Journal of Medical Systems最新文献

{"title":"Comparative Analysis of Large Language Models and Machine Learning for ASA Classification Using Structured Electronic Health Record Data.","authors":"Shih-Yu Ko, Ting-Yun Huang, Ming-Siang Huang, Yi-Hsuan Lin, Yung-Cheng Su, Yung-Chun Chang","doi":"10.1007/s10916-026-02393-2","DOIUrl":"10.1007/s10916-026-02393-2","url":null,"abstract":"<p><p>The American Society of Anesthesiologists Physical Status classification suffers from substantial inter-rater variability, compromising perioperative risk assessment. Large language models have demonstrated remarkable capabilities in medical text processing, but their application to anesthetic risk prediction remains unexplored. This study compared large language model performance against traditional machine learning approaches for automated ASA classification. This retrospective study analyzed 21,049 surgical records from the MOVER database (2017–2022), utilizing structured preoperative data including patient demographics, laboratory values, medication records, and procedural information for ASA I-IV classifications. We compared five large language models (GPT-4, GPT-4 Turbo, GPT-4o, Gemini 1.5 Flash, Gemini 1.5 Pro) against eight traditional machine learning algorithms using 10-fold cross-validation. Four training approaches were implemented: zero-shot learning, few-shot learning, few-shot with machine learning features, and few-shot with large language model features. Performance was evaluated using precision, recall, F₁-score, AUROC, and AUPRC. GPT-4o Few Shot Feature LLM achieved superior performance with 66% accuracy, F₁-score of 0.65, and AUROC of 0.79, substantially outperforming the best traditional method LightGBM (54% accuracy), representing a 12%-point improvement in classification accuracy. Large language models demonstrated excellent performance in intermediate risk categories (ASA II-III) with precision 0.60–0.75 and recall 0.60–0.78, covering 85% of surgical patients. However, all models showed low recall rates for ASA IV patients (28%), indicating challenges identifying highest-risk patients. Feature importance analysis revealed convergent validity between approaches, consistently identifying weight, age, and medication complexity as primary predictors. Large language models outperformed traditional machine learning for ASA classification in this dataset, particularly in intermediate-risk patients comprising most surgical cases. Their superior ability to interpret structured clinical text, including procedure names, medication records, and laboratory results, demonstrates potential for reducing inter-rater variability. However, low recall for high-risk patients necessitates human oversight and thus positions these models as decision support tools. Multi-institutional validation is essential before clinical implementation.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"50 1","pages":""},"PeriodicalIF":8.8,"publicationDate":"2026-04-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773721","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Designing Operating Rooms as an Integrated Socio-Technical Ecosystem: Practical Lessons from a High-Volume Tertiary Center.","authors":"Dina Orkin, Ofir Morag, Hadas Moshonov-Cohavi, Tania Fishman, Gad Segal","doi":"10.1007/s10916-026-02387-0","DOIUrl":"10.1007/s10916-026-02387-0","url":null,"abstract":"","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"50 1","pages":""},"PeriodicalIF":8.8,"publicationDate":"2026-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773850","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Correction: The Limits of Humans in Data Gathering: Documentation Error Rates in the Electronic Health Record in the Operating Room.","authors":"Andrew R Bradley, Abner Barbosa, Logan Younk, Naila Rocha, Peter F Nichol","doi":"10.1007/s10916-026-02396-z","DOIUrl":"10.1007/s10916-026-02396-z","url":null,"abstract":"","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"50 1","pages":""},"PeriodicalIF":8.8,"publicationDate":"2026-04-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13111492/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773788","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Validating Digital Scribes: A Scoping Review of Evaluation Practices and Clinical Use.","authors":"Ekin Kerimoğlu, Fleur V Notermans, Milou E W M Silkens, Marleen de Mul, Kees C T B Ahaus, Robert M A van der Boon","doi":"10.1007/s10916-026-02392-3","DOIUrl":"10.1007/s10916-026-02392-3","url":null,"abstract":"<p><p>Clinical documentation accounts for a substantial share of clinicians’ working time, contributing to administrative burden and reduced patient-facing care. Artificial intelligence has stimulated the development of digital scribes that combine speech recognition (ASR) and large language models (LLMs) to generate clinical notes from patient–provider conversations with the aim to automate and support this process and reduce this burden. This scoping review explores how digital scribes are currently validated, both technical and clinical, and whether they reliably support clinical workflows. Using the Technology Readiness Level (TRL) framework, we show that most systems remain in early development stages (typically TRL 3&4), with only a small number progressing to workflow integration. While digital scribes show potential to improve documentation efficiency, validation methods are highly heterogeneous, most studies rely on simulated or retrospective data, and real-world testing is limited. Consequently, cross-system comparisons and conclusions about clinical performance remain limited. We identified three motivational frames: human-, performance-, and system-oriented, which shape evaluation practices and outcome expectations. These findings suggest that successful implementation depends not only on scribes’ technical capability but also on alignment with clinical needs and documentation styles. Overall, our review underscores the need for standardised validation frameworks and prospective real-world studies to ensure that digital scribes progress beyond their current low TRL and move from experimental promise to safe, effective, and sustainable integration into clinical care.