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

Zhongying Deng, Cheng Tang, Ziyan Huang et al.

Foundation models have demonstrated remarkable success across diverse domains and tasks, primarily due to the thrive of large-scale, diverse, and high-quality datasets. However, in the field of medical imaging, the curation and assembling of such medical datasets are highly challenging due to the reliance on clinical expertise and strict ethical and privacy constraints, resulting in a scarcity of large-scale unified medical datasets and hindering the development of powerful medical foundation models. In this work, we present the largest survey to date of medical image datasets, covering over 1,000 open-access datasets with a systematic catalog of their modalities, tasks, anatomies, annotations, limitations, and potential for integration. Our analysis exposes a landscape that is modest in scale, fragmented across narrowly scoped tasks, and unevenly distributed across organs and modalities, which in turn limits the utility of existing medical image datasets for developing versatile and robust medical foundation models. To turn fragmentation into scale, we propose a metadata-driven fusion paradigm (MDFP) that integrates public datasets with shared modalities or tasks, thereby transforming multiple small data silos into larger, more coherent resources. Building on MDFP, we release an interactive discovery portal that enables end-to-end, automated medical image dataset integration, and compile all surveyed datasets into a unified, structured table that clearly summarizes their key characteristics and provides reference links, offering the community an accessible and comprehensive repository. By charting the current terrain and offering a principled path to dataset consolidation, our survey provides a practical roadmap for scaling medical imaging corpora, supporting faster data discovery, more principled dataset creation, and more capable medical foundation models.

Nichapha Chongthavonsatit, Kamonchanok Khachintararod, Pongsakorn Atiksawedparit et al.

Abstract Background Dyspnea is an individual’s sensation of discomfort during breathing. For patients with dyspnea who are unable to communicate, the Respiratory Distress Observation Scale (RDOS) was used to rate the severity based on eight parameters observed. In the emergency department, emergency nurses triage the patients with dyspnea and monitor their symptom severities, while emergency physicians evaluate the patients to determine treatment decisions. We aim to study the inter-rater agreement of RDOS measurement between emergency physicians and nurses. Method Between March 2024 and January 2025, an observational cross-sectional study was conducted in resuscitation rooms of two university academic hospitals. The participants were emergency physicians and nurses who were the first responders to adult patients presenting to the resuscitation room with dyspnea. The RDOS assessment was done individually by the data record forms within 20 min after arriving in the resuscitation rooms. The primary outcome was inter-rater agreement on RDOS measurement between emergency physicians and nurses. Result By 176 patients with dyspnea (N = 176), 44 emergency physicians and 55 nurses were included with no difference in either age or clinical experiences. The overall physicians and nurses reported a fair agreement of RDOS severity (58%) with a Kappa statistic of 0.54 (95% CI: 0.47–0.61, SE 0.075, p < 0.001), with 57% agreement in patients with intact communication and 61% agreement in those with impaired communication. The ratings were internally consistent and homogeneous among each profession. (the overall IIC 0.737 and 0.767, respectively). The inter-rater reliability was poor to moderate across both professions when scoring seven of eight RDOS parameters. Conclusions Emergency physicians and nurses have fair inter-rater agreement on RDOS measurement. Seven of the eight RDOS parameters revealed poor to moderate inter-rater reliability in both professions. Therefore, implementing RDOS in the ED requires customized training and calibrating the RDOS assessment.

Jorge García-Torres, Øyvind Meinich-Bache, Sara Brunner et al.

Around 10% of newborns require some help to initiate breathing, and 5\% need ventilation assistance. Accurate Time of Birth (ToB) documentation is essential for optimizing neonatal care, as timely interventions are vital for proper resuscitation. However, current clinical methods for recording ToB often rely on manual processes, which can be prone to inaccuracies. In this study, we present a novel two-stream fusion system that combines the power of image and video analysis to accurately detect the ToB from thermal recordings in the delivery room and operating theater. By integrating static and dynamic streams, our approach captures richer birth-related spatiotemporal features, leading to more robust and precise ToB estimation. We demonstrate that this synergy between data modalities enhances performance over single-stream approaches. Our system achieves 95.7% precision and 84.8% recall in detecting birth within short video clips. Additionally, with the help of a score aggregation module, it successfully identifies ToB in 100% of test cases, with a median absolute error of 2 seconds and an absolute mean deviation of 4.5 seconds compared to manual annotations.

