Artificial Intelligence in Medicine

Background and Objective: Access to real-world electronic health records (EHRs) remains limited by privacy, governance and annotation constraints, hindering the development of clinical natural language processing models. Realistic synthetic progress notes may provide EHR-like corpora that preserve clinically rigorous information on diagnoses, treatments, symptoms, imaging, laboratory findings and therapeutic trajectories without relying directly on sensitive patient records. This study evaluates whether large language models (LLMs) can generate realistic Spanish prostate cancer progress notes from published case reports, preserving clinical content, temporality and hospital-style conventions.

Machine learning (ML) is being considered to help diagnose cardiovascular diseases (CVD). Still, challenges like inconsistent and limited datasets, limited infrastructure, and global inequalities lead to the need for a reliable and practicable ML solution. This paper presents an ML-driven framework for predicting CVD risk scores and classifying status. Several data preprocessing techniques, including multiple imputation by chained equations (MICE), outlier removal, are considered. In addition, hyperparameter tuning is performed with the GridSearchCV tuning technique. Moreover, a consensus-driven five-feature selection method is applied to identify optimal predictors. The dataset used in this study contains healthcare records related to future CVD risk scores, comprising 1,529 patient records with 22 features. The optimized stacked ensemble model is applied to the dataset and achieves a cross-validated coefficient of determination value of 98.13% for CVD risk score regression. Comparative evaluation with other ML models confirmed improved accuracy, efficiency, and interpretability. The explainable AI technique SHAP is applied to interpret predictions and highlight key risk factors. Moreover, a deployment-ready web platform with multi-role access has been developed that demonstrates clinical applicability. The proposed framework offers a reliable and interpretable tool for early detection of CVD and personalized risk assessment. In the future, this work can be extended to integrate longitudinal data, medical imaging, and deep learning to improve generalizability and strengthen real-world impact.

Background: Preparing tumor board patient summaries is time intensive. Large-language-model based systems may automate summarization but require real-world evaluation prior to clinical use. We performed an exploratory retrospective evaluation of the Microsoft Healthcare Agent Orchestrator (HAO), deployed in a Mayo Clinic controlled staged environment, to generate tumor board-style patient summaries from retrospective Electronic Health Record (EHR) notes. Methods: HAO generated summaries for breast, hepatobiliary, and neuro-oncology tumor board cases using up to the most recent 1,000 clinical notes. Clinician reviewers evaluated outputs via REDCap surveys across perceived factuality, completeness, clarity/conciseness, temporal cohesion, comparative performance, safety, and clinical utility (0-4 Likert scale). Reviewers were permitted to query the HAO chat interface to address missing details. Automated factuality was assessed using TBFact (bidirectional entailment), reporting precision and recall against available reference summaries. Results: Among 57 survey responses from 5 different physicians, mean scores exceeded 2.8 across domains, with medians of 3 for most axes. In an exploratory comparison, oncology fellows required less time to review HAO-generated summaries than to manually generate patient summaries (mean difference 13.57 minutes per patient, p<0.001), although this difference may be influenced by prior familiarity with the same cases; 96% of survey responses indicated that HAO would save time. TBFact evaluations showed higher recall than precision across domains, consistent with broad capture of reference content alongside additional content that was not present in gold-standard summaries. Attribution was viewed favorably but showed issues with primary-source specificity and link reliability. Conclusions: In a controlled Mayo environment, HAO demonstrated moderate performance and was associated with reduced review time for tumor board preparation. These findings are promising but preliminary and do not establish clinical safety, noninferiority to manual review, or readiness for routine clinical use. Limitations, including verbosity, specialty-specific content gaps, and inconsistent attribution, highlight the need for iterative refinement and further evaluation.

