Journal of Medical Systems最新文献

{"title":"Assessment and Integration of Large Language Models for Automated Electronic Health Record Documentation in Emergency Medical Services.","authors":"Enze Bai, Xiao Luo, Zhan Zhang, Kathleen Adelgais, Humaira Ali, Jack Finkelstein, Jared Kutzin","doi":"10.1007/s10916-025-02197-w","DOIUrl":"https://doi.org/10.1007/s10916-025-02197-w","url":null,"abstract":"<p><p>Automating Electronic Health Records (EHR) documentation can significantly reduce the burden on care providers, particularly in emergency care settings where rapid and accurate record-keeping is crucial. A critical aspect of this automation involves using natural language processing (NLP) techniques to convert transcribed conversations into structured EHR fields. For instance, extracting temperature values like \"102.4 Fahrenheit\" from the transcribed text \"His temperature is 39.1, which is 102.4 Fahrenheit.\" However, traditional rule-based and single-model NLP approaches often struggle with domain-specific medical terminology, contextual ambiguity, and numerical extraction errors. This study investigates the potential of integrating multiple Large Language Models (LLMs) to enhance EMS documentation accuracy. We developed an LLM integration framework and evaluated four state-of-the-art LLMs-Claude 3.5, GPT-4, Gemini, and Mistral-on a dataset comprising transcribed conversations from 40 EMS training simulations. The evaluation focused on precision, recall, and F1 score across zero-shot and few-shot learning scenarios. Results showed that the integrated LLM framework outperformed individual models, achieving overall F1 scores of 0.78 (zero-shot) and 0.81 (few-shot). In addition to quantitative evaluation, a preliminary user study was conducted with domain experts to assess the perceived usefulness and challenges of the integrated framework. The findings suggest that this approach has the potential to reduce documentation effort compared to traditional manual documentation. However, challenges such as misinterpretation of medical context and occasional omissions were noted, highlighting areas for further refinement and future work. This research is the first to systematically explore and evaluate the use of LLMs for real-time EMS EHR documentation. By addressing key challenges in automated transcription and structured data extraction, our work lays a foundation for real-world implementation, improving efficiency and accuracy in emergency medical documentation.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"49 1","pages":"65"},"PeriodicalIF":3.5,"publicationDate":"2025-05-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144086242","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":"Predicting Immunotherapy Response in Unresectable Hepatocellular Carcinoma: A Comparative Study of Large Language Models and Human Experts.","authors":"Jun Xu, Junjie Wang, Junjun Li, Zhangxiang Zhu, Xiao Fu, Wei Cai, Ruipeng Song, Tengfei Wang, Hai Li","doi":"10.1007/s10916-025-02192-1","DOIUrl":"https://doi.org/10.1007/s10916-025-02192-1","url":null,"abstract":"<p><p>Hepatocellular carcinoma (HCC) is an aggressive cancer with limited biomarkers for predicting immunotherapy response. Recent advancements in large language models (LLMs) like GPT-4, GPT-4o, and Gemini offer the potential for enhancing clinical decision-making through multimodal data analysis. However, their effectiveness in predicting immunotherapy response, especially compared to human experts, remains unclear. This study assessed the performance of GPT-4, GPT-4o, and Gemini in predicting immunotherapy response in unresectable HCC, compared to radiologists and oncologists of varying expertise. A retrospective analysis of 186 patients with unresectable HCC utilized multimodal data (clinical and CT images). LLMs were evaluated with zero-shot prompting and two strategies: the 'voting method' and the 'OR rule method' for improved sensitivity. Performance metrics included accuracy, sensitivity, area under the curve (AUC), and agreement across LLMs and physicians.GPT-4o, using the 'OR rule method,' achieved 65% accuracy and 47% sensitivity, comparable to intermediate physicians but lower than senior physicians (accuracy: 72%, p = 0.045; sensitivity: 70%, p < 0.0001). Gemini-GPT, combining GPT-4, GPT-4o, and Gemini, achieved an AUC of 0.69, similar to senior physicians (AUC: 0.72, p = 0.35), with 68% accuracy, outperforming junior and intermediate physicians while remaining comparable to senior physicians (p = 0.78). However, its sensitivity (58%) was lower than senior physicians (p = 0.0097). LLMs demonstrated higher inter-model agreement (κ = 0.59-0.70) than inter-physician agreement, especially among junior physicians (κ = 0.15). This study highlights the potential of LLMs, particularly Gemini-GPT, as valuable tools in predicting immunotherapy response for HCC.