BMC Medical Informatics and Decision Making

BackgroundType 2 diabetes mellitus (T2DM) is a leading global public health challenge. Machine learning (ML) combined with Explainable AI (XAI) is increasingly applied to T2DM risk prediction, but the field lacks a quantitative overview of methodological trends and integration gaps. MethodsWe present a structured synthesis and critical analysis of the XAI literature on T2DM risk prediction, combining (i) quantitative bibliometric analysis of a two-database corpus (N = 2,048 documents from Scopus and PubMed/MEDLINE, deduplicated via a transparent three-tier pipeline) and (ii) an in-depth selective review of 15 highly cited papers. Reporting follows PRISMA 2020, adapted for metadata-based synthesis; analyses include keyword frequency, rule-based thematic clustering, and publication trend analysis. ResultsThe field grew rapidly, from 36 documents (2020) to 866 (2025). SHAP and LIME dominate XAI methods; XGBoost and Random Forest dominate ML models. Critically, KG/GNN terms appeared in only 17 documents ([~]0.83%) compared with 906 for XAI methods, a 53.3:1 disparity. This gap is consistent across both databases, which share 33.2% of their records, ruling out a single-database artifact. The selective review confirmed that none of the 15 highly cited papers combined all three components, ML, XAI, and KG, in T2DM risk prediction. ConclusionsThe XAI for T2DM risk prediction field exhibits a clinical interpretability gap: statistical explanations are rarely linked to structured clinical pathways. We propose a three-layer conceptual framework (Predictive [->] Explainability [->] Knowledge) that integrates KG as a supplementary semantic layer, with potential applications in clinical decision support and population-level screening. The framework does not perform true causal inference but structures explanations around established pathophysiological knowledge. This study contributes a transferable methodology and a quantified research gap to guide future work integrating ML, XAI, and structured medical knowledge.

Background: China has implemented policies to strengthen its pharmacist workforce since the 2009 healthcare reform, yet a comprehensive evaluation of their long-term systemic effects is lacking. Objective: To systematically analyze the evolution of Chinas pharmacist workforce in healthcare institutions from 2007 to 2023 across four dimensions: quantity, quality, structure, and distribution, providing an empirical foundation for policy optimization. Methods: A retrospective analysis was conducted using longitudinal data from the China Health Statistics Yearbooks. Trends were delineated via descriptive statistics. Equity and spatial evolution were assessed using the Gini coefficient, Theil index decomposition, and spatial autocorrelation analyses (Global Morans I and hotspot analysis). Results: From 2007 to 2023, the total number of pharmacists increased from 357,700 to 569,500 (average annual growth: 2.2%). This growth lagged behind physicians (4.6%) and nurses (7.4%), causing the pharmacist-to-physician ratio to decline from 1:5.15 to 1:8.39. The workforce showed trends of feminization (female proportion rose from 59.7% to 70.8%) and aging. While quality improved, 51.1% still held an associate degree or below, and only 6.6% held senior titles. Equity analysis revealed the provincial Gini coefficient improved from 0.145 to 0.093. Theil index decomposition confirmed intra-provincial disparities as the primary inequality driver. Spatial analysis showed a non-significant global Morans I by 2023 (0.154, P*>0.05), down from 0.254 (P<0.01) in 2007. Hotspot analysis confirmed this transition, revealing a contraction of high-confidence clusters and a trend toward balanced distribution. Conclusions: China has made measurable progress in expanding pharmacist workforce size and improving inter-provincial equity since 2007. However, persistent structural challenges remain: relative workforce contraction compared to other health professions, an aging demographic, a shortage of senior talent, and significant intra-provincial inequity. Future policies must prioritize optimizing workforce structure and enhancing clinical service capabilities to catalyze a shift toward patient-centered pharmaceutical care.

