Hasil untuk "Medical emergencies. Critical care. Intensive care. First aid"

E. Oh, T. Fong, Tammy T. Hshieh et al.

J. Brierley, J. Carcillo, K. Choong et al.

Dongshen Peng, Yi Wang, Austin Schoeffler et al.

Large language models (LLMs) show promise in clinical decision support yet risk acquiescing to patient pressure for inappropriate care. We introduce SycoEval-EM, a multi-agent simulation framework evaluating LLM robustness through adversarial patient persuasion in emergency medicine. Across 20 LLMs and 1,875 encounters spanning three Choosing Wisely scenarios, acquiescence rates ranged from 0-100\%. Models showed higher vulnerability to imaging requests (38.8\%) than opioid prescriptions (25.0\%), with model capability poorly predicting robustness. All persuasion tactics proved equally effective (30.0-36.0\%), indicating general susceptibility rather than tactic-specific weakness. Our findings demonstrate that static benchmarks inadequately predict safety under social pressure, necessitating multi-turn adversarial testing for clinical AI certification.

Jiaqi Shi, Xulong Zhang, Xiaoyang Qu et al.

Recent advances in Vision-Language-Action (VLA) models have shown promise for robot control, but their dependence on action supervision limits scalability and generalization. To address this challenge, we introduce CARE, a novel framework designed to train VLA models for robotic task execution. Unlike existing methods that depend on action annotations during pretraining, CARE eliminates the need for explicit action labels by leveraging only video-text pairs. These weakly aligned data sources enable the model to learn continuous latent action representations through a newly designed multi-task pretraining objective. During fine-tuning, a small set of labeled data is used to train the action head for control. Experimental results across various simulation tasks demonstrate CARE's superior success rate, semantic interpretability, and ability to avoid shortcut learning. These results underscore CARE's scalability, interpretability, and effectiveness in robotic control with weak supervision.

Nikolaos Dionelis, Jente Bosmans, Nicolas Longépé

Performing accurate confidence quantification and assessment in pixel-wise regression tasks, which are downstream applications of AI Foundation Models for Earth Observation (EO), is important for deep neural networks to predict their failures, improve their performance and enhance their capabilities in real-world applications, for their practical deployment. For pixel-wise regression tasks, specifically utilizing remote sensing data from satellite imagery in EO Foundation Models, confidence quantification is a critical challenge. The focus of this research work is on developing a Foundation Model using EO satellite data that computes and assigns a confidence metric alongside regression outputs to improve the reliability and interpretability of predictions generated by deep neural networks. To this end, we develop, train and evaluate the proposed Confidence-Aware Regression Estimation (CARE) Foundation Model. Our model CARE computes and assigns confidence to regression results as downstream tasks of a Foundation Model for EO data, and performs a confidence-aware self-corrective learning method for the low-confidence regions. We evaluate the model CARE, and experimental results on multi-spectral data from the Copernicus Sentinel-2 satellite constellation to estimate the building density (i.e. monitoring urban growth), show that the proposed method can be successfully applied to important regression problems in EO and remote sensing. We also show that our model CARE outperforms other baseline methods.

Juncheng Dong, Yiling Liu, Ahmed Aloui et al.

Large language models (LLMs) have recently demonstrated impressive capabilities across a range of reasoning and generation tasks. However, research studies have shown that LLMs lack the ability to identify causal relationships, a fundamental cornerstone of human intelligence. We first conduct an exploratory investigation of LLMs' behavior when asked to perform a causal-discovery task and find that they mostly rely on the semantic meaning of variable names, ignoring the observation data. This is unsurprising, given that LLMs were never trained to process structural datasets. To first tackle this challenge, we prompt the LLMs with the outputs of established causal discovery algorithms designed for observational datasets. These algorithm outputs effectively serve as the sufficient statistics of the observation data. However, quite surprisingly, we find that prompting the LLMs with these sufficient statistics decreases the LLMs' performance in causal discovery. To address this current limitation, we propose CARE, a framework that enhances LLMs' causal-reasoning ability by teaching them to effectively utilize the outputs of established causal-discovery algorithms through supervised fine-tuning. Experimental results show that a finetuned Qwen2.5-1.5B model produced by CARE significantly outperforms both traditional causal-discovery algorithms and state-of-the-art LLMs with over a thousand times more parameters, demonstrating effective utilization of its own knowledge and the external algorithmic clues.