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"49 1","pages":"64"},"PeriodicalIF":3.5,"publicationDate":"2025-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144078622","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":"A Novel Dashboard To Assess Anesthesiologist Staffing of Peripartum Obstetrical Services.","authors":"Ronald A Kahn, Andrew B Leibowitz","doi":"10.1007/s10916-025-02196-x","DOIUrl":"https://doi.org/10.1007/s10916-025-02196-x","url":null,"abstract":"<p><p>Anesthesiologist staffing of procedural areas must anticipate locations requiring coverage, typical utilization of those locations, potential emergencies, and the need for provision of care in the after hours. While many dashboards exist for management of key performance indicators (KPIs) in general operating rooms, labor and delivery (L&D) suites are unique and difficult to characterize. In this report we present a novel L&D dashboard that presents relevant KPIs. We found that this dashboard tool is a valuable tool for executive decision-making and can facilitate discussions with hospital leadership and departmental staff.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"49 1","pages":"63"},"PeriodicalIF":3.5,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144024087","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":"Neurological Orientation and Verbal Assessment of Delirium-Hospital Clinical Decision Support Tool.","authors":"Kyle Hendrie, Pooja Rangan, Sumit K Agarwal, Sandeep Pagali, Babar Khan, Samreen Kidwai, Nimit Agarwal","doi":"10.1007/s10916-025-02195-y","DOIUrl":"https://doi.org/10.1007/s10916-025-02195-y","url":null,"abstract":"<p><strong>Importance: </strong>Delirium is a complex medical condition that is underdiagnosed. We present findings of a novel electronic medical record (EMR) clinical decision support tool called the Neurologic, Orientation and Verbal Assessment of Delirium (NOVAD) that can optimize delirium identification and management. In this retrospective observational study, we report the performance of NOVAD as a clinical decision support tool for delirium.</p><p><strong>Objective: </strong>We present an innovative EMR based clinical decision support tool for delirium called NOVAD. NOVAD utilizes variables from nursing assessments that are documented in EMR consistently and accurately. We aim to study the NOVAD (sensitivity, specificity, and predictive value) ability in detecting delirium among hospitalized adults and assess its potential as a clinical decision support tool for delirium.</p><p><strong>Design: </strong>A retrospective observational study of consecutive hospital admissions to Banner Health System between January 1, 2020, and December 31, 2020. 464,395 participants were included in this study. The mean age of study participants was 56.18 years (SD 20.82), with 56.2% women (n = 260,856).</p><p><strong>Results: </strong>NOVAD was 86.6% sensitive (95% CI 86.5, 86.7) and 78.1% specific (95% CI: 77.9, 78.2) based on ICD-10 diagnosis codes for delirium. The positive predictive value for NOVAD was 33.5% (95% CI 33.3, 33.6) and the negative predictive value was 97.8% (95% CI 97.8, 97.9).</p><p><strong>Conclusions and relevance: </strong>We demonstrate that an innovative EMR tool that leverages nursing assessments, NOVAD, has the potential to be used as a clinical tool to predict and screen for delirium in hospitalized adults.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"49 1","pages":"62"},"PeriodicalIF":3.5,"publicationDate":"2025-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144014616","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":"A Kidney Transplant Support System for Patient-Clinician Shared Decision-Making.","authors":"Yunwei Zhang, Marni Torkel, Samuel Muller, Germaine Wong, Jean Yang","doi":"10.1007/s10916-025-02175-2","DOIUrl":"https://doi.org/10.1007/s10916-025-02175-2","url":null,"abstract":"<p><p>An optimal deceased donor allocation program requires a fair, ethical, and transparent algorithm to ensure efficient and effective allocation of deceased donor kidneys to recipients that will benefit most, by maximizing utility of the donor organs, but at the same time, ensuring all potential candidates have equitable access and equal opportunity to this scarce resource. In response to the increasing demand and the limited availability of donor organs, there has been a global concerted effort to increase the use of less optimal donor kidneys in suitable recipients. During this complex allocation process, organ acceptance decision-making is the final step. Transplant nephrologists are typically the gatekeeper of this process and make the ultimate decision regarding organ suitability for the intended patients. However, with considerable evidence suggesting the value of shared decision making, engaging patients, caregivers and their primary nephrologists prior to accepting the allocated organ, ensures the healthcare decisions align with the patients' values and their preferences. To tackle this challenge, we developed a visualisation guided simulation-based tool to assist shared decision-making. We have shown that the three-dimensional clinical information required for organ acceptance can be represented using an intuitive and user-friendly interface. By utilizing our published allocation simulation process, simKAP, this decision support system called Kidney Transplant Support System has the capacity to forecast a sequence of potential kidneys offered to a candidate on the waiting list, with the provision of estimated waiting time, donor quality and the expected post-transplant patient survivals for each consecutive offer. Implementation of this tool may inform shared decision-making and reduce organ discards.