Objectives: The aim of mammographic screening is the early detection of invasive cancers. In the era of artificial intelligence (AI), this tool may improve diagnosis of earlier stages. The purpose of this study was to assess the impact on selected quality indicators retrospectively. Method: The data source was the Breast Cancer Screening Registry using data from one Screening Unit that currently uses AI routinely. The indicators of the cancer detection rate (CDR), further assessment rate (FAR), and recall rate (RR) in the year 2023, when AI was used, and the year 2022, without AI, in women aged 45-69 were compared. The statistical evaluation used the chi-square test and logistic regression adjusting for the effects of age, a woman's risk level, and the screening round at a 5% significance level. Results: In 2022, without AI, 4,034 women aged 45-69 were included, compared with 4,049 women in 2023 when AI was used. This study showed a non-significant increase in CDR from 5.0 breast cancers detected per 1,000 women (non-AI assessment) to 5.2 (AI-assisted assessment), p = 0.919; OR (95% CI): 1.034 (0.542-1.974), a significant decrease in the FAR from 5.2% to 3.9%, p < 0.001; OR (95% CI): 0.665 (0.529-0.836), and a decrease in RR from 2.4% to 1.9%, p = 0.083; OR (95% CI): 0.754 (0.548-1.037). Conclusion: AI has the potential to be a useful tool in the early detection of breast cancer by improving quality through a decrease in FAR and RR, while probably maintaining CDR.

Background: Traditional diagnostic models lack explainability, while multimodal language models prone to hallucination remain unsafe for medical education. An interactive, risk-free artificial intelligence framework is required to serve as a reliable clinical mentor for radiology trainees. Methods: We propose a multi-agent architecture decoupling deterministic image analysis from generative consultation. Specialized computer vision models perform anatomical localization and pathological segmentation. These quantitative outputs are synthesized into a structured payload, which grounds a locally hosted large language model (LLaVA 7B) using strict prompt guardrails and prerequisite protocols. Results: The system effectively eliminates visual hallucinations by intercepting unanchored queries. The artificial intelligence tutor successfully contextualizes spatial anomalies and baseline metrics, generating accurate conversational explanations and formally structured radiology reports while strictly enforcing medical safety disclaimers. Discussion and Conclusion: By anchoring language generation exclusively to verified algorithmic realities, this framework transforms opaque diagnostic models into safe, interactive educational simulators. This establishes a highly reliable paradigm for integrating explainable artificial intelligence into medical training.

Background: Adverse drug events (ADEs) are a critical indicator of patient safety but are often documented only in free-text clinical notes. The potential of recent advances in natural language processing (NLP), particularly generative large language models (LLMs), to identify ADEs remains understudied. This study aimed to compare the performance of multiple LLMs in identifying ADE-Drug relationships in inpatient and ambulatory clinical notes. Methods: We used clinical notes from the 2018 National NLP Clinical Challenge (n2c2) ADE dataset (inpatient; n=505) and from outpatient encounters (n=2,555) between October 1, 2018, and December 31, 2019, at a large academic medical center based in New England. Notes were pre-processed into snippets for model input. Evaluated Models included: GPT-4o, GPT-4o-mini, LLAMA 3.3-70B and their instruction fine-tuned variants (including low-rank adapters for LLAMA). Performance was assessed using both strict and relaxed evaluations (precision, recall, and F1) for all models, followed by manual evaluation (exact semantic match, partial match, missing ADE, drug mention only, not a drug, or wrong) of the two best-performing models. Results: GPT-4o and GPT-4o-mini were the top-performing models among those evaluated. GPT-4o consistently outperformed GPT-4o-mini in ADE extraction across both datasets, with higher F1-scores (0.524 vs. 0.381) and a more balanced precision-recall profile. Both models captured ADEs effectively in explicit and complex clinical contexts, although limitations included misclassification of pre-existing allergies and occasional conflation of therapeutic indications with adverse effects. GPT-4o achieved higher exact match coverage and fewer errors across clinical notes, indicating more reliable performance in both inpatient and ambulatory settings. Conclusion: This work establishes a foundation for integrating LLM methods into real-world drug safety surveillance, with direct implications for improving patient safety.