Objectives: Large language models (LLMs) have shown promise in creating clinical codelists for research purposes, a time-consuming task requiring expert domain knowledge. Here, we evaluate the performance and assess failure modes of a retrieval augmented generation (RAG) approach to creating clinical codelists for the large and complex medical terminology used by the Clinical Practice Research Datalink (CPRD). Materials & Methods: We set up a RAG system using a database of word embeddings of the medical terminology that we created using a general-purpose word embedding model (gemini-embedding). We developed 7 reference codelists presenting different challenges and tagged required and optional codes. We ran 168 evaluations (7 codelists, 2 different database subsets, 4 models, 3 epochs each). Scoring was based on the omission of required codes, and inclusion of irrelevant codes. We used model-grading (i.e., grading by another LLM with the reference codelists provided as context) to evaluate the output codelists (a score of 0% being all incorrect and 100% being all correct). Results: We saw varying accuracy across models and codelists, with Gemini 3 Pro (Score 43%) generally performing better than Claude Sonnet 4.6 (36%), Gemini 3 Flash, and OpenAI GPT 5.2 performing worst (14%). Models performed better with shorter target codelists (e.g., Eosinophilic esophagitis with four codes, and Hidradenitis suppurativa with 14 codes). For example, all models consistently failed to produce a complete Wrist fracture codelist (with 214 required codes). We further present evaluation summaries, and failure mode evaluations produced by parsing LLM chat logs. Discussion: Besides demonstrating that a single-shot RAG approach is currently not suitable for codelist generation, we demonstrate failure modes including hallucinations, retrieval failures and generation failures where retrieved codes are not used. Conclusions: Our findings suggest that while RAG systems using current frontier LLMs may create correct clinical codelists in some cases, they still struggle with large and complex terminologies and codelists with a large number of codes. The failure mode we highlight can inform the creation of future workflows to avoid failures.

Importance: Physicians routinely prognosticate to guide care delivery and shared decision making, particularly when caring for patients with critical illnesses. Yet, these physician estimates are prone to inaccuracy and uncertainty. Artificial intelligence, including large language models (LLMs), show promise in supporting or improving this prognostication. However, the performance of contemporary LLMs in prognosticating for the heterogeneous population of critically ill patients remains poorly understood. Objective: To characterize and compare the performance of LLMs and physicians when predicting 6-month mortality for hospitalized adults who survived critical illness. Design: Embedded mixed methods study with elicitation and comparison of prognostic estimates and reasoning from LLMs and practicing physicians. Setting: The publicly available, deidentified Medical Information Mart for Intensive Care (MIMIC)-IV v2.2 dataset. Participants: We randomly selected 100 hospitalizations of adult survivors of critical illness. Four contemporary LLMs (Open AI GPT-4o, o3- and o4-mini, and DeepSeek-R1) and 7 physicians provided independent prognostic estimates for each case (1,100 total estimates; 400 LLM and 700 physician). Main outcomes and measures: For each case, LLMs and physicians used the hospital discharge summary and demographics to predict 6-month mortality (yes/no) and provide their reasoning (free text). We assessed prognostic performance using accuracy, sensitivity, and specificity, and used inductive, qualitative content analysis to characterize reasonings. Results: Mean physician accuracy for predicting mortality was 70.1% (95% CI 63.7-76.4%), with sensitivity of 59.7% (95% CI 50.6-68.8%) and specificity of 80.6% (95% CI 71.7-88.2%). The top-performing LLM (OpenAI o4-mini) accuracy was 78.0% (95% CI 70.0-86.0%), with sensitivity of 80.0% (95% CI 67.4-90.2%) and specificity of 76.0% (95% CI 63.3-88.0%). The difference between mean physician and top-performing LLM accuracy was not statistically significant (p = 0.5). Qualitative analysis revealed similar patterns in LLM and physician expressed reasoning, except that physicians regularly and explicitly reported uncertainty while LLMs did not. Conclusion and Relevance: In this study, LLMs and physicians achieved comparable, moderate performance in predicting 6-month mortality after critical illness, with similar patterns in expressed reasoning. Our findings suggest LLMs could be used to support prognostication in clinical practice but also raise safety concerns due to the lack of LLM uncertainty expression.

