193 papers
This study demonstrates that while zero-shot sentiment analysis of clinical notes yields only a weak correlation with length of stay, directly prompting large language models to estimate hospitalization duration provides a significantly stronger predictive capability, suggesting future systems should prioritize direct outcome extraction over sentiment analysis.
This research proposes a reproducible, lightweight, and interpretable runtime anomaly detection framework for AI-driven Nurse Call Systems that utilizes an Isolation Forest model to ensure operational safety and reliability in resource-constrained medical environments without relying on complex deep learning models.
This pre-registered falsification study across four clinical datasets demonstrates that while sepsis prediction models at elite academic centers primarily detect genuine biological signals rather than care-process intensity, they reveal a systematic and consequential divergence between clinical sepsis definitions and administrative coding that undermines the validity of regulatory metrics and AI benchmarks relying on the latter.
This study evaluates the performance of 11 different filters for removing animal and nonhuman records in Ovid Embase by analyzing 3,000 records, revealing significant variations in sensitivity and specificity among the methods and highlighting the need for search teams to carefully select filters based on their specific research context.
This study demonstrates that a zero-shot, pairwise comparison approach using a large language model and Bradley-Terry aggregation for emergency department triage significantly outperforms the standard Emergency Severity Index and maintains cross-site stability without requiring site-specific training, matching the performance of supervised models on external data.
The authors developed an AI-driven pipeline that leverages multimodal EHR data and an ensemble meta-learner with a dual-threshold confidence framework to automate the abstraction of Society of Thoracic Surgeons Adult Cardiac Surgery Database variables, achieving over 99% accuracy while significantly reducing the manual data collection burden across two healthcare systems.
This national fairness audit of 386,775 US adults demonstrates that enforcing demographic parity in HIV testing prediction models is inappropriate for differential-burden clinical contexts, as it significantly reduces screening access for high-risk populations and underscores the need for fairness metrics like equalized odds and calibration that align with clinical needs.
This paper, titled "Determinants of Digital Health Technology Acceptance Among Healthcare Caregivers: A Structural Equation Modeling Approach," has been withdrawn from medRxiv due to the submission of false information.
This paper is a withdrawn study from medRxiv that claimed to analyze the causal effects of natural language processing-enhanced clinical decision support on early cognitive impairment detection, but was retracted because it was submitted with false information.
This withdrawn study utilized Stochastic Frontier Analysis with Bayesian Model Averaging on a panel of 247 South Korean hospital networks to demonstrate that blockchain-enabled health information exchange systems significantly improve technical efficiency compared to conventional platforms, even after controlling for endogeneity and model uncertainty.
This paper, which originally aimed to analyze the causal impact of NLP-driven clinical decision support on sepsis mortality in England, has been withdrawn from medRxiv due to the submission of false information.
This paper has been withdrawn from medRxiv because it was submitted with false information, rendering its original findings on NLP-augmented clinical decision support invalid.
The paper introduces RLSYN+REG, a reinforcement learning-driven generative model that significantly enhances the scientific utility of synthetic biomedical data by ensuring regression models trained on it accurately reproduce the coefficients and predictions of models trained on real data, while maintaining high fidelity and privacy.
This study demonstrates that novel mathematical functional features of spatial proximity and circularity derived from digital clock drawing tests can predict cognitive impairment with accuracy comparable to traditional summary statistics, offering a promising approach to enhance early detection strategies.
This paper establishes Ollivier–Ricci curvature as a quantitative geometric biomarker for biomedical networks by demonstrating its ability to distinguish structural phases in medical ontologies, map monotonic aging trajectories in disease comorbidity, and complement sedenion-based features in differentiating brain disorders, all while validating core mathematical claims through machine verification.
This study demonstrates that both logistic regression and gradient boosting models achieve near-perfect accuracy in detecting national influenza outbreaks using U.S. CDC ILINet surveillance data from 2010 to 2025, validating the operational utility of framing early-warning as a threshold-based binary classification problem.
This study analyzed over 410,000 Reddit posts to reveal that while gastrointestinal symptoms are the most common self-reported side effects of semaglutide and tirzepatide, online communities also highlight emerging concerns like reproductive and temperature-related issues that are often underrepresented in clinical trials and official labeling.
This paper introduces a modular, adapter-based framework that decouples reasoning supervision from reward tuning to stabilize the alignment of small clinical language models, achieving robust, auditable, and accurate performance without sacrificing privacy or efficiency.
This paper presents the Orphanet Nomenclature and Classification system as a comprehensive, multilingual, and interoperable standard for rare diseases, detailing its July 2025 update which includes nearly 10,000 clinical entities with extensive mappings to major medical terminologies to facilitate global data sharing, accurate patient identification, and improved healthcare outcomes.
The GLOW UP study is a prospective observational case-control trial in Switzerland designed to evaluate the feasibility of using wearable and smartphone-derived digital biomarkers to screen for prediabetes and characterize lifestyle-metabolic relationships in adults aged 45 and older with elevated BMI.