Jie Zhu, Yuanchen Zhou, Shuo Jiang et al.

Emotional Support Conversation (ESC) plays a vital role in alleviating psychological stress and providing emotional value through dialogue. While recent studies have largely focused on data augmentation and synthetic corpus construction, they often overlook the deeper cognitive reasoning processes that underpin effective emotional support. To address this gap, we propose \textbf{CARE}, a novel framework that strengthens reasoning in ESC without relying on large-scale synthetic data. CARE leverages the original ESC training set to guide models in generating logically coherent and supportive responses, thereby explicitly enhancing cognitive reasoning. Building on this foundation, we further employ reinforcement learning to refine and reinforce the reasoning process. Experimental results demonstrate that CARE significantly improves both the logical soundness and supportive quality of responses, advancing the development of empathetic, cognitively robust, and human-like emotional support systems.

Jean C. Nieto-Merino, Lucero S. Pérez-Gómez

Cristiana Olaru, Sam Langberg, Nicole Streiff McCoin

Increased intracranial pressure (ICP) is encountered in numerous traumatic and non-traumatic medical situations, and it requires immediate recognition and attention. Clinically, ICP typically presents with a headache that is most severe in the morning, aggravated by Valsalva-like maneuvers, and associated with nausea or vomiting. Papilledema is a well-recognized sign of increased ICP; however, emergency physicians often find it difficult to visualize the optic disc using ophthalmoscopy or to accurately interpret digital fundus photographs when using a non-mydriatic retinal camera. Emergency ultrasound can evaluate the optic nerve sheath diameter (ONSD) and optic disc elevation to determine whether increased ICP is present, however, the studies have been small with different definitions and measurements of the ONSD. The ONSD threshold values for increased ICP have been reported anywhere from 4.8 to 6.3 millimeters. Neuroimaging is the next step in the evaluation of patients with papilledema or high clinical suspicion of increased ICP, as it can identify most structural causes or typical radiological patterns of increased ICP. Neuroradiographic signs of increased ICP can be helpful in suggesting idiopathic intracranial hypertension (IIH), especially when papilledema is absent. Patients with papilledema and normal neuroimaging may undergo lumbar puncture as part of their clinical workup. The cerebrospinal fluid (CSF) opening pressure remains one of the most important investigations to establish the diagnosis of IIH. A CSF evaluation is also required to exclude other etiologies of elevated ICP such as infectious, inflammatory, and neoplastic meningitis. Invasive ICP measurement remains the standard to measure and monitor this condition.