Coronary artery disease is a leading cause of morbidity and mortality. Invasive coronary angiography is currently the gold standard in disease diagnosis. Several studies have attempted to use artificial intelligence (AI) to automate their interpretations with varying levels of success. However, most existing studies cannot generate detailed angiographic reports beyond simple classification or segmentation. This study aims to fine-tune and evaluate the performance of a Vision-Language Model (VLM) in coronary angiogram interpretation and report generation. Using twenty-thousand angiogram keyframes of 1987 patients collated across four unique datasets, we finetuned InternVL2-4B model with Low-Rank Adaptor weights that can perform stenosis detection, anatomy labelling, and report generation. The fine-tuned VLM achieved a precision of 0.56, recall of 0.64, and F1-score of 0.60 for stenosis detection. In anatomy segmentation, it attained a weighted precision of 0.50, recall of 0.43, and F1-score of 0.46, with higher scores in major vessel segments. Report generation integrating multiple angiographic projection views yielded an accuracy of 0.42, negative predictive value of 0.58 and specificity of 0.52. This study demonstrates the potential of using VLM to streamline angiogram interpretation to rapidly provide actionable information to guide management, support care in resource-limited settings, and audit the appropriateness of coronary interventions. AUTHOR SUMMARYCoronary artery disease has heavy disease burden worldwide and coronary angiogram is the gold standard imaging for its diagnosis. Interpreting these complex images and producing clinical reports require significant expertise and time. In this study, we fine-tuned and investigated an open-source VLM, InternVL2-4B, to interpret and report coronary angiogram images in key tasks including stenosis detection, anatomy identification, as well as full report generation. We also referenced the fine-tuned InternVL2-4B against state-of-the-art segmentation model, YOLOv8x, which was evaluated on the same test sets. We examined how machine learning metrics like the intersection over union score may not fully capture the clinical accuracy of model predictions and discussed the limitations of relying solely on these metrics for evaluating clinical AI systems. Although the model has not yet achieved expert-level interpretation, our results demonstrate the potential and feasibility of automating the reporting of coronary angiograms. Such systems could potentially assist cardiologists by improving reporting efficiency, highlightning lesions that may require review, and enabling automated calculations of clinical scores such as the SYNTAX score.

Objective The aim of this study was to develop and validate an automated, scalable framework to harmonise fragmented UK primary care prescription records into a research-ready dataset by mapping four diverse medical ontologies to a unified, historically comprehensive reference standard. Materials and Methods We used raw prescription records for consented participants in the UK Biobank, in which participants are uniquely characterized by multiple data modalities. Primary care data were preprocessed by selecting one drug code if multiple were recorded, cleaning codes to match reference presentations, expanding code granularity based on drug descriptions, and updating outdated codes to a single reference version. Harmonisation entailed mapping British National Formulary (BNF) and Read2 codes to dm+d, the universal NHS standard vocabulary for uniquely identifying and prescribing medicines. Harmonised dm+d records were then homogenised to a single concept granularity, the Virtual Medicinal Product (VMP). We validated our methods by creating medication profiles mapping contemporary drug prescribing patterns in 312 physical and mental health conditions. Results We preprocessed 57,659,844 records (100%) from 221,868 participants (100%). Of those, 48,950 records were dropped due to lack of drug code. 7,357,572 records (13%) used multiple ontologies. Most (76%) records were encoded in BNF and most had the code granularity expanded via the drug description (N=28,034,282; 49%). 41,244,315 records (72%) were harmonised to dm+d and 99.98% of these were converted to VMP as a homogeneous dataset. Across 312 diseases, we identified 23,352 disease-drug associations with 237 medications (represented as BNF subparagraphs) that survived statistical correction of which most resembled drug - indication pairs. Conclusion Our methodology converts highly fragmented and raw prescription records with inconsistent data quality into a streamlined, enriched dataset at a single reference, version, and granularity of information. Harmonised prescription records can be easily utilised by researchers to perform large-scale analyses in research.

Machine learning models have seen a lot of success in medical image segmentation domain. However, one of the challenges that they face are confounders or shortcuts: spurious correlations or biases in the training data that affect the resulting models. One example of such confounders for surgical machine learning is the setup of surgical equipment, including tools and lighting. Using the task of identification of safe and dangerous zones of dissection in laparoscopic cholecystectomy images and videos as a use-case, we inspect two equipment-induced biases: the presence of surgical tools in the field of view and the position of lighting. We propose methods for evaluating the severity of these biases and augmentation-based methods for mitigating them. We show that our tool bias mitigations improve the models' consistency under tool movements by 9 percentage points in the most inconsistent cases, and by 4 percentage points on average. Our lighting bias mitigations help reduce fraction of true dangerous zone pixels that may be predicted as safe under light changes from 5% to 1.5%, without compromising segmentation quality.