Andrea E. Davidson, Jessica M. Ray, Ayush K. Patel et al.

This study reports the findings of qualitative interview sessions conducted with ICU clinicians for the co-design of a system user interface of an artificial intelligence (AI)-driven clinical decision support (CDS) system. This system integrates medical record data with wearable sensor, video, and environmental data into a real-time dynamic model that quantifies patients' risk of clinical decompensation and risk of developing delirium, providing actionable alerts to augment clinical decision-making in the ICU setting. Co-design sessions were conducted as semi-structured focus groups and interviews with ICU clinicians, including physicians, mid-level practitioners, and nurses. Study participants were asked about their perceptions on AI-CDS systems, their system preferences, and were asked to provide feedback on the current user interface prototype. Session transcripts were qualitatively analyzed to identify key themes related to system utility, interface design features, alert preferences, and implementation considerations. Ten clinicians participated in eight sessions. The analysis identified five themes: (1) AI's computational utility, (2) workflow optimization, (3) effects on patient care, (4) technical considerations, and (5) implementation considerations. Clinicians valued the CDS system's multi-modal continuous monitoring and AI's capacity to process large volumes of data in real-time to identify patient risk factors and suggest action items. Participants underscored the system's unique value in detecting delirium and promoting non-pharmacological delirium prevention measures. The actionability and intuitive interpretation of the presented information was emphasized. ICU clinicians recognize the potential of an AI-driven CDS system for ICU delirium and acuity to improve patient outcomes and clinical workflows.

Imaan Hameed, Huma Umar, Fozia Umber et al.

In recent years, LLM-based maternal health chatbots have been widely deployed in low-resource settings, but they often ignore real-world contexts where women may not own phones, have limited literacy, and share decision-making within families. Through the deployment of a WhatsApp-based maternal health chatbot with 48 pregnant women in Lahore, Pakistan, we examine barriers to use in populations where phones are shared, decision-making is collective, and literacy varies. We complement this with focus group discussions with obstetric clinicians. Our findings reveal how adoption is shaped by proxy consent and family mediation, intermittent phone access, silence around asking questions, infrastructural breakdowns, and contested authority. We frame barriers to non-use as culturally conditioned rather than individual choices, and introduce the Relational Chatbot Design Grammar (RCDG): four commitments that enable mediated decision-making, recognize silence as engagement, support episodic use, and treat fragility as baseline to reorient maternal health chatbots toward culturally grounded, collective care.

Tommaso Pettenuzzo, Honoria Ocagli, Nicolò Sella et al.

Abstract Background In the last decades, veno-arterial extracorporeal membrane oxygenation (V-A ECMO) has been gaining in popularity for intraoperative support during lung transplant (LT), being advocated for routinely use also in uncomplicated cases. Compared to off-pump strategy and, secondarily, to traditional cardiopulmonary bypass (CPB), V-A ECMO seems to offer a better hemodynamic stability and oxygenation, while data regarding blood product transfusions, postoperative recovery, and mortality remain unclear. This systematic review and network meta-analysis aims to evaluate the comparative efficacy and safety of V-A ECMO and CPB as compared to OffPump strategy during LT. Methods A comprehensive literature search was conducted across multiple databases (PubMed Embase, Cochrane, Scopus) and was updated in February 2024. A Bayesian network meta-analysis (NMA), with a fixed-effect approach, was performed to compare outcomes, such as intraoperative needing of blood products, invasive mechanical ventilation (IMV) duration, intensive care unit (ICU) length of stay (LOS), surgical duration, needing of postoperative ECMO, and mortality, across different supports (i.e., intraoperative V-A (default (d) or rescue (r)) ECMO, CPB, or OffPump). Findings Twenty-seven observational studies (6113 patients) were included. As compared to OffPump surgery, V-A ECMOd, V-A ECMOr, and CPB recorded a higher consumption of all blood products, longer IMV durations, prolonged ICU LOS, surgical duration, and higher mortalities. Comparing different extracorporeal supports, V-A ECMOd and, secondarily, V-A ECMOr overperformed CPB in nearly all above mentioned outcomes, except for RBC transfusions. The lowest rate of postoperative ECMO was recorded after OffPump surgery, while no differences were found comparing different extracorporeal supports. Finally, older age, male gender, and body mass index ≥ 25 kg/m2 negatively impacted on RBC transfusions, ICU LOS, surgical duration, need of postoperative ECMO, and mortality, regardless of the intraoperative extracorporeal support investigated. Interpretation This comparative network meta-analysis highlights that OffPump overperformed ECMO and CPB in all outcomes of interest, while, comparing different extracorporeal supports, V-A ECMOd and, secondarily, V-A ECMOr overperformed CPB in nearly all above mentioned outcomes, except for RBC transfusions. Older age, male gender, and higher BMI negatively affect several outcomes across different intraoperative strategies, regardless of the intraoperative extracorporeal support investigated. Future prospective studies are necessary to optimize and standardize the intraoperative management of LT.