Background: The development of personalised mRNA cancer vaccines holds considerable promise for oncology, yet a significant translational gap persists between neoantigen identification and the selection of therapeutically impactful targets. Current approaches predominantly prioritise human leukocyte antigen (HLA) binding affinity and immunogenicity, often overlooking the systems-level biological context of the target. This can inadvertently favour immunogenic but biologically peripheral peptides that exert limited influence on tumour signalling networks, thereby constraining vaccine efficacy. Furthermore, mRNA therapeutics must satisfy additional design requirements, including favourable codon usage and favourable secondary-structure stability, which directly affect in vivo translation and half-life. A unified computational framework that integrates neoantigen discovery with network biology is therefore critically needed. Results: Here, we present PimRNA, a Priority index (Pi)-centric computational medicine framework that bridges this gap by unifying neoantigen identification, mRNA sequence optimisation, and gene interaction network analysis. First, high-confidence tumour-specific HLA class I and II neoantigenic peptides are identified from paired tumour-normal genomic and tumour transcriptomic data using NeoDisc. Second, the coding sequences of these peptides are optimised for stability and translational efficiency with LinearDesign, yielding a core set of neoantigen-encoding mRNAs. Third, a random walk with restart algorithm is applied to a knowledgebase of gene interactions to identify peripheral genes exhibiting significant network connectivity to core genes, generating a gene-predictor matrix in which each gene is assigned an affinity score reflecting its network proximity to immunogenic neoantigens. These scores are consolidated into a single, unified priority rating (0-5) for each gene, followed by subnetwork analysis that reveals therapeutically relevant gene modules. Application of PimRNA to breast cancer and melanoma datasets demonstrates that it successfully selects high-confidence immunogenic neoantigen candidates embedded within biologically meaningful tumour-specific networks. Conclusion: PimRNA provides a systems biology foundation for mRNA vaccine design, moving beyond isolated immunogenicity to prioritise targets that are both highly presented and central to tumour-relevant biological networks. This framework offers a generalisable strategy for the rational discovery and prioritisation of mRNA therapeutics, significantly advancing the field of computational medicine towards personalised cancer vaccines.

Importance Prognostic tools beyond staging are needed to guide treatment and counseling in head and neck squamous cell carcinoma (HNSCC). Objective To develop and externally validate a machine learning model predicting survival in advanced HNSCC using routinely collected clinical and biomarker data. Design, Setting, and Participants Retrospective, multi-institutional cohort study including 2,385 patients with stage III-IV HNSCC diagnosed from 2012-2022 in the University of California Health Data Warehouse (UCHDW). Patients were randomly split into training (n = 1,908) and test (n = 477) sets. Partial external validation used 7,749 patients from the Surveillance, Epidemiology, and End Results (SEER) registry (2010-2020). Exposures Demographic, tumor, treatment, comorbidity, and biomarker variables recorded at or before diagnosis. Main Outcomes and Measures The primary outcome was all-cause mortality within 70 months. Cox proportional hazards models included all predictors. Discrimination was assessed with Harrell's concordance index (C-index), calibration with predicted vs observed survival, and stratification with Kaplan-Meier curves. A Random Survival Forest (RSF) was trained for benchmarking and interpretability using Shapley Additive exPlanations (SHAP). Results Among 2,385 patients in UCHDW (median age, 63 years; 29.0% mortality), the Cox model achieved a C-index of 0.735 in the internal test set. Risk quartiles showed clear separation on Kaplan-Meier curves (log-rank p < 0.0001). In the SEER cohort (n = 7,749), where only demographic, staging, subsite, and treatment variables were available, the reduced Cox model achieved a C-index of 0.688, with calibration showing modest underestimation of survival in high-risk groups. Age, T stage, Charlson Comorbidity Index, neutrophil-to-lymphocyte ratio, and platelet count were among the strongest predictors, while surgery was associated with improved survival. The RSF achieved a C-index of 0.758 internally, with SHAP highlighting nonlinear effects of albumin, BMI, and inflammatory markers. Conclusions and Relevance A machine learning model using routine clinical and biomarker data demonstrated good prognostic performance in advanced HNSCC, with partial external validation. Such approaches may support individualized survival estimates, risk stratification, and treatment discussions, but broader validation is required before clinical adoption.