Introduction Secondary use of electronic health records (EHRs) often requires transforming raw clinical information into research-grade data. A central step in this process is EHR phenotyping - the identification of patient cohorts defined by specific medical conditions. Although numerous approaches exist, from ICD-based heuristics to supervised learning and large language models (LLMs), the field lacks standardized benchmark datasets, limiting reproducibility and hindering fair comparison across methods. Methods We developed the MIMIC-IV Phenotype Atlas (MIPA) dataset, an adaptation of MIMIC-IV that provides expert-annotated discharge summaries across 16 phenotypes of varying prevalence and complexity. Two independent clinicians reviewed and labeled the discharge summaries, resolving disagreements by consensus. In parallel, we implemented a processing pipeline that extracts multimodal EHR features and generates training, validation, and testing datasets for supervised phenotyping. To illustrate MIPA's utility, we benchmarked four phenotyping methods : ICD-based classifiers, keyword-driven Term Frequency-Inverse Document Frequency (TF-IDF) classifiers, supervised machine learning (ML) models, and LLMs on the task. Results The final MIPA corpus consists of 1,388 expert-annotated discharge summaries. Annotation reliability was high (mean document-level kappa = 0.805, mean label-level kappa = 0.771), with 91% of disagreements resolved through consensus review. MIPA provides high-quality phenotype labels paired with structured EHR features and predefined train/validation/test splits for each phenotype. In the benchmarking case study, LLMs achieved the highest F1 scores in 13 of 16 phenotypes, particularly for conditions requiring contextual interpretation of clinical narrative, while supervised ML offered moderate improvements over rule-based baselines. Conclusion MIPA is the first publicly available benchmark dataset dedicated to EHR phenotyping, combining expert-curated annotations, broad phenotype coverage, and a reproducible processing pipeline. By enabling standardized comparison across ICD-based heuristics, ML models, and LLMs, MIPA provides a durable reference resource to advance methodological development in automated phenotyping.

Abstract Objective To address the unresolved bottleneck of selecting cohort-relevant clinical concepts for treatment trajectory analysis in observational health data, we introduce CohortContrast, an OMOP-compatible R package for enrichment-based concept identification, temporal and semantic noise reduction, and concept aggregation, enabling cohort-level characterization and downstream trajectory analysis. Materials and Methods We developed CohortContrast and applied it to OMOP-mapped observational data from the Estonian nationwide OPTIMA database, which includes all cases of lung, breast, and prostate cancer, focusing here on lung and prostate cancer cohorts. The workflow combines target-control statistical enrichment, temporal/global noise filtering, hierarchical concept aggregation and correlation-based merging, with optional patient clustering for downstream trajectory exploration. We validated the approach with a clinician-based plausibility assessment of extracted diagnosis-concept pairs and evaluated a large language model (LLM) as an auxiliary filtering step. Results We analyzed 7,579 lung cancer and 11,547 prostate cancer patients. The workflow reduced concept dimensionality from 5,793 to 296 concepts (94.9%) in lung cancer and from 5,759 to 170 concepts (97.0%) in prostate cancer, and identified three exploratory patient subgroups in both cohorts. In a plausibility assessment of 466 diagnosis-concept pairs, validators rated 31.3% as directly linked and 57.5% as indirectly linked. Discussion CohortContrast reduces manual concept curation by prioritizing and aggregating cohort-relevant concepts while preserving clinically interpretable treatment patterns in OMOP-based real-world data. Conclusion CohortContrast enables scalable reduction of broad OMOP concept spaces into clinically interpretable, cohort-specific representations for exploratory trajectory analysis and real-world evidence research.