Katherine Vlasica, Amanda Hall

Introduction: Rectus sheath blocks have been used for decades in the operating room for analgesia following umbilical surgical procedures. We present the first reported case series of a rectus sheath block used in the emergency department (ED) for the reduction of an umbilical hernia. Case Series: Four patients presented to the ED for painful, non-reducible umbilical hernias. An ultrasound-guided bilateral rectus sheath block was used in all four patients with complete pain relief and an easy hernia reduction. Conclusion: Rectus sheath blocks are an excellent addition to a multimodal analgesic regimen in periumbilical pain and painful procedures. This block is easy to perform and implement for pain control in umbilical hernias in an ED setting.

B. Shuriya, S. Vimal Kumar, K. Bagyalakshmi

In this paper, we introduce the Fully Homomorphic Integrity Model (HIM), a novel approach designed to enhance security, efficiency, and reliability in encrypted data processing, primarily within the health care industry. HIM addresses the key challenges that noise accumulation, computational overheads, and data integrity pose during homomorphic operations. Our contribution of HIM: advances in noise management through the rational number adjustment; key generation based on personalized prime numbers; and time complexity analysis details for key operations. In HIM, some additional mechanisms were introduced, including robust mechanisms of decryption. Indeed, the decryption mechanism ensures that the data recovered upon doing complex homomorphic computation will be valid and reliable. The healthcare id model is tested, and it supports real-time processing of data with privacy maintained concerning patients. It supports analytics and decision-making processes without any compromise on the integrity of information concerning patients. Output HIM promotes the efficiency of encryption to a greater extent as it reduces the encryption time up to 35ms and decryption time up to 140ms, which is better when compared to other models in the existence. Ciphertext size also becomes the smallest one, which is 4KB. Our experiments confirm that HIM is indeed a very efficient and secure privacy-preserving solution for healthcare applications

Musa Taib, Jiajun Wu, Steve Drew et al.

The top priority of a Housing and Homelessness System of Care (HHSC) is to connect people experiencing homelessness to supportive housing. An HHSC typically consists of many agencies serving the same population. Information technology platforms differ in type and quality between agencies, so their data are usually isolated from one agency to another. Larger agencies may have sufficient data to train and test artificial intelligence (AI) tools but smaller agencies typically do not. To address this gap, we introduce a Federated Learning (FL) approach enabling all agencies to train a predictive model collaboratively without sharing their sensitive data. We demonstrate how FL can be used within an HHSC to provide all agencies equitable access to quality AI and further assist human decision-makers in the allocation of resources within HHSC. This is achieved while preserving the privacy of the people within the data by not sharing identifying information between agencies without their consent. Our experimental results using real-world HHSC data from Calgary, Alberta, demonstrate that our FL approach offers comparable performance with the idealized scenario of training the predictive model with data fully shared and linked between agencies.

Jennie Karlsson, Marisa Wodrich, Niels Christian Overgaard et al.

Deep learning has shown to have great potential in medical applications. In critical domains as such, it is of high interest to have trustworthy algorithms which are able to tell when reliable assessments cannot be guaranteed. Detecting out-of-distribution (OOD) samples is a crucial step towards building a safe classifier. Following a previous study, showing that it is possible to classify breast cancer in point-of-care ultrasound images, this study investigates OOD detection using three different methods: softmax, energy score and deep ensembles. All methods are tested on three different OOD data sets. The results show that the energy score method outperforms the softmax method, performing well on two of the data sets. The ensemble method is the most robust, performing the best at detecting OOD samples for all three OOD data sets.

Guojun Tang, Jason E. Black, Tyler S. Williamson et al.