Haruka Taira, Masahiro Kashiura, Takashi Moriya

Uriya A. First, Mathieu Florence, Zev Rosengarten

Let $k$ be a field and let $G$ be an affine algebraic group over $k$. Call a $G$-torsor weakly versal for a class of $k$-schemes $\cal C$ if it specializes to every $G$-torsor over a scheme in $\cal C$. A recent result of the first author, Reichstein and Williams says that for any $d\geq 0$, there exists a $G$-torsor over a finite type $k$-scheme that is weakly versal for finite type affine $k$-schemes of dimension at most $d$. The first author also observed that if $G$ is unipotent, then $G$ admits a torsor over a finite type $k$-scheme that is weakly versal for all affine $k$-schemes, and that the converse holds if $\operatorname{char} k=0$. In this work, we extend this to all fields, showing that $G$ is unipotent if and only if it admits a $G$-torsor over a quasi-compact base that is weakly versal for all finite type regular affine $k$-schemes. Our proof is characteristic-free and it also gives rise to a quantitative statement: If $G$ is a non-unipotent subgroup of $\mathbf{GL}_n$, then a $G$-torsor over a quasi-projective $k$-scheme of dimension $d$ is not weakly versal for finite type regular affine $k$-schemes of dimension $n(d+1)+2$. This means in particular that every such $G$ admits a nontrivial torsor over a regular affine $(n+2)$-dimensional variety. When $G$ contains a nontrivial torus, we show that nontrivial torsors already exist over $3$-dimensional smooth affine varieties (even when $G$ is special), and this is optimal in general. In the course of the proof, we show that for every $m,\ell\in\mathbb{N}\cup\{0\}$ with $\ell\neq 1$, there exists a smooth affine $k$-scheme $X$ carrying an $\ell$-torsion line bundle that cannot be generated by $m$ global sections. We moreover study the minimal possible dimension of such an $X$ and show that it is $m$, $m+1$ or $m+2$.

Georgina Cosma, Mohit Kumar Singh, Patrick Waterson et al.

This study applies Natural Language Processing techniques, including Latent Dirichlet Allocation, to analyse anonymised maternity incident investigation reports from the Healthcare Safety Investigation Branch. The reports underwent preprocessing, annotation using the Safety Intelligence Research taxonomy, and topic modelling to uncover prevalent topics and detect differences in maternity care across ethnic groups. A combination of offline and online methods was utilised to ensure data protection whilst enabling advanced analysis, with offline processing for sensitive data and online processing for non-sensitive data using the `Claude 3 Opus' language model. Interactive topic analysis and semantic network visualisation were employed to extract and display thematic topics and visualise semantic relationships among keywords. The analysis revealed disparities in care among different ethnic groups, with distinct focus areas for the Black, Asian, and White British ethnic groups. The study demonstrates the effectiveness of topic modelling and NLP techniques in analysing maternity incident investigation reports and highlighting disparities in care. The findings emphasise the crucial role of advanced data analysis in improving maternity care quality and equity.

Willemijn Klaassen, Bram van Dijk, Marco Spruit

Artificially intelligent systems optimized for speech conversation are appearing at a fast pace. Such models are interesting from a healthcare perspective, as these voice-controlled assistants may support the elderly and enable remote health monitoring. The bottleneck for efficacy, however, is how well these devices work in practice and how the elderly experience them, but research on this topic is scant. We review elderly use of voice-controlled AI and highlight various user- and technology-centered issues, that need to be considered before effective speech-controlled AI for elderly care can be realized.

Ludovica Golino, Michela Saracco, Marco Caiazzo et al.

Major trauma is bound to be managed in highly specialized centers. Due to logistics needs or due to an initial clinical stabilization, these patients happen to be managed in hospitals that are not fully equipped for trauma. We handled a patient, major trauma to dynamics following a high-speed collision between two cars in which she was behind the wheel. The patient was also complex due to cardiovascular and respiratory comorbidities. After ‘ABCDE’ and radiological evaluation, the patient was managed in our hospital with ICU recovery and multiple orthopedic interventions to which she was subjected with neuraxial and peripheral regional anesthesia. The patient was managed successfully and with excellent pain control. The risks associated with her comorbidities were limited to a minimum and she was discharged for rehabilitation 5 days after the operations.

Bart J. Verhoef, Xixi Lu

Previous studies have used prescriptive process monitoring to find actionable policies in business processes and conducted case studies in similar domains, such as the loan application process and the traffic fine process. However, care processes tend to be more dynamic and complex. For example, at any stage of a care process, a multitude of actions is possible. In this paper, we follow the reinforcement approach and train a Markov decision process using event data from a care process. The goal was to find optimal policies for staff members when clients are displaying any type of aggressive behavior. We used the reinforcement learning algorithms Q-learning and SARSA to find optimal policies. Results showed that the policies derived from these algorithms are similar to the most frequent actions currently used but provide the staff members with a few more options in certain situations.