Timely identification of intensive care unit (ICU) patients likely to exit the unit can support anticipatory workflows such as chart review, eligibility screening, and patient outreach prior to transfer. Most ICU discharge prediction studies report discrimination and calibration, but these metrics do not quantify the decision consequences of acting on predictions. Using adult ICU admissions from MIMIC-IV, we represented each ICU stay as a sequence of daily clinical summaries and trained logistic regression, random forest, and XGBoost models to predict next day ICU transfer. Models achieved ROC AUC of 0.80-0.84 with differing calibration. We evaluated decision utility using decision curve analysis (DCA), where positive predictions trigger proactive review. Across thresholds, model guided strategies outperformed review-all, review-none, and a simple clinical rule. To translate net benefit into implementable operations, we modeled a clinical trial recruitment workflow with an 8 hour daily time constraint, incorporating chart review and consent effort. At a feasible operating threshold (0.23), the model flagged [~]23 charts/day and yielded [~]1.23 enrollments/day under conservative eligibility and consent assumptions. These results demonstrate that DCA provides a transparent framework for determining when ICU transfer predictions are worth using and how thresholds should be selected to align with real world workflow constraints. Data and Code AvailabilityThis research has been conducted using data from MIMIC-IV. Researchers can request access via PhysioNet. Implementation code is available upon request.

BackgroundClinical pharmacists, trainees, and educators rely on multi-database literature retrieval and structured evidence synthesis to answer drug-information questions. Existing workflows require navigation across PubMed, DailyMed, LactMed, interaction checkers, and specialty guideline repositories with manual de-duplication, appraisal, and synthesis. Commercial platforms that integrate these functions are costly and often unavailable in community, rural, and international training contexts. ObjectiveThis report describes the architecture of AuditMed, a single-file, browser-based clinical evidence audit platform, and reports preliminary stress-test results against a complex multi-morbidity case corpus. AuditMed is intended for research and educational use and is not a substitute for clinical judgment or validated commercial clinical decision-support systems. MethodsAuditMed integrates nineteen free, publicly available clinical and biomedical application programming interfaces into a six-stage Search [-&gt;] Select [-&gt;] Parse [-&gt;] Analyze [-&gt;] Infer [-&gt;] Create pipeline and supports browser-local patient-case ingestion with regex-based HIPAA Safe Harbor de-identification. Preliminary stress-testing was conducted against eleven cases (Cases 30 through 40) from the Complex Clinical Case Compendium Software Validation Suite, each featuring over twenty concurrent active disease states. For each case, the one-click inference pipeline was executed with default settings and the full Clinical Inference Report was captured verbatim. No retrieval-sensitivity, synthesis-fidelity, or time-to-answer endpoints were pre-specified; the exercise was qualitative and oriented toward pipeline behavior under extreme multi-morbidity. ResultsThe pipeline completed without fatal errors for all eleven cases and produced a structured Clinical Inference Report in each instance. Quantitative-finding detection performed as designed for hematologic parameters and cardiac biomarkers. Two parser defects were identified and are reproduced in the appendix: an age-as-fever regex-precedence defect affecting seven cases and a diagnosis-versus-medication parsing defect affecting one case. Evidence-linkage rate varied from zero evidence-linked statements in seven cases to eleven in one case, reflecting dependence of the inference layer on MeSH-indexed literature coverage of the specific case diagnoses. ConclusionsAuditMed is an early-stage, open-source platform whose value at this stage is in providing a free, transparent, auditable workflow for multi-source evidence synthesis with explicit uncertainty flagging. The preliminary results document both robust end-to-end completion under extreme case complexity and specific, reproducible parser defects that will be addressed before formal evaluation. Planned evaluation studies are described.

Diabetic retinopathy (DR) is a leading cause of vision impairment, requiring accurate and scalable diagnostic tools. Foundation models are increasingly applied to clinical imaging, but concerns remain about their calibration. We evaluated DINOv3, RETFound, and VisionFM for DR classification using different transfer learning strategies in BRSET (n = 16,266) and mBRSET (n = 5,164). Models achieved high discrimination in binary classification (normal vs retinopathy) in BRSET (AUROC 0.90-0.98), with DINOv3 achieving the best under full fine-tuning (AUROC 0.98 [95% CI: 0.97-0.99]). External validation on mBRSET showed decreased performance for all models regardless of the fine-tuning strategy (AUROC 0.70-0.85), though fine-tuning improved performance. Foundation models achieved strong discrimination but poor calibration, generally overestimating DR risk. While the generalist model, DINOv3, benefited from deeper fine-tuning, miscalibration remained evident. These findings underscore the need to improve calibration and the comprehensive evaluation of foundation models, which are essential in clinical settings. Author summaryArtificial intelligence is increasingly being used to detect eye diseases such as diabetic retinopathy from retinal images. Recent advances have introduced "foundation models," which are trained on large datasets and can be adapted to new tasks. We aimed to evaluate how well these models perform in a clinical prediction context, with a focus not only on accuracy but also on how reliably they estimate disease risk. In this study, we compared different types of foundation models using two independent datasets from Brazil. We found that while these models were generally good at distinguishing between healthy and diseased eyes, their predicted risks were often poorly calibrated. In other words, the estimated probabilities did not consistently reflect the true likelihood of disease. We also examined whether adapting the models to the target population could improve performance. Although this approach led to improvements, calibration issues remained. However, post-training correction improved the agreement between predicted risks and observed outcomes. Our findings highlight an important gap between model performance and clinical usefulness. We suggest that improving the reliability of risk estimates is essential before such systems can be safely used in healthcare.