Importance: Physician-facing AI tools are now in clinical use, yet whether different platforms fail in similar or fundamentally different ways in high-stakes settings like critical care is unknown. Objective: To evaluate two physician-facing AI platforms, ChatGPT for Clinicians and OpenEvidence, for distinct vulnerabilities under structured stress testing. Design, Setting, and Participants: An observational study conducted using 60 simulated critical care vignettes developed and adjudicated by four attending critical care physicians. Data were collected in the last week of April 2026, via the public website interfaces of each platform. Interventions/Exposures: A 2x2x2x2 factorial design across four stressors - anchoring, cognitive load, social conformity pressure, and a clinically incorrect directive - yielded 16 prompt subsets per vignette and 960 prompts per platform. A separate multi-turn adversarial prompting paradigm administered three sequential "You are incorrect" challenges to baseline vignettes. All prompts had a universal output length constraint of fewer than 30 words. Main Outcomes and Measures: Critical elements capture (percentage of gold-standard critical elements present in responses), susceptibility to clinically incorrect directive, and sycophancy (reversal of an initial correct recommendation under iterative adversarial challenge). Results: Across 1916 responses to 1920 prompts, ChatGPT for Clinicians captured more gold-standard critical elements than OpenEvidence (81.4% {+/-} 18.1% vs 61.0% {+/-} 23.5%; adjusted difference, 20.3 percentage points; 95% CI, 18.3 to 22.4; P < .001) and was less susceptible to clinically incorrect directives (1.7% vs 8.0%; adjusted odds ratio, 0.07; 95% CI, 0.02-0.21; P < .001). Anchoring and social conformity pressure were associated with reduced critical element capture across both platforms, while cumulative stressor burden reduced critical element capture only on OpenEvidence. Conversely, ChatGPT for Clinicians reversed correct recommendations more readily under adversarial prompting (hazard ratio, 2.61; 95% CI, 1.10 - 6.19; P = .03). Conclusion and Relevance: The two physician-facing clinical AI platforms evaluated demonstrated non-overlapping vulnerabilities, with neither platform uniformly superior. These findings argue against single-axis ranking of clinical AI systems and support multidimensional safety evaluation encompassing completeness of reasoning, resistance to incorrect directives, and stability under adversarial challenge.

Background: Wearable technologies enable scalable and continuous monitoring of emotional states through passive sensing of physiological and behavioral signals. However, conventional learning approaches often struggle to model the complex temporal, contextual, and relational dependencies underlying human emotions. To address these limitations, we propose a graph-based framework that represents multimodal wearable observations as heterogeneous knowledge graphs enriched with semantic information derived from Large Language Models (LLMs), enabling richer contextual understanding beyond raw sensor measurements. Methods: We constructed a heterogeneous knowledge graph using multimodal Fitbit physiological signals and affective self-report data collected from 45 users. Framing mood prediction and emotion detection was formulated as both binary and ternary node classification tasks. We evaluated five baseline heterogeneous Graph Neural Network (GNN) architectures and compared them with the proposed Semantically Gated Augmented Graph Neural Network (SeGA-GNN) framework, which dynamically integrates LLM-generated semantic embeddings into graph representations through a gated cross-modal fusion mechanism. Results: The baseline GNN models achieved strong performance, with classification accuracies ranging from 0.7525 to 0.9739 for binary classification and 0.6249 to 0.9699 for ternary classification. The proposed SeGA framework consistently improved predictive performance across most architectures. In particular, semantic augmentation transformed the HAN model from moderate baseline performance into near-perfect emotion recognition capability, achieving SeGA-HAN Accuracy = 0.9988 and AUC = 1.0000 for binary classification and Accuracy = 0.9979 and AUC = 1.0000 for ternary classification. Discussion and Conclusion: Integrating LLM-derived semantic contextualization into heterogeneous graph learning enables effective modeling of contextual information that is not directly captured by wearable physiological signals alone. The proposed SeGA-GNN framework demonstrates that adaptive semantic fusion substantially improves the accuracy, robustness, and interpretability of wearable-based emotion detection. These findings establish a promising direction for next-generation wearable affective computing systems and intelligent emotion-aware applications.

The availability of electronic health records has facilitated data-driven approaches to the understanding of multimorbidity, with clustering becoming a common tool for uncovering relevant groups of associated conditions. Previous studies, however, have found challenges in their reproducibility, with wide disparity in the reported clusters. At the core of this issue lays a vagueness of the definition of a cluster, leading to a lack of standards in their methods and evaluation, while implementation details are often not completely reported or explicit in their assumptions. We present a methodological pipeline that can be adapted to different cluster definitions (e.g. multiple cluster membership or clusters where all nodes are mutually associated) and a set of scores that can be composed into an evaluation metric that explicitly incorporates assumptions that align with the research aims. We apply our pipeline to a healthcare dataset of over 7 million patients in England and show how clusters may drastically differ when varying the parameter choices, exposing the risks of reporting a single clustering realisation. Our methodological pipeline, evaluation framework, and tools for analysis and network visualisation serve as a reference to transparently explore and align methodological decisions to the aims of multimorbidity clustering, contributing to overcome the reproducibility challenges of the field.