Integrating Electronic Health Records (EHR) and the application of machine learning present opportunities for enhancing the accuracy and accessibility of data-driven diabetes prediction. In particular, developing data-driven machine learning models can provide early identification of patients with high risk for diabetes, potentially leading to more effective therapeutic strategies and reduced healthcare costs. However, regulation restrictions create barriers to developing centralized predictive models. This paper addresses the challenges by introducing a federated learning approach, which amalgamates predictive models without centralized data storage and processing, thus avoiding privacy issues. This marks the first application of federated learning to predict diabetes using real clinical datasets in Canada extracted from the Canadian Primary Care Sentinel Surveillance Network (CPCSSN) without crossprovince patient data sharing. We address class-imbalance issues through downsampling techniques and compare federated learning performance against province-based and centralized models. Experimental results show that the federated MLP model presents a similar or higher performance compared to the model trained with the centralized approach. However, the federated logistic regression model showed inferior performance compared to its centralized peer.

Mya Cubitt, Rachel Keys

China Atsumi, Katsuhiro Aikawa, Keita Takahashi et al.

Abstract Introduction Recently, modified thoracoabdominal nerve block through perichondrial approach (M-TAPA) has been introduced as a novel trunk block. To date, studies comparing its clinical advantages with those of existing local anesthetic techniques are scarce. We aimed to compare the analgesic efficacy of M-TAPA to that of wound infiltration analgesia (WIA) in patients who underwent gynecological laparoscopic surgeries. Methods We studied medical records from January 2020 to July 2021 at Hokkaido University Hospital. The primary outcome was the number of analgesic requirements in the first 24 h postoperatively. Secondary outcomes were the time until the first analgesic requirement and adverse events regarding local anesthetic techniques. To address confounding, a regression model was used. Results Data from 90 of 231 patients were analyzed (M-TAPA group, n = 40; WIA group, n = 50). For the primary outcome, means and 95% confidence intervals for each group and between-group differences were as follows: 2.25 (1.74, 2.76), 2.28 (1.81, 2.75), and −0.03 (−0.72, 0.66), respectively. Adjusted mean difference was 0.39 (−0.32, 1.11). There were no significant differences in means between groups, with or without adjustment for covariates (p = 0.93, 0.28). Furthermore, no significant difference was detected in the time until the first analgesic requirement and adverse events related to local anesthesia. Conclusion Our results demonstrate that M-TAPA did not reduce postoperative analgesic requirements compared to WIA. In a future clinical trial, sufficient visceral pain control may be required to evaluate the effectiveness of M-TAPA over WIA in patients undergoing laparoscopic gynecological surgery.

Christopher T. Edmunds, Kate Lachowycz, Sarah McLachlan et al.

Abstract Background Helicopter Emergency Medical Services (HEMS) are a limited and expensive resource, and should be intelligently tasked. HEMS dispatch was identified as a key research priority in 2011, with a call to identify a ‘general set of criteria with the highest discriminating potential’. However, there have been no published data analyses in the past decade that specifically address this priority, and this priority has been reaffirmed in 2023. The objective of this study was to define the dispatch criteria available at the time of the initial emergency call with the greatest HEMS utility using a large, regional, multi-organizational dataset in the UK. Methods This retrospective observational study utilized dispatch data from a regional emergency medical service (EMS) and three HEMS organisations in the East of England, 2016–2019. In a logistic regression model, Advanced Medical Priority Dispatch System (AMPDS) codes with ≥ 50 HEMS dispatches in the study period were compared with the remainder to identify codes with high-levels of HEMS patient contact and HEMS-level intervention/drug/diagnostic (HLIDD). The primary outcome was to identify AMPDS codes with a > 10% HEMS dispatch rate of all EMS taskings that would result in 10–20 high-utility HEMS dispatches per 24-h period in the East of England. Data were analysed in R, and are reported as number (percentage); significance was p < 0.05. Results There were n = 25,491 HEMS dispatches (6400 per year), of which n = 23,030 (90.3%) had an associated AMPDS code. n = 13,778 (59.8%) of HEMS dispatches resulted in patient contact, and n = 8437 (36.6%) had an HLIDD. 43 AMPDS codes had significantly greater rates of patient contact and/or HLIDD compared to the reference group. In an exploratory analysis, a cut-off of ≥ 70% patient contact rate and/or ≥ 70% HLIDD (with a > 10% HEMS dispatch of all EMS taskings) resulted in 17 taskings per 24-h period. This definition derived nine AMPDS codes with high HEMS utility. Conclusion We have identified nine ‘golden’ AMPDS codes, available at the time of initial emergency call, that are associated with high-levels of whole-system and HEMS utility in the East of England. We propose that UK EMS should consider immediate HEMS dispatch to these codes.