Background Depressive symptoms among reproductive-aged women represent a major public health concern in low- and middle-income countries, yet systematic screening remains limited. In most population survey datasets, the low prevalence of depression results in severe class imbalance, which challenges conventional machine learning models. Therefore, we develop and evaluate a bagging-based ensemble machine learning framework to predict depressive symptoms among reproductive-aged women using highly imbalanced Bangladesh demographic and health survey (BDHS) 2022 data. Methods The sample comprised women aged 15-49 years drawn from BDHS 2022 data. Depressive symptoms were defined using the Patient Health Questionnaire (PHQ-9 [&ge;]10). Candidate predictors were drawn from sociodemographic, reproductive, nutritional, psychosocial, healthcare access, and environmental domains. Feature selection was performed using Elastic Net (EN), Random Forest (RF), and XGBoost model. Five classifiers (EN, RF, Support Vector Machine (SVM), K-nearest neighbors (KNN), and Gradient Boosting Machine (GBM)) were trained using both oversampling-based approaches and the proposed ensemble framework. Model performance was evaluated on an independent test set using accuracy, sensitivity, specificity, F1-score, and the normalized Matthews correlation coefficient (normMCC). Results Approximately 4.8% of women were identified with depressive symptoms. The proposed bagging ensemble framework consistently achieved more balanced predictive performance than oversampling-based models. Average normMCC improved from 0.540 (oversampling) to 0.557 (ensemble). RF and GBM ensembles demonstrated notable improvements in identifying depressive cases, while the EN ensemble achieved the highest overall performance and sensitivity. Threshold optimization yielded stable normMCC across models, indicating robust trade-offs between sensitivity and specificity. Conclusions Bagging-based ensemble learning provides a more robust and balanced approach than synthetic oversampling for predicting depressive symptoms in highly imbalanced population survey data. This approach has important implications for improving early identification and population-level mental health surveillance in resource-constrained settings.

With Japans rapidly aging population, demand for home healthcare is projected to increase by 62% by 2040. This study quantitatively evaluated accessibility to 24-hour home healthcare and regional disparities across all 335 secondary medical areas (SMAs) in Japan using the Enhanced Two-Step Floating Catchment Area (E2SFCA) method. We conducted a nationwide cross-sectional study analyzing approximately 430,000 population points at 500-meter mesh resolution. The E2SFCA integrated demand (age-adjusted population), supply (24-hour home care support clinics and hospitals), and transportation (road networks). Accessibility scores (ASs) and Gini coefficients were calculated for each SMA. Wards hierarchical cluster analysis classified regional types, and multiple regression based on the Penchansky and Thomas five-dimensional access framework identified factors associated with the median AS (ASM) and Gini coefficient. The median ASM was 45.71 (0.00-153.49), and the median Gini coefficient was 0.33 (0.06-0.93). Cluster analysis identified six types ranked by descending ASM, from C1 (high access, equitable; n = 48) to C6 (access desert; n = 23). C6 had a median ASM of 0.00 and Gini coefficient of 0.74, indicating virtually no access within a 30-minute catchment. Home-visit standardized claim ratios, used as external validation, declined monotonically from C1 (125.6) to C6 (17.6). For ASM, 24-hour visiting nursing stations ({beta} = +0.369) and clinic physicians ({beta} = +0.342) showed the strongest positive associations, with non-residential area negatively associated ({beta} = -0.273). For the Gini coefficient, non-residential area showed the strongest positive association ({beta} = +0.523). Taxable income per taxpayer was not significantly associated with either outcome. Non-residential area was associated with both lower accessibility and greater intra-regional inequality, suggesting that geographic constraints may limit the effectiveness of resource investment alone. Uniform nationwide implementation of policies shifting care from long-term care beds to home healthcare may not be feasible; region-specific approaches considering geographic characteristics are necessary.