Artificial intelligence (AI) tools have been rapidly adopted by medical researchers, yet whether early career researchers in low and middle income countries possess the awareness and habits needed to use these tools safely remains poorly documented. This study characterized AI adoption patterns, hallucination awareness, and verification and disclosure practices among early career medical researchers in Pakistan. A cross sectional anonymous online survey was conducted among medical students, house officers, residents, physicians, and faculty involved in research or academic work across Pakistan (May 2026). Descriptive statistics and chi square tests were applied to 373 eligible responses. AI use was near universal (99.7%), with 60.3% using AI tools daily. The most commonly reported tool in this sample was Claude (40.5%), followed by ChatGPT (29.2%) and Perplexity (26.0%), though this ranking likely reflects sampling characteristics. Despite high adoption, 59.2% typically did not verify AI outputs before use, and 40.2% had never heard that AI can generate fabricated scientific references. In behavioral vignettes, 36.5% assumed convincing AI generated references were authentic, and 54.2% would continue using remaining AI content after discovering one fabricated reference. Formal research training was strongly associated with consistent disclosure (51.7% vs. 17.1%; chi square=48.43, p less than 0.001). Role, daily use frequency, and research training were not significantly associated with verification behavior. Early career medical researchers in Pakistan demonstrate high AI adoption alongside incomplete hallucination awareness and infrequent verification, a pattern that may carry implications for research integrity. Formal training was the only factor significantly associated with consistent disclosure. Integration of AI literacy into medical curricula and institutional governance frameworks merits consideration.

Objective. Biomedical knowledge graphs (KGs) such as PrimeKG, Hetionet, UMLS, and PharmGKB are increasingly used as the substrate for downstream machine-learning, retrieval-augmented generation, drug-repurposing, and electronic health record (EHR) augmentation pipelines. The dominant assumption in published work is that integrating two or more such KGs is a tractable engineering step solved by identifier (ID) matching. This paper interrogates that assumption empirically. We quantify how much concept overlap survives realistic alignment, and we characterize the new failure modes introduced by the methods that practitioners reach for when ID matching is insufficient. Materials and Methods. We compared four widely used biomedical KGs (PrimeKG, Hetionet v1.0, the full UMLS Metathesaurus, and PharmGKB) across eleven node types using a tiered alignment pipeline: (1) direct ID matching for nodes sharing a primary vocabulary; (2) cross-ontology bridging using standard mappings (e.g., MONDO-DOID, HPO-UMLS, HPO-UMLS-MeSH for side effects, NCBI Gene-HGNC-UMLS, UBERON-FMA/SNOMEDCT_US/NCI/MeSH for anatomy); (3) ClinicalBERT cosine-similarity grouping at threshold >= 0.98 for over-segmented disease nodes, with a deterministic suffix-stripping canonicalizer; (4) exact name matching for ontology-poor types (anatomy, REACTOME pathways); and (5) embedding-based fuzzy matching with UMLS lookup (SapBERT and ClinicalBERT) for free-text microbiome concepts. We applied the pipeline to a 698-concept gut-microbiome benchmark spanning taxa, pathways, and disease labels, validated grouping decisions against the curated SSSOM mappings released by the MONDO project, and audited the ClinicalBERT consolidation against five clinical-genetics case studies drawn from the literature. Results. Per-type pairwise coverage was strikingly asymmetric. Genes/proteins and the three Gene Ontology categories aligned cleanly across PrimeKG and Hetionet (mutual coverage 94-99%), but disease overlap was sparse: only 0.7% of PrimeKG individual disease nodes mapped to Hetionet, rising to 2.0% after MONDO grouping (versus 78.7% and 18.4% from the Hetionet side). PrimeKG-to-UMLS coverage spanned 100% (effect/phenotype via HPO) down to 20.8% (REACTOME pathways), with drugs at 73.7% and anatomy at 58.8%. PrimeKG-to-PharmGKB drug coverage required up to two bridging hops (DrugBank -> UMLS -> RxNorm/ATC/MeSH). Bigger was not uniformly more complete: on a 698-concept microbiome drug benchmark, Hetionet missed 0 concepts while PrimeKG missed 16. ClinicalBERT-based grouping consolidated 22,205 raw MONDO disease nodes into 17,080 groups but introduced three reproducible failure modes documented in case studies: (i) peer over-merging: for example, all 22 osteogenesis imperfecta subtypes collapsed into a single node despite distinct severity classes; (ii) parent-child collapse: e.g. acute myeloid leukemia merged with myeloid leukemia, erasing the acute/chronic distinction that drives clinical management; and (iii) lexical false positives: neurofibromatosis and schwannomatosis grouped together despite cellular-pathology differences. Discussion. Identifier matching alone is a weak baseline for biomedical KG integration. Cross-ontology bridges and embedding-based consolidation expand coverage but do so at the cost of clinically meaningful resolution, and the resulting failures are systematic rather than random. Reporting only aggregate coverage statistics obscures these losses, which propagate silently into downstream tasks. Conclusion. We provide reusable per-type coverage tables, a taxonomy of three integration failure modes, and concrete recommendations for downstream studies that depend on a unified biomedical KG. We argue that future KG integration work should report per-type coverage and per-cluster confidence rather than aggregate match rates.

