256 papers
Health informatics sits at the vibrant intersection of medicine, data science, and technology, transforming how we store, analyze, and utilize health information. This rapidly evolving field empowers clinicians and researchers to uncover patterns in patient data, improve diagnostic accuracy, and personalize treatment plans without getting lost in complex databases. By turning raw medical records into actionable insights, these innovations are reshaping the future of healthcare delivery and population health management.
At Gist.Science, we bridge the gap between cutting-edge research and public understanding by curating the latest preprints from medRxiv specifically within this domain. Our team processes every new submission in this category, providing both accessible plain-language explanations and detailed technical summaries to ensure the science is clear for everyone, from policymakers to curious readers. Below are the latest papers in health informatics, freshly distilled and ready for you to explore.
This paper demonstrates that vision-language models exhibit inattentional blindness similar to humans but also display a unique confabulation failure mode, necessitating signal-detection analysis with matched-control baselines to accurately assess their visual perception capabilities.
This study introduces the Chemical-Gene Atlas (CGA), an interactive resource integrating 4,110 chemicals with 928 developmental lethality-associated genes to map exposure-specific vulnerabilities across gestational windows and identify key gene-environment interactions underlying recurrent pregnancy loss.
This manuscript, which proposed an explainable deep learning model for breast cancer diagnosis using 2D grayscale ultrasound, has been withdrawn by the authors due to conflicts of interest and a co-author's disagreement with the methodology.
This paper introduces SPIRIT-CONSORT-ELM, a novel framework and dataset that extends existing reporting guidelines to the element level, utilizing a hybrid pipeline of PubMedBERT and generative large language models to automatically assess the completeness and transparency of randomized controlled trial reports with high accuracy.
This cross-sectional study of healthcare professionals in Freetown, Sierra Leone, reveals a low prevalence of private health insurance coverage despite high enrollment willingness, with no significant demographic or socioeconomic factors identified as determinants, highlighting the need to address unexamined barriers like cost and accessibility to advance Universal Health Coverage in the region.
This study validates the Empatica E4 wristband's utility for sleep research by analyzing the Wearanize+ dataset, confirming its accuracy in tracking heart rate and skin temperature across sleep stages while identifying specific limitations in electrodermal activity measurements due to wrist sweat accumulation.
This paper proposes an explainable multimodal AI framework that integrates supervised deep learning with conservative offline reinforcement learning to support post-surgical clinical decisions, while simultaneously critiquing synthetic data practices and validating the system's architecture through a rigorous real-data methodology.
This registered report presents an artifact index for capacitive ECG signals acquired from an instrumented armchair, which utilizes a voting approach of three signal quality indices to effectively distinguish clean from artifact segments during reading and TV-watching activities, thereby enabling reliable continuous health monitoring in unsupervised real-world settings.
The paper introduces PolyCLIP-T, a topology-guided multimodal framework that leverages whole-genome sequencing and persistent homology to identify stable clusters of polygenic variants across familial multimorbid disorders, effectively overcoming the limitations of traditional rule-based pipelines in detecting non-coding, structural, and low-penetrance genetic drivers.
This study evaluates three photoplethysmography-derived microvascular reactivity signals in the MIMIC-IV-WDB v0.1.0 dataset using human and AI-assisted visual inspection, finding that two signals failed to capture their intended physiology in most cases and the third was limited by sensor placement, thereby highlighting the necessity of preliminary raw-data validation before downstream modeling.