MedSafe-Dx (v0), introduces a new safety-focused benchmark for evaluating large language models in clinical diagnostic decision support using a filtered subset of the DDx Plus dataset (N=250). MedSafe-Dx evaluates three dimensions: escalation sensitivity, avoidance of false reassurance, and calibration of uncertainty. Models were tasked with providing a ranked differential (ICD-10), an escalation decision (Urgent vs. Routine), and a confidence flag. Performance was measured via a "Safety Pass Rate," a composite metric penalizing three hard failure modes: missed escalations of life-threatening conditions, overconfident incorrect diagnoses, and unsafe reassurance in ambiguous cases. Eleven models were evaluated and revealed a significant disconnect between diagnostic recall and safety. GPT-5.2 achieved the highest Safety Pass Rate (97.6%), while several models exhibited high rates of missed escalations or unsafe reassurance. MedSafe-Dx provides a robust stress test for identifying high-risk failure modes in diagnostic decision support and shows that high diagnostic accuracy does not guarantee clinical safety. While the benchmark is currently limited by synthetic data and proxy labels, it provides a reproducible, auditable framework for testing AI behavior before clinical deployment. Our findings suggest that interventions such as safety-focused prompting and reasoning-token budgets could be essential components for the safe deployment of LLMs in clinical workflows.

Social determinants of health (SDoH) are important for clinical care, but it remains unclear how much AI-captured social context is preserved after clinician editing in ambient documentation workflows. We retrospectively analyzed 75,133 paired ambient AI-drafted and clinician-finalized note sections from ambulatory care at a large academic health system. Using a rule-based NLP pipeline, we extracted 21 SDoH categories and quantified retention, deletion, and addition. SDoH appeared in 25.2% of AI drafts versus 17.2% of final notes. At the mention level, AI captured 29,991 SDoH mentions, of which 45.1% were deleted, 54.9% were retained with clinicians adding 3,583 new mentions. Insurance and marital status were most often deleted, whereas substance use and physical activity were more often retained. Deletion patterns also varied by specialty, supporting the need for specialty-aware ambient AI systems.

Multimodal machine learning leverages complementary information from diverse data sources and has shown strong promise in medical imaging, where multimodal data is critical for clinical decision making. In glioma grading, integrating MRI modalities with clinical data can improve diagnostic accuracy, yet systematic comparisons of fusion strategies remain limited. This study evaluates early, intermediate, and late fusion approaches, addressing the question: How does the inclusion of clinical data alongside MRI modalities influence grading performance? To assess modality contributions, we design adaptable fusion layers and employ interpretability techniques, including attention-based analysis. Our results show that incorporating clinical data consistently outperforms unimodal and MRI-only baselines, with intermediate fusion yielding the most reliable gains. Beyond accuracy, the framework reveals how MRI and clinical features jointly shape predictions, underscoring the importance of both fusion design and interpretability for clinical adoption.