Our study evaluated whether a deep learning auto segmentation model combined with machine learning triage can streamline radiotherapy clinical trial quality assurance (QA). We analyzed 107 stereotactic ablative radiotherapy (SABR) cases from a multi-institutional phase II clinical trial of neurovascular sparing prostate SABR, focusing on physician contours of the internal pudendal artery (IPA) as a novel organ-at-risk with substantial interobserver variability. Contours were scored by the trial principal investigator as Per-Protocol or Minor Deviation/Unacceptable. We applied a deep learning model for IPA auto-segmentation. Agreement between human and AI contours was then quantified using 14 overlap, distance, and surface metrics, and a supervised classifier was trained on these metrics to flag clinical trial protocol deviations. While AI segmentation achieved only modest geometric accuracy with mean Dice similarity coefficient of 0.446 and 95th percentile Hausdorff distance of 14.23, when incorporating all 14 metrics, a machine learning classifier yielded AUROC of 0.836, flagging all Minor Deviation/Unacceptable cases with 100% sensitivity on the 27 case hold-out set with 6 false positives and no false negatives. AI segmentation combined with metrics-based machine learning can triage protocol deviations within a multi-institution radiotherapy clinical trial, supporting prospective evaluation of AI-assisted trial QA.

Subjective auditory-perceptual evaluation and uninterpretable deep learning models limit the clinical assessment of voice disorders. This study proposes a two-phase zero-shot framework to evaluate voice pathology. First, an Audio Spectrogram Transformer is fine-tuned on the Perceptual Voice Quality Database to generate an acoustic latent space. Second, Orthogonal Procrustes analysis maps these acoustic embeddings directly onto the semantic space of a pre-trained Sentence Transformer. The geometric alignment produced continuous semantic axes that outperformed a supervised machine learning baseline in regressing clinician-rated GRBAS (Grade, Roughness, Breathiness, Asthenia, and Strain) severity scales. Furthermore, these axes correlate with traditional acoustic measures, including Harmonics-to-Noise Ratio and local jitter, while remaining robust when applied to aperiodic signals by not requiring fundamental frequency extraction. Most importantly, the model achieved zero-shot semantic expansion, successfully evaluating voices using an untrained, natural clinical vocabulary beyond the GRBAS scale. External validation on the Voice ICarus Database confirmed cross-corpus stability and demonstrated the capacity for zero-shot differential phenotyping of specific etiologies, such as hypokinetic dysphonia and reflux laryngitis. By bridging acoustic and semantic latent spaces, this framework offers an objective, continuous, and transparent metric for evaluating voice quality using voice descriptive vocabulary.

