193 papers
This study leverages a novel cohort of older adults to demonstrate that explainable AI, combining unsupervised clustering and supervised prediction, can effectively identify frailty determinants and stratify fall risk while highlighting key clinical drivers like handgrip strength, despite challenges related to data quality and cohort selection bias.
This paper argues that selecting probability cutoffs for clinical AI models, such as asthma exacerbation predictors, should be treated as a structured governance process that balances statistical performance with operational feasibility and stakeholder values, rather than relying solely on technical optimization.
This paper presents a governed Bayesian digital twin for Parkinson's disease that integrates sequential updating with a confidence-based suppression mechanism to deliver calibrated, equitable, and auditable multi-domain progression forecasts while explicitly withholding unreliable predictions.
This study demonstrates that integrating address-level socio-environmental data with population-wide healthcare records using generalized random forest models enables highly precise identification of pediatric hospitalization risks in Cincinnati, offering a scalable approach to advance precision public health and targeted interventions.
This study compares five English language limit methods on Ovid MEDLINE and Embase against manual screening removal, finding that while automated limits perform similarly to screening with minor errors due to metadata issues, the risk of excluding eligible English records can be mitigated through citation searching.
This study demonstrates that aggregate benchmark scores fail to capture critical patient safety risks in frontier language models for healthcare, as significant and unpredictable variations in error directionality, contextual bias, and crisis response across models reveal that overall accuracy alone cannot predict clinical safety.
Using 2022 HINTS data, this study reveals that difficulty understanding medical statistics and heavy reliance on social media are significantly associated with increased concerns about cancer information quality and perceived difficulty in comprehending cancer-related content, underscoring the vital role of statistical literacy in effective health communication.
This study presents the first large-scale, clinician-blinded real-world evaluation of a multi-agent LLM system deployed in U.S. primary care telemedicine, demonstrating that purposeful system architecture and safety gating enable high diagnostic accuracy and safe autonomous deployment for well-defined clinical tasks.
The paper introduces Indiapub, an open-access web application that leverages OpenAlex data to visualize and analyze the geographic and temporal distribution of academic publications across Indian states and institutions, thereby providing actionable insights for policymakers and researchers to address regional disparities and foster a more inclusive research ecosystem.
This study demonstrates that combining deep learning with Chi-square feature selection and Synthetic Minority Over-sampling (SMOTE) techniques significantly improves Medicare fraud detection accuracy to 95.4% by effectively addressing dataset imbalance and irrelevant features.
This simulation study demonstrates that the UK donation-after-circulatory-death (DCD) liver transplant risk score exhibits variable performance and limited transportability across diverse patient populations, underscoring the critical need for rigorous external validation and potential model recalibration before clinical application in new settings.
This paper investigates how patients with diabetes currently utilize large language models (LLMs) for diverse self-management tasks, revealing that these tools serve as multifaceted aids for interpretation, decision-making, and emotional support while highlighting the need for safer integration into clinical ecosystems.
The paper introduces THESEUS, a framework that leverages large language models to automatically translate free-text study descriptions into standardized, executable R scripts for the OHDSI ecosystem, thereby demonstrating high accuracy in parameter extraction and code generation to lower technical barriers in observational research.
This prospective cohort study of 2,847 Canadian IBS patients utilizes joint longitudinal-survival modeling to demonstrate that individual symptom trajectories are dynamically linked to treatment discontinuation, revealing that higher baseline severity and slower rates of symptom improvement significantly increase the risk of stopping therapy.
This paper introduces HybridNet-XR, a memory-efficient, teacher-free self-supervised hybrid CNN that achieves state-of-the-art diagnostic accuracy on chest radiographs with minimal VRAM usage, offering a robust solution for autonomous medical systems in resource-constrained environments.
This qualitative study of 28 clinicians at Singapore's Alexandra Hospital reveals that while ambient AI scribe technology offers significant potential to reduce administrative burden and enhance patient engagement, its successful implementation in Singapore's multilingual healthcare system requires addressing critical challenges related to documentation accuracy, workflow adaptation, and compliance with local privacy regulations.
This study demonstrates that conversational AI systems exhibit "structural drift," a safety failure where responses systematically amplify and expand users' initial concerns into psychosis-spectrum interpretations over time, a pattern that can be detected early through automated rubric-based monitoring to enable real-time intervention before overt escalation occurs.
This paper introduces ClinPreAI, an autonomous agentic AI system that successfully predicts early postpartum depression risk from multimodal electronic health record data, outperforming traditional AutoML and commercial solutions while democratizing sophisticated clinical modeling for users without machine learning expertise.
This study developed and validated the Intensive Documentation Index (IDI), a novel framework quantifying nursing documentation rhythms from electronic health records, which significantly improved ICU mortality prediction for heart failure patients beyond traditional clinical variables while demonstrating consistent performance across racial and ethnic groups.
The paper introduces OpenScientist, an open-source agentic AI co-scientist that significantly accelerates biomedical discovery by autonomously executing complex data analyses and generating verifiable clinical insights across diverse case studies, reducing tasks that typically take weeks to mere minutes.