Yunxiang Li, Hua-Chieh Shao, Xiao Liang et al.

Recently, the diffusion model has emerged as a superior generative model that can produce high quality and realistic images. However, for medical image translation, the existing diffusion models are deficient in accurately retaining structural information since the structure details of source domain images are lost during the forward diffusion process and cannot be fully recovered through learned reverse diffusion, while the integrity of anatomical structures is extremely important in medical images. For instance, errors in image translation may distort, shift, or even remove structures and tumors, leading to incorrect diagnosis and inadequate treatments. Training and conditioning diffusion models using paired source and target images with matching anatomy can help. However, such paired data are very difficult and costly to obtain, and may also reduce the robustness of the developed model to out-of-distribution testing data. We propose a frequency-guided diffusion model (FGDM) that employs frequency-domain filters to guide the diffusion model for structure-preserving image translation. Based on its design, FGDM allows zero-shot learning, as it can be trained solely on the data from the target domain, and used directly for source-to-target domain translation without any exposure to the source-domain data during training. We evaluated it on three cone-beam CT (CBCT)-to-CT translation tasks for different anatomical sites, and a cross-institutional MR imaging translation task. FGDM outperformed the state-of-the-art methods (GAN-based, VAE-based, and diffusion-based) in metrics of Frechet Inception Distance (FID), Peak Signal-to-Noise Ratio (PSNR), and Structural Similarity Index Measure (SSIM), showing its significant advantages in zero-shot medical image translation.

Louise Rose, Andreas Xyrichis, Joel Meyer

Silvana Soares dos Santos, Antonio Paulo Nassar Junior

Kwang Wook Jo

Traumatic brain injury (TBI) is a critical cause of disability and death worldwide. Many studies have been conducted aimed at achieving favorable neurologic outcomes by reducing secondary brain injury in TBI patients. However, ground-breaking outcomes are still insufficient so far. Because mild-to-moderate hypothermia (32°C–35°C) has been confirmed to help neurological recovery for recovered patients after circulatory arrest, it has been recognized as a major neuroprotective treatment plan for TBI patients. Thereafter, many clinical studies about the effect of therapeutic hypothermia (TH) on severe TBI have been conducted. However, efficacy and safety have not been demonstrated in many large-scale randomized controlled studies. Rather, some studies have demonstrated an increase in mortality rate due to complications such as pneumonia, so it is not highly recommended for severe TBI patients. Recently, some studies have shown results suggesting TH may help reperfusion/ischemic injury prevention after surgery in the case of mass lesions, such as acute subdural hematoma, and it has also been shown to be effective in intracranial pressure control. In conclusion, TH is still at the center of neuroprotective therapeutic studies regarding TBI. If proper measures can be taken to mitigate the many adverse events that may occur during the course of treatment, more positive efficacy can be confirmed. In this review, we look into adverse events that may occur during the process of the induction, maintenance, and rewarming of targeted temperature management and consider ways to prevent and address them.

Eduardo Navarro, Pavel Mazirka, Tera Thigpin et al.

Split-thickness skin grafting (STSG) is one of the most frequently used approaches to wound coverage in large size deep partial thickness and full thickness burns. The technique, however, is often degraded by partial, and at times complete, failure of graft to take. Though often multifactorial; operative technique, quality of wound bed and the type of dressing used are key factors in the ability of the STSG to heal. Various dressing methods have been described in the literature with varying benefits and degree of success but no consensus on a gold standard has been reached. This study describes the use of intra-operatively placed Lactic-Acid-Based Copolymer (LABC) as a dressing for meshed STSG and its impact on skin graft survival.A retrospective chart review of patients who received LABC dressing over their STSG was performed. A total of 22 patients, 16 male (73%) and 6 female (27%), with a mean age of 36.5 years, were included. Burn injuries were due to fire (10), grease (8), scald (2) and electrical (2), with a median treated area of 375 cm2 and follow up ranging from 21 to 232 days. Complete wound healing with no graft loss was documented in 17 of 22 (77%) patients; with the remaining 5 patients being lost to follow up.Intraoperative LABC dressing application after split-thickness skin grafting in burn patients was shown to be an effective strategy in protecting the STSG site and allowing graft take.

Michael Bauer, Reinhard Wetzker