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"49 1","pages":"60"},"PeriodicalIF":3.5,"publicationDate":"2025-05-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12064594/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143997150","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":"Should Patients Be Able To Delete their Electronic Health Records??","authors":"Clyde T Matava, Yvonne Fahy, Gregory Johnson","doi":"10.1007/s10916-025-02194-z","DOIUrl":"https://doi.org/10.1007/s10916-025-02194-z","url":null,"abstract":"","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"49 1","pages":"61"},"PeriodicalIF":3.5,"publicationDate":"2025-05-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143999455","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":"ChatOCT: Embedded Clinical Decision Support Systems for Optical Coherence Tomography in Offline and Resource-Limited Settings.","authors":"Chang Liu, Haoran Zhang, Zheng Zheng, Wenjia Liu, Chengfu Gu, Qi Lan, Weiyi Zhang, Jianlong Yang","doi":"10.1007/s10916-025-02188-x","DOIUrl":"https://doi.org/10.1007/s10916-025-02188-x","url":null,"abstract":"<p><p>Optical Coherence Tomography (OCT) is a critical imaging modality for diagnosing ocular and systemic conditions, yet its accessibility is hindered by the need for specialized expertise and high computational demands. To address these challenges, we introduce ChatOCT, an offline-capable, domain-adaptive clinical decision support system (CDSS) that integrates structured expert Q&A generation, OCT-specific knowledge injection, and activation-aware model compression. Unlike existing systems, ChatOCT functions without internet access, making it suitable for low-resource environments. ChatOCT is built upon LLaMA-2-7B, incorporating domain-specific knowledge from PubMed and OCT News through a two-stage training process: (1) knowledge injection for OCT-specific expertise and (2) Q&A instruction tuning for structured, interactive diagnostic reasoning. To ensure feasibility in offline environments, we apply activation-aware weight quantization, reducing GPU memory usage to ~ 4.74 GB, enabling deployment on standard OCT hardware. A novel expert answer generation framework mitigates hallucinations by structuring responses in a multi-step process, ensuring accuracy and interpretability. ChatOCT outperforms state-of-the-art baselines such as LLaMA-2, PMC-LLaMA-13B, and ChatDoctor by 10-15 points in coherence, relevance, and clinical utility, while reducing GPU memory requirements by 79%, while maintaining real-time responsiveness (~ 20 ms inference time). Expert ophthalmologists rated ChatOCT's outputs as clinically actionable and aligned with real-world decision-making needs, confirming its potential to assist frontline healthcare providers. ChatOCT represents an innovative offline clinical decision support system for optical coherence tomography (OCT) that runs entirely on local embedded hardware, enabling real-time analysis in resource-limited settings without internet connectivity. By offering a scalable, generalizable pipeline that integrates knowledge injection, instruction tuning, and model compression, ChatOCT provides a blueprint for next-generation, resource-efficient clinical AI solutions across multiple medical domains.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"49 1","pages":"59"},"PeriodicalIF":3.5,"publicationDate":"2025-05-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143996612","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 a Real-Time Medication Intake and Adherence Dashboard: Features, Functionality, and Data Display to Meet the Needs of Patients, Care Partners and Healthcare Providers.","authors":"Sadaf Faisal, Bincy Baby, Devine Samoth, Faisal Ghaffar, Yusra Aslam, Alexander Stavropoulos, Brent McCready-Branch, Ghada Elba, Jasdeep Kaur Gill, Prapti Choudhuri, Tejal Patel","doi":"10.1007/s10916-025-02189-w","DOIUrl":"https://doi.org/10.1007/s10916-025-02189-w","url":null,"abstract":"<p><p>To determine the key features, utilities, and functionalities of an adherence dashboard so that patients, care partners, and healthcare providers can effectively use it to identify and improve medication nonadherence. A qualitative study was conducted in four stages. In Stage 1, semi-structured interviews and focus groups were conducted with patients, care partners, and healthcare providers after showing existing dashboards to gain feedback and determine their needs and preferences for the dashboard. In Stage 2, data gathered from Stage 1 were used to develop paper prototypes. In Stage 3, these prototypes were evaluated by participants to refine the design. Finally, in Stage 4, a framework for the key components of the adherence dashboard was developed based on user feedback. Some of the key features identified include individualized medication adherence, overall adherence summaries, customizable notifications, real-time data visualization, and integration with existing healthcare systems such as electronic health records (EHR). Participants also highlighted the importance of intuitive design, ease of navigation, secure data handling, and the ability to customize the dashboard according to user preferences. This study identifies key features and functionalities and provides a user-centered framework for designing a real-time medication adherence dashboard tailored to the needs of patients, caregivers, and healthcare providers. Future work will focus on developing a fully functional dashboard, integrating it into clinical practice, and evaluating its effectiveness in improving medication adherence.