Yuanfang Ren, Tyler J. Loftus, Ziyuan Guan et al.

In the United States, more than 5 million patients are admitted annually to ICUs, with ICU mortality of 10%-29% and costs over $82 billion. Acute brain dysfunction status, delirium, is often underdiagnosed or undervalued. This study's objective was to develop automated computable phenotypes for acute brain dysfunction states and describe transitions among brain dysfunction states to illustrate the clinical trajectories of ICU patients. We created two single-center, longitudinal EHR datasets for 48,817 adult patients admitted to an ICU at UFH Gainesville (GNV) and Jacksonville (JAX). We developed algorithms to quantify acute brain dysfunction status including coma, delirium, normal, or death at 12-hour intervals of each ICU admission and to identify acute brain dysfunction phenotypes using continuous acute brain dysfunction status and k-means clustering approach. There were 49,770 admissions for 37,835 patients in UFH GNV dataset and 18,472 admissions for 10,982 patients in UFH JAX dataset. In total, 18% of patients had coma as the worst brain dysfunction status; every 12 hours, around 4%-7% would transit to delirium, 22%-25% would recover, 3%-4% would expire, and 67%-68% would remain in a coma in the ICU. Additionally, 7% of patients had delirium as the worst brain dysfunction status; around 6%-7% would transit to coma, 40%-42% would be no delirium, 1% would expire, and 51%-52% would remain delirium in the ICU. There were three phenotypes: persistent coma/delirium, persistently normal, and transition from coma/delirium to normal almost exclusively in first 48 hours after ICU admission. We developed phenotyping scoring algorithms that determined acute brain dysfunction status every 12 hours while admitted to the ICU. This approach may be useful in developing prognostic and decision-support tools to aid patients and clinicians in decision-making on resource use and escalation of care.

Zihao Zhao, Yuxiao Liu, Han Wu et al.

Contrastive Language-Image Pre-training (CLIP), a simple yet effective pre-training paradigm, successfully introduces text supervision to vision models. It has shown promising results across various tasks due to its generalizability and interpretability. The use of CLIP has recently gained increasing interest in the medical imaging domain, serving as a pre-training paradigm for image-text alignment, or a critical component in diverse clinical tasks. With the aim of facilitating a deeper understanding of this promising direction, this survey offers an in-depth exploration of the CLIP within the domain of medical imaging, regarding both refined CLIP pre-training and CLIP-driven applications. In this paper, we (1) first start with a brief introduction to the fundamentals of CLIP methodology; (2) then investigate the adaptation of CLIP pre-training in the medical imaging domain, focusing on how to optimize CLIP given characteristics of medical images and reports; (3) further explore practical utilization of CLIP pre-trained models in various tasks, including classification, dense prediction, and cross-modal tasks; and (4) finally discuss existing limitations of CLIP in the context of medical imaging, and propose forward-looking directions to address the demands of medical imaging domain. Studies featuring technical and practical value are both investigated. We expect this survey will provide researchers with a holistic understanding of the CLIP paradigm and its potential implications. The project page of this survey can also be found on https://github.com/zhaozh10/Awesome-CLIP-in-Medical-Imaging.

W. Chao, C. Tseng, Chieh-Liang Wu et al.

Background High glycemic variability (GV) is common in critically ill patients; however, the prevalence and mortality association with early GV in patients with sepsis remains unclear. Methods This retrospective cohort study was conducted in a medical intensive care unit (ICU) in central Taiwan. Patients in the ICU with sepsis between January 2014 and December 2015 were included for analysis. All of these patients received protocol-based management, including blood sugar monitoring every 2 h for the first 24 h of ICU admission. Mean amplitude of glycemic excursions (MAGE) and coefficient of variation (CoV) were used to assess GV. Results A total of 452 patients (mean age 71.4 ± 14.7 years; 76.7% men) were enrolled for analysis. They were divided into high GV (43.4%, 196/452) and low GV (56.6%, 256/512) groups using MAGE 65 mg/dL as the cut-off point. Patients with high GV tended to have higher HbA1c (6.7 ± 1.8% vs. 5.9 ± 0.9%, p   65 mg/dL. Higher GV within 24 h of ICU admission was independently associated with increased 30-day mortality. These findings highlight the need to monitor GV in septic patients early during an ICU admission.