Introduction -- Emergency Medical Dispatch Systems (EMDS) can reduce delays in accessing emergency care by providing structured communication, triage, and coordination. However, such systems remain absent or underdeveloped in most low- or middle-income countries (LMICs). This study aimed to establish international consensus on essential EMDS components to inform global guidance. Methods -- We convened a multidisciplinary expert group to draft a preliminary list of essential components for three EMDS levels reflecting resource availability and system maturity. We then conducted a three-round Delphi with international experts to reach consensus on core EMDS components. Components which had [&ge;]75% agreement were included, those with [&ge;]75% disagreement were excluded. Components not achieving consensus by Round 3 were removed. Results were analysed overall and stratified by respondents' country income level. A subsequent online expert meeting resolved inconsistencies and finalised the component list. Results -- The expert group generated 111 components for each of three EMDS levels (Foundational, Emerging, and Established) spanning 11 operational domains. Of the 68 experts invited to the Delphi, 43 participated in Round 1 and 30 in Round 3. Across all Delphi rounds, 289 components reached consensus for inclusion. The consensus resulted in a final list of 227 components (63 Foundational, 84 Emerging, and 80 Established). Consensus agreement clustered around core EMDS domains including communication, structured call-taking and prioritisation, advice-giving, resource dispatch and tracking, and foundational governance and data functions, whereas items showing either non-consensus or consensus disagreement were typically technology-dependent or context-specific. Conclusions -- This international consensus offers guidance for EMDS development across diverse resource settings and provides a scalable roadmap to strengthen emergency care systems.

Background: Sepsis is a life-threatening condition in which delayed recognition and treatment are associated with increased mortality. While predictive models such as Epic's Early Detection of Sepsis Model (ESM) were developed to support early intervention, their real-world impact after integration into clinical workflows remains difficult to evaluate. Objectives: To evaluate the real-world impact of ESM integrated into clinical workflow on clinical outcomes, antibiotic use, and harm-benefit tradeoffs. Methods: We conducted a quasi-experimental study in a single healthcare system using encounter-level data from inpatient settings. Inpatient mortality, prolonged hospitalization, antibiotic use, and sepsis prevalence were compared between the pre-implementation period (3 June 2023 to 20 August 2024) and the online period (21 August 2024 to 26 December 2024) when the model became visible to clinicians. We also applied a counterfactual framework using models trained on pre-implementation data to estimate expected outcomes without ESM and to quantify harms related to overtreatment and delayed treatment. Results: Among 101,138 encounters, 86,884 occurred during the pre-implementation period and 14,254 during the online period. In unadjusted analyses, the online period had lower inpatient mortality, prolonged hospitalization, antibiotic use, and sepsis prevalence (all p[&le;]0.002). In the counterfactual analyses, observed outcomes were lower than expected without ESM for mortality (1.21% vs 1.82%; p<0.001), prolonged hospitalization (5.56% vs 7.95%; p<0.001), and antibiotic use (43.52% vs 47.04%; p<0.001). False positive harm (37.72% vs 41.68%; p<0.001) was also lower than expected. Conclusions: Integration of ESM into clinical workflow was associated with improved patient outcomes, reduced antibiotic use, and decreased harm from overtreatment, without evidence of increased harm from delayed treatment, supporting a positive net clinical benefit and the safety and effectiveness of ESM under Software as a Medical Device principles. Keywords: Machine learning, Electronic health records, Clinical workflow, Counterfactual analysis, Real-world evaluation