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"49 1","pages":"58"},"PeriodicalIF":3.5,"publicationDate":"2025-05-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143988787","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":"The Role of Virtual Reality in Personalized Medicine: Advancing Prediction, Prevention, and Participation.","authors":"Ilias Miltiadis, Apostolos Skarlis, Pavel Burko","doi":"10.1007/s10916-025-02191-2","DOIUrl":"https://doi.org/10.1007/s10916-025-02191-2","url":null,"abstract":"<p><p>The integration of virtual reality (VR) technology into medicine represents a transformative step toward the realization of 4P medicine, characterized by personalization, prediction, prevention, and participation. VR provides an immersive platform for addressing neurological and mental disorders by tailoring virtual environments to individual patient needs, enhancing diagnostic precision, and improving therapeutic outcomes. This paper explores the applications of VR within the framework of 4P medicine, emphasizing its potential in neurorehabilitation, stress reduction, and patient engagement. The personalization aspect of VR enables the design of customized scenarios for therapeutic interventions. Predictive capabilities allow for early detection and mitigation of complications through simulated environments that analyze patient responses. In preventive medicine, VR fosters stress and anxiety reduction, which directly influences patient well-being and quality of life. Participation is enhanced by VR's interactive nature, transforming patients from passive recipients of care to active participants in their treatment journey. Despite its promise, VR faces limitations, including simulator sickness, technical challenges, and accessibility barriers. These factors highlight the need for methodological standardization, improved hardware, and enhanced training for medical professionals. Additionally, a lack of longitudinal studies and safety monitoring systems restricts the widespread clinical adoption of VR. As healthcare systems continue to adapt to technological advancements, VR has the potential to emerge as a pivotal tool in personalized medicine, offering innovative solutions for complex neurological and mental health challenges. By addressing current limitations, VR may redefine patient care and mark a significant milestone in the evolution of evidence-based medical practice.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"49 1","pages":"56"},"PeriodicalIF":3.5,"publicationDate":"2025-04-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143970312","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":"Predicting Visual Acuity after Retinal Vein Occlusion Anti-VEGF Treatment: Development and Validation of an Interpretable Machine Learning Model.","authors":"Chunlan Liang, Lian Liu, Tianqi Zhao, Weiyun Ouyang, Guocheng Yu, Jun Lyu, Jingxiang Zhong","doi":"10.1007/s10916-025-02190-3","DOIUrl":"https://doi.org/10.1007/s10916-025-02190-3","url":null,"abstract":"<p><p>Accurate prediction of post-treatment visual acuity in macular edema secondary to retinal vein occlusion (RVO-ME) is critical for optimizing anti-VEGF therapy and improving clinical outcomes. While machine learning (ML) has shown promise in ophthalmic prognostication, existing models often lack interpretability and clinical applicability for RVO management. This study developed and validated an interpretable ML model to predict visual acuity changes in RVO patients following anti-VEGF treatment. Using retrospective data from 259 RVO patients at the First Affiliated Hospital of Jinan University, we identified key predictive features through the Boruta algorithm and evaluated eight ML algorithms. The Extreme Gradient Boosting (XGBoost) model emerged as optimal, achieving an AUC of 0.91 (95% CI: 0.85-0.96) in the testing cohort with 0.83 accuracy, 0.88 sensitivity, 0.73 specificity, 0.87 F1 score, and 0.14 Brier score. Critical predictors included baseline visual acuity, systolic blood pressure (SBP), age, diabetic retinal inner layer dysfunction (DRIL), and disease subtype. Shapley Additive exPlanations (SHAP) analysis revealed baseline visual acuity as the most influential prognostic factor, followed by SBP and age. Our model seeks to bridge the critical gaps in current research: (1) systematically comparing the applicability and effects of different ML algorithms in RVO-ME visual acuity prediction, and (2) inherent interpretability through SHAP value visualization. The combination of high predictive performance (AUC > 0.9) with inherent clinical transparency may enable the practical implementation of this tool in guiding anti-VEGF treatment decisions. Future validation in multicenter cohorts could further strengthen its generalizability for personalized RVO management.</p>","PeriodicalId":16338,"journal":{"name":"Journal of Medical Systems","volume":"49 1","pages":"57"},"PeriodicalIF":3.5,"publicationDate":"2025-04-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144015714","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}