Background: The increasing availability of routinely collected health data offers new opportunities for population-level research, yet access to comprehensive, linked, and standardised datasets remains limited. We describe EST-Health-30, a large-scale, population-representative health data resource from Estonia. Methods: EST-Health-30 comprises a random 30% sample of the Estonian population (~500,000 individuals), with longitudinal data from 2012 to 2024 and annual updates planned through 2026. Individual-level records are linked across five nationwide databases, including electronic health records, health insurance claims, prescription data, cancer registry, and cause of death records. A privacy-preserving hashing approach ensures consistent cohort inclusion over time while maintaining pseudonymisation. All data are harmonised to the Observational Medical Outcomes Partnership (OMOP) Common Data Model (version 5.4) using international standard vocabularies. Data quality was assessed using established OMOP-based validation frameworks. Results: The dataset contains rich multimodal information on diagnoses, procedures, laboratory measurements, prescriptions, free-text clinical notes, healthcare utilisation, and costs, with high population coverage and longitudinal depth. Data quality assessment showed high completeness and consistency, with 99.2% of applicable checks passing. The age-sex distribution closely reflects the national population, supporting representativeness, though coverage is marginally below the target 30% (29.2%), primarily attributable to recent immigrants without health system contact. The dataset enables construction of detailed clinical cohorts, analysis of disease trajectories, and evaluation of healthcare utilisation and outcomes across the life course. Conclusions: EST-Health-30 is a comprehensive, standardised, and population-representative real-world data resource that supports epidemiological, clinical, and methodological research. Its alignment with the OMOP CDM facilitates reproducible analytics and participation in international federated research networks, while secure access infrastructure ensures compliance with data protection regulations.

Abstract: Background: Vascular access is essential in treating patients undergoing prolonged endovenous therapy such as chemotherapy, antibiotics, and parenteral nutrition. Since the 1990s, when PICCs (peripherally inserted central catheters) appeared, vascular access options have expanded significantly, revolutionizing the treatment landscape for all types of patients. Objective: To analyze and describe the profile of the use of PICCs in a Brazilian quaternary hospital over 10 years with data collected by the infusion therapy team. Evaluating the number of PICCs implanted over the years, patients epidemiology and clinical characteristics, insertion details, associated complications, and the reason for removal. Methods: A retrospective cohort study that employs a quantitative, non-experimental approach to classify and statistically analyze past events associated with 21,652 PICCs implanted from January 2012 to December 2021 in a quaternary hospital at Sao Paulo - Brazil. All the catheters were implanted, and the data was collected by a team of nurses specializing in infusion therapy. We analyzed the number of catheters implanted over the years, insertion characteristics, patients epidemiology and clinical data, possible associated complications, and the reason for removal. Statistical analyses were conducted using R software (version 4.4.1) and SPSS (version 29) for Windows (IBM Corp, Armonk, NY). Results: During the specified period, 21,652 catheters were analyzed. The patients gender distribution was nearly balanced (48.2% versus 51.8%), and the average age was 66 years. Cardiovascular and metabolic issues were the most common comorbidities, and between 2020 and 2021, 29.3% of the sample tested positive for COVID-19. The most common location of hospitalization and implantation was the medical-surgical clinic (31.6% - 41.4%), and the most used type of catheter was the Power Picc (83.9%). The estimated complication incidence density is 2.94 complications per 1,000 catheter-days. Almost all the PICCs (98,2%) were adequately located at the cavo-atrial junction after the first attempt, 82.2% of catheters were removed after therapy, and the median duration of catheter use was 12 days. Conclusion: PICCs are widely employed for drug infusion, with their use growing progressively due to specialized teams greater availability and training. The high efficiency of these devices with a relatively low risk of complications already observed in previous studies was reinforced by the findings of this study of more than 20,000 catheters.

Ambulatory electrocardiograms (ECG) provides continuous monitoring of the hearts electrical activity. However, many existing machine learning and artificial intelligence models for analyzing ambulatory ECG traces are often unimodal and do not incorporate patient clinical context. In this study, we propose a multimodal framework integrating ambulatory ECG-derived representations with clinical text embeddings to predict two cardiac outcomes: sudden cardiac death and pump failure death. Ambulatory ECG traces are preprocessed, segmented, and encoded via a multiple instance learning and temporal convolutional neural network framework. In parallel, patient clinical features are parsed into structured prompts, which are passed through a large language model to generate clinical reasoning; this reasoning passes through a biomedical language encoder to generate a text embedding. With the ECG and text embeddings, we systematically evaluate multiple fusion strategies, including concatenation- and gating-based approaches, to integrate these two data modalities. Our results demonstrate that multimodal models consistently outperform unimodal baselines, with adaptive fusion mechanisms providing the greatest improvements in predictive performance. Decision curve analysis highlights the potential clinical utility of the proposed framework for risk stratification. Finally, we visualize model attention across modalities, including ECG attention patterns, segment-level saliency, heart rate variability features, and clinical reasoning, to contextualize patient-specific predictions.