Karin Erwander, Kjell Ivarsson, Mona Landin Olsson et al.

Introduction: The Emergency Department (ED) is a common route to hospitalization for critically ill and older adults. Older patients are admitted to hospital at a higher rate and have longer length of stay (LOS) when hospitalized. To be able to confront an increasing aging population, meet their medical needs and influence rising costs of health care, there is a need to focus on the older population. In Scandinavia, few studies are made that focus on the geriatric population at the ED. It is essential to early identify risk factors for hospitalization at the ED to improve the medical care for older adults and the influence of prehospital comorbidities. Methods: This is a retrospective observational study of older adults visiting the ED in southwest Sweden. The aim of this study was to examine if routinely collected patient demographics and prehospital comorbidities were associated with ED disposition and in-patient process outcomes. The data collection was generated from the Regional Healthcare Information Platform. The variables extracted were age, gender, ED-visits, LOS at ED, admission rate, in-hospital LOS and comorbidities before visiting the ED. Results: A total of 15 528 patients aged ≥ 65 years visited the ED during 2016, 8 098 (52%) were female and 7 430 (48%) were male, 6 631 (41%) were 65–74 years of age, 5 585 (36%) were 75–84 years of age and 3 612 (23%) were 85 years or older. LOS at the ED were over 4 hours for 45% of the population. Patients aged 85 or older had a Hazard ratio of 2.56 (CI 2.33–2.82) for admission and patients with HF had a Hazard ratio of 1.75 (CI 1.46–2.09). Conclusion: Patients with old age, HF and comorbidities as prehospital conditions have a significant higher risk for admission to the hospital and a longer in-hospital stay regardless reason for the ED visit. The awareness of this could help physicians identify older patients with high risk for admission and early initiate an admission plan to be able to reduce LOS at the ED.

Serekara Gideon Christian*1, Orokwu Eziaku Chukuigwe-Igbere1, Ransom Baribefii Jacob1 Happiness Nkiruka Ejimmadu1

Soot is a mass of impure particles of carbon obtained from incomplete hydrocarbon combustion. Soot is an ultrafine air-borne pollutant and enters into body through ingestion, skin contact and inhalation cause devastating effects on the blood cells. The aim of the study was to determine the effect of soot inhalation on methaemoglobin and oxyhaemoglobin levels of individuals resident in both Iwofe, Port Harcourt (exposed subjects) and Ihiala, Anambra state (control subjects). The study is a case control study involving residents of Iwofe, Rumuolumeni, who have been exposed to soot pollution in the environment for an average period of one year. Iwofe is in Port Harcourt, the capital of Rivers State, Nigeria. A total of fifty (50) test samples were obtained. Thirty control samples were obtained from subjects in Ihiala a city located in the South of Anambra state where illegal oil refineries and other major means of soot generation are not as comparable to what is present in Port Harcourt and its environs. Methaemoglobin and Oxyhaemoglobin concentrations were analyzed using spectrophotometric method. The data obtained was analyzed using SPSS for descriptive statistics (mean and standard deviations) and inferential statistics (t-test). The student t-test was used to test for difference in the methaemoglobin and oxyhaemoglobin levels between the exposed subjects and non-exposed controls and based on age groups and gender. An error of probability (p ≤ 0.05) is considered statistically significant. Generally, there was a significant increase in the methaemoglobin concentrations and decrease in oxyhaemoglobin levels of exposed subjects showing the effect of soot inhalation.  Activities of illegal oil refineries (a major source of soot pollution in the city) should be stopped along with other activities like burning of tyres, indiscriminate burning of wastes and gas flaring etc.