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"50 1","pages":""},"PeriodicalIF":8.8,"publicationDate":"2026-04-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13106249/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773865","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Signal Detection Using Change Point Analysis to Identify Safety Signals from Spontaneous Reports in Aortic Stent Grafts in Japan.","authors":"Kohei Takahashi, Hideto Yokoi, Fumiaki Mikami, Tomomi Satomi, Morikazu Seki, Takako Takayama Niwa, Hideki Hanaoka, Yosuke Inaba, Daisuke Koide","doi":"10.1007/s10916-026-02388-z","DOIUrl":"10.1007/s10916-026-02388-z","url":null,"abstract":"<p><p>Signal detection is essential for identifying potential safety issues with drugs and medical devices. Change point analysis (CPA) detects changes in longitudinal data for a medical device without incorporating comparisons with other devices. This study investigated the temporal relationship between endoleak reports for aortic stent grafts identified using CPA and the issuance of the \"Pharmaceuticals and Medical Devices Agency (PMDA) Alert for Proper Use of Medical Devices\" in Japan. Device malfunction and adverse event reports for aortic stent grafts submitted to the PMDA between April 2008 and September 2022 were analyzed. The reports were mapped to the Medical Device Problem Terminology from the Japan Federation of Medical Devices Associations. CPA was used to identify temporal changes in report counts, with change points having confidence levels ≥ 90% being considered significant. A total of 20,678 reports were analyzed, including 23,541 device malfunctions and 30,500 adverse events. Endoleaks accounted for 10,779 events. CPA identified a significant increase in endoleak events in April 2015 (2 years before the regulatory alert). Confidence levels exceeded 90% in 76 out of 174 evaluated time points, first in July 2008 (9 years before the alert) and next in April 2016 (1 year before the alert). Based on these findings, CPA may help prioritize investigations of medical-device-related malfunctions and adverse events. However, the results do not indicate earlier or superior signal detection relative to regulatory authorities. Further research is needed to assess CPA's applicability to other medical devices and malfunctions.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"50 1","pages":""},"PeriodicalIF":8.8,"publicationDate":"2026-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC13106239/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773879","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Patient-Initiated Permanent Deletion of Their Electronic Health Record Data: implications for Artificial Intelligence and Big Data in Healthcare : Clyde T. Matava, MBChB, MMed, MHSc.","authors":"Clyde T Matava, Yvonne Fahy, Asad Siddiqui, Gregory Johnson, Melissa McCradden, Allan F Simpao","doi":"10.1007/s10916-026-02385-2","DOIUrl":"10.1007/s10916-026-02385-2","url":null,"abstract":"","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"50 1","pages":""},"PeriodicalIF":8.8,"publicationDate":"2026-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773938","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Telemedicine Devices in Prehospital Trauma Care: A Scoping Review of Telemedical Actions, Clinical Utilities and Implementation Barriers.","authors":"Ziqiang Xu, Weijing Sui, Weilin Jiang, Jingbang Liu, Hong Zhu, Yeru Xia, Xiaoyan Gong, Yiyu Zhuang","doi":"10.1007/s10916-026-02389-y","DOIUrl":"10.1007/s10916-026-02389-y","url":null,"abstract":"<p><p>Device-enabled telemedicine is increasingly used in prehospital trauma, but reported workflows remain heterogeneous across operational settings, device types, and study designs. A workflow-oriented synthesis is needed to clarify which telemedical actions are being evaluated, what device categories enable them, what clinical utilities are reported, and which implementation barriers recur across studies. To map device-enabled telemedicine in prehospital trauma by linking telemedical actions, enabling device categories, reported clinical utilities, and implementation barrier domains. We conducted a scoping review following the Arksey and O’Malley framework and PRISMA-ScR reporting guidance. Four databases were searched for peer-reviewed primary studies published from 1 January 2010 to 1 July 2025. Eligible studies evaluated prehospital trauma workflows in which a device-enabled telemedical function informed triage, diagnosis, monitoring, or treatment decisions. Using inductive content analysis, we mapped included studies to a four-layer