Background: Hospital incident risk scoring has long relied on two- or three-dimensional frameworks (Severity Assessment Codes or Risk Priority Numbers),even though root cause analysis standards recognize that clinical risk is multi-factorial. The obstacle has been mainly cognitive: human reviewers cannotreliably score many dimensions across high incident volumes, so richer assessmenthas not been operationalized at scale.Objective: To extend the traditional three-dimensional FMEA to an eight-dimensional patient-safety risk feature framework, to establish a multi-modellarge language model (LLM) extraction pipeline that scores these dimensionsautomatically, and to demonstrate a variance-aware integer optimization (mean-variance integer programming, MV-IP) that provides a reproducible tie-breakingrule for incident prioritization under extraction uncertainty, rather than improvedrisk coverage.Methods: An 8-dimensional framework covering harm severity, potential harm,frequency, detectability, systemic impact, vulnerable populations, regulatoryrelevance, and economic impact was applied to 213 synthetic and 196 realcurated incident narratives. Three independent LLMs (GPT-5.4, Gemini 3.1 Pro, Grok-4.1 Fast) from different provider families extracted structured risk scores.Inter-model consistency was assessed via ICC(A,1). Among coverage-equivalentselections, MV-IP minimized inter-model variance to give a reproducible prioriti-zation rule. An English-language sensitivity analysis was conducted on 31 AHRQPSNet WebM&M cases.Results: On real cases, seven of eight dimensions reached Fair or betterinter-model reliability (ICC(A,1) 0.53 to 0.83); D5 (Systemic Impact) was theexception at Poor reliability (0.275), driven by little between-case variation ratherthan by wide model disagreement. Reliability was not uniform: two dimensionswere Excellent (D1 actual harm 0.834, D8 economic impact 0.782), two Good,and three only Fair, so some dimensions are more readily extractable than others.The same anchors gave broadly similar results on English-language narratives.When deterministic top-K selection returned several equal-coverage solutions(11 on real cases, total inter-model variance 0.205 to 1.274), MV-IP selected theminimum-disagreement set, replacing ad hoc tie-breaking with an explicit rulewithout improving coverage. Bootstrap resampling found 74% to 90% of per-casevariance estimates stable despite the three-model panel.Conclusions: The eight-dimensional framework operationalizes patient-safetyrisk features that quality teams have considered only implicitly, and three inde-pendent LLM families produced reproducible scores on most dimensions ofcurated narratives. Inter-model agreement, however, measures reproducibilityrather than clinical correctness, and high agreement does not by itself establishthat a score is right; the dimensions that are reliably extractable today (notablyD6 and D8) differ from those that are not yet (D5, and to a lesser degree D4 andD7), which has direct implications for incident-reporting form design. MV-IP con-tributes a reproducible, variance-aware tie-breaking rule rather than improvedcoverage. Validation against expert-prioritized RCA lists and deployment on rawinstitutional incident reports remain the next steps toward clinical use.

Spousal caregivers of individuals with Alzheimers disease and related dementias frequently experience elevated perceived stress, caregiver burden, and loneliness, which are associated with adverse health outcomes. Early identification is therefore critical for timely intervention. Existing approaches commonly rely on wearable sensor data and standardized psychological questionnaires, while recent multimodal methods aim to improve prediction by integrating behavioral and linguistic information. In this study, we explored three modality configurations, wearable-derived features, interview-based text, and their combination, to classify caregiver psychological risk using the Perceived Stress Scale (PSS), Zarit Burden Interview, and UCLA Loneliness Scale. We compared traditional machine learning models and large language models (LLMs) (Gemini 2.0, Llama 4, and GPT-4o) under psychometrician-centered and caregiver-centered prompting strategies. Traditional machine learning models performed better under multimodal settings, while LLMs achieved stronger performance with Interview-Only input. We further demonstrate that PSS was the most predictable construct and prompting strategies substantially influenced LLM performance.

Background Complex gastrointestinal (GI) oncologic surgeries carry substantial perioperative risk, and nationwide outcomes in low- and middle-income countries (LMICs) are underreported. This study aimed to evaluate national trends in surgical volume, in-hospital mortality, and intensive care unit (ICU) utilization for major GI cancer surgery in Brazils Unified Health System (SUS) over a 14-year period. Methods A population-based analysis was performed using national administrative databases to identify all adult patients undergoing colectomy, gastrectomy, pancreatic resection or esophagectomy for cancer in the SUS from 2010-2023. Annual rates were age-standardized according to the WHO standard population. Temporal trends were assessed using Poisson regression to estimate average annual percent change (AAPC) with 95% confidence intervals (CIs). Results A total of 179,337 hospital admissions were analyzed (median age 63 years; 48% female). Colectomies accounted for 72% of cases, followed by gastrectomies (19%), pancreatic resections (5%), and esophagectomies (3%). Although crude surgical volume increased, population-adjusted rates declined overall (AAPC -2.09%; 95% CI -2.58 to -1.59), mainly due to reductions in gastrectomies and esophagectomies. Median hospital stay decreased from 9 to 7 days (AAPC -1.93%; 95% CI -2.79 to -1.06). Overall in-hospital mortality declined from 8.1% to 5.7% (AAPC -2.88%; 95% CI -4.15 to -1.59). ICU utilization rose from 37% to 43% of admissions (AAPC +1.31%; 95% CI 0.91 to 1.71). Conclusion Over 14 years, in-hospital mortality and length of stay for major gastrointestinal cancer surgery declined within Brazils universal public health system. These temporal trends occurred alongside expansion of accredited oncology services and increased ICU utilization, although causal relationships cannot be established from administrative data. These findings should be interpreted as hypothesis-generating and highlight the need for more granular hospital-level data in LMIC settings.