Most accessed

  • Published in last 1 year
  • In last 2 years
  • In last 3 years
  • All

Please wait a minute...
  • Select all
    |
  • LI Haiyu, FU Yi, BAO Guofeng, ZHAO Jiming, SI Fenggang.
    China Digital Medicine. 2026, 21(5): 1-11. https://doi.org/10.3969/j.issn.1673-7571.2026.05.001
    Medical agents have evolved from the stage of concept validation to clinical application, demonstrating
    significant potential for development and application. This study systematically reviews the current research status,
    technical architecture, development platforms, deployment models, and main application scenarios of medical agents
    both domestically and internationally. It also analyzes their application models and core values in scenarios such as
    patient services, clinical diagnosis and treatment, operation management, and medical research. The study discusses
    the main challenges currently faced and the corresponding strategies. Looking ahead, it is necessary to further
    enhance the reliability of the decision-making capabilities of agents, promote research on multimodal data and multi
    agent collaboration, and establish a sound evaluation and regulatory system for medical agents to truly achieve their
    standardized and large-scale application.
  • TANG Caixing, YANG Wen, WU Zhen, DUAN Mengqi.
    China Digital Medicine. 2026, 21(5): 12-17. https://doi.org/10.3969/j.issn.1673-7571.2026.05.002
    Objective This study aims to construct an online triage and registration assistance system based on multi-modal multi-agent collaborative technology. The objective is to address the current inadequacy of intelligence in online hospital triage and registration, thereby enhancing the efficiency of medical resource allocation and improving the patient experience with online appointment scheduling. Methods The system was designed with a hierarchical architecture, comprising an interaction layer, a processing layer (with four collaborating core agents), and a data layer. Four core agents were developed: tool invocation, report parsing, department recommendation, and doctor recommendation. Based respectively on the Doubao series of models and the DeepSeek-R1-671B model, these agents integrate prompt engineering and a Retrieval-Augmented Generation (RAG) architecture to enable multi-modal information processing and precise recommendations. A controlled experiment on triage performance, a NASA-TLX workload assessment for triage nurses, and a patient Net Promoter Score (NPS) survey were conducted at a Grade A Tertiary hospital in Guangzhou. Results The system achieved a triage accuracy of 96.67%, an increase of 12.88 percentage points compared to manual triage (83.79%) (P < 0.01). The accuracy for doctor recommendations was 83.33%. Following the system's implementation, the NASA-TLX workload index for triage nurses decreased by 8.32 (P= 0.03), and the patient Net Promoter Score (NPS) increased from 12% to 44%. Conclusion The multi-modal multi-agent triage and registration assistance system developed in this study successfully facilitates the intelligent transformation of online medical triage. Its triage accuracy surpasses that of human, effectively reduces the workload for triage nurses, and enhances the patient experience with online registration. This system offers a viable technical solution for the advancement of smart hospitals.
  • JIA Zhihao, LI Junwei, ZHANG Xu