framework of telemedical actions, device categories, clinical utilities, and implementation barriers. Twenty-two studies were included. Most evaluations were simulation-based (15/22, 68.2%). Within the retrieved evidence base, most explicitly described telemedical workflows were synchronous (20/22, 90.9%), whereas asynchronous and hybrid functions were less frequently represented. Telemedical actions clustered around remote consultation and telementoring, remote image acquisition with remote interpretation, algorithm-supported triage, and remote physiological monitoring. Reported utilities most often aligned with diagnostic and triage impact and with system coordination and access. Common and overlapping barriers included evidence maturity gaps, connectivity and technical robustness, usability and training burden, governance and system readiness, and operational context constraints. The evidence base for prehospital trauma telemedicine remains early-stage and is dominated by simulation studies, with limited implementation-ready evaluation. An action-oriented evidence map clarifies the telemedical pathways most often studied and the recurrent barriers to translation, suggesting that future progress will depend on real-world evaluations of workflow resilience and implementation readiness rather than device capability alone.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"50 1","pages":""},"PeriodicalIF":8.8,"publicationDate":"2026-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773873","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Safety Envelopes and Continuous Assurance for Updates and Postmarket Surveillance in Digital Mental Health Technologies.","authors":"Siddharth Nandagopal","doi":"10.1007/s10916-026-02375-4","DOIUrl":"10.1007/s10916-026-02375-4","url":null,"abstract":"<p><p>Digital mental health technologies (DMHTs) increasingly behave as continuously updated services, yet most oversight mechanisms still evaluate them as comparatively static products. This creates a governance gap where software change, deployment context, and mental health risk interact. This perspective presents SECA (Safety Envelope and Continuous Assurance) as a governance framework for regulated software as a medical device (SaMD) DMHTs, with extension to the wider “adjacent app” market through procurement and accreditation levers. Rather than treating oversight as review of a fixed version, SECA combines a bounded Safety Envelope, which pre-specifies change classes and evidence expectations, with Continuous Assurance, which monitors real-world performance and supports escalation, pause, or rollback when needed. The framework is mapped to the EU/UK, US, and India by aligning with existing post-market and vigilance structures while highlighting context triggers especially relevant in mental health, including population vulnerability, language and cultural setting, and interaction design changes that may alter help-seeking or crisis behavior.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"50 1","pages":""},"PeriodicalIF":8.8,"publicationDate":"2026-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773947","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Artificial Intelligence For 6P Medicine: Consolidating AI Needs of Predictive, Preventive, Personalized, Participatory, Precision, and Public Health Trajectories.","authors":"Aly Khalifa, Rada Hussein","doi":"10.1007/s10916-026-02391-4","DOIUrl":"10.1007/s10916-026-02391-4","url":null,"abstract":"<p><p>Artificial intelligence (AI) is increasingly integrated into healthcare, and it aligns naturally with the 6P medicine model (predictive, preventive, personalized, participatory, precision, and public health). In this narrative review, we synthesize recent evidence on how AI methods-including machine learning, deep learning, large language models, and digital twins-support each “P”, and we highlight where evidence is mature versus still exploratory. Across the literature, the strongest evidence focuses on predictive, preventive, and precision applications (e.g., imaging, risk stratification, and treatment decision support). In contrast, participatory and public health applications are less consistently evaluated and introduce additional challenges around equity, trust, and governance. We propose a practical 6P-AI implementation blueprint that links common use cases to data requirements, workflow integration steps, evaluation designs, and post-deployment monitoring. Responsible adoption requires attention to data quality, privacy, bias mitigation, transparency, human oversight, and alignment with regulatory frameworks (e.g., the European Health Data Space and the EU AI Act).</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"50 1","pages":""},"PeriodicalIF":8.8,"publicationDate":"2026-04-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147716920","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Beyond Root Mean Squared Error: Expanding Operational Validation for Surgical Bed Forecasting Models.","authors":"Jorge Valle-Reconco, Patricia Soriano, Isaac Zablah","doi":"10.1007/s10916-026-02378-1","DOIUrl":"10.1007/s10916-026-02378-1","url":null,"abstract":"","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"50 1","pages":""},"PeriodicalIF":8.8,"publicationDate":"2026-04-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147698910","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}