Subscribe to updates
  1. arXiv

    Clinical imaging decisions get a harm gate: HC-ROCP cuts severe-case misrouting from 4.68% to 0.54%

    A clinical image classifier that returns a set of possible diagnoses does not guarantee that a severe case avoids a harmful action. HC-ROCP adds a harm penalty weighted by predicted severe-class probability to frozen classifier probabilities, tests coverage and severe-action error on disjoint calibration splits, and then certifies, reports descriptively, or refuses each claim. On COVID-19 radiographs across ten calibration seeds, severe-action errors fall from the exact ROCP baseline's 4.68% to 0.54%, at higher modeled loss.

  2. Patient-safety reasoning without the knowledge graph loses 0.047 F1, while retrieval's gain straddles zero

    Patient-safety reasoning requires linking clinical narratives, medications, laboratory findings and adverse events, and a language model alone can produce plausible but unsupported conclusions. The same MIMIC-III/IV cases ran under four conditions with one LLaMA 2 7B model and one prompt: model only, plus retrieval-augmented generation, plus a patient-safety knowledge graph and ontology, and both. The full framework reached F1 0.89 for concept extraction, 0.046 above baseline; removing the knowledge graph cost the most, with F1 0.047, AUROC 0.084 and AUPRC 0.110, while removing retrieval gave an F1 difference of 0.029 whose interval crossed zero.

  3. Patient-safety reasoning leans on a knowledge graph: removing it drops F1 by 0.047 and worsens all three metrics

    A patient-safety reasoning system combines a knowledge graph, retrieval-augmented generation and a LLaMA model to extract safety concepts from MIMIC-III and MIMIC-IV records and produce traceable explanations. The complete system reached an F1 of 0.89 for concept extraction; removing the knowledge graph lowered F1 by 0.047, AUROC by 0.084 and AUPRC by 0.110, all significant. Removing retrieval or LoRA adaptation gave smaller, non-significant drops, and retrieval relevance was confirmed only qualitatively by experts.

  4. Sepsis triage stops handing clinicians a black-box score: ShAprox turns Shapley values into a readable sum, reaching 0.824 AUC on external test

    Sepsis is often flagged at emergency triage, but a neural network returns only a probability that clinicians cannot check. ShAprox fits each feature's Shapley value to a curve or mean and adds them to a base rate, so the prediction is a readable sum. Across 28,072 emergency attendances at two NHS trusts, ShAprox reached 0.824 external-test AUC, comparable to the neural network's 0.818, with training and internal-test performance closer together.

  5. Medication recommendation stops at a drug list: RES-MR writes the diagnostic reasoning first, then prescribes inside each patient's own risk boundary

    RES-MR splits medication recommendation into two steps, first generating a checkable diagnostic reasoning chain and then adjusting the prescription to each patient's own risk boundary, reaching a 7.7% higher Jaccard than FLAME on MIMIC-III with lower DDI and RIS.

  6. This short communication proposes that AI could ease seafarers' limited healthcare access, fatigue, and mental health risks through telemedicine, clinical decision support, and predictive analytics, provided data privacy, transparency, and equitable access are safeguarded.

    This short communication explores the potential role of artificial intelligence in seafarers' occupational health, noting that seafarers face persistent challenges including limited access to healthcare, fatigue, and mental health risks, and suggesting that AI-enabled applications in telemedicine, clinical decision support, and predictive analytics could enhance prevention, early detection, and continuity of care at sea, while stressing that effective integration requires ethical safeguards for data privacy, transparency, and equitable access, which may contribute to safer and more resilient maritime health systems.

  7. JBJS Open Access

    Generative Artificial Intelligence in Hip and Knee Arthroplasty: A Systematic Review of Emerging Clinical Applications in Patient Communication and Education, Documentation, and Decision Support

    This systematic review searched PubMed and Embase (July 9, 2025) and included 23 studies to assess generative AI, mainly ChatGPT 3.5/4, in total hip and knee arthroplasty across patient communication and education (n=19), clinical documentation (n=2), and clinical decision support (n=2): blinded ratings found FAQ responses comparable to surgeon-written answers in accuracy, clarity, and completeness with better readability; consent documents showed better readability and completeness than surgeon versions; operative-report extraction reached 97.5%–100% agreement; decision support showed higher accuracy for surgical candidacy but low specificity for outcome prediction, alongside fabricated citations and limited patient trust.

  8. Large Language Models in Multidisciplinary Decision-Making for Hepatopancreatobiliary Oncology: Retrospective Comparative Feasibility Study

    This retrospective study submitted standardized summaries of 107 single-center hepatopancreatobiliary (HPB) multidisciplinary team (MDT) cases four times each to four large language models (GPT-4o, GPT-5.2, Gemini 3 Pro, and Claude Sonnet 4.5), finding that response stability differed significantly across models (Gemini 3 Pro lowest discordance, mean 12.8%, Fleiss κ=0.737; GPT-4o highest, mean 30.2%, Fleiss κ=0.430), that concordance with MDT decisions ranged from 48.6% to 72.9% using the initial response and from 66.3% to 74.5% using the modal response, and that class-wise F1 was consistently lower for surgery (0.400–0.520) than for chemotherapy (0.621–0.836), indicating that concordance alone is insufficient for evaluating LLMs as clinical decision support tools.

  9. Research Square

    Medical Artificial Intelligence: A Multimodal, Human-Centered, Domain-Adaptive Framework for Surgical Decision Support and Patient Safety

    This paper proposes an evidence-informed conceptual framework that links multimodal clinical data (electronic health records, laboratory data, imaging, physiological monitoring, surgical video, device telemetry) through a modular AI architecture, domain adaptation, uncertainty estimation, and safety controls to a clinician-facing decision-support interface, explicitly separating the proposed architecture from evidence already reported in the literature and noting that clinical effectiveness of surgical AI remains to be established prospectively.

  10. Fine-Tuning, Retrieval-Augmented Generation, and Hybrid Adaptation of Language Models for Clinical Decision-Making in Health Care: Systematic Review

    Following PRISMA 2020 guidelines and searching PubMed/MEDLINE, Scopus, and Web of Science (January 2018 through May 2026), this systematic review screened 1890 records and included 35 studies published between 2024 and 2026, grouping them into fine-tuning or parameter-efficient fine-tuning (7/35, 20%), retrieval-augmented generation (17/35, 48.6%), and hybrid approaches (11/35, 31.4%) for descriptive synthesis, finding that fine-tuning performed strongly on narrow task-specific applications (area under the receiver operating characteristic curve up to 0.912 for cancer detection and area under curve 0.892 for major depressive disorder prediction), that retrieval-augmented generation improved guideline adherence and diagnostic accuracy (from 71.1% to 92.1% and from 78.9% to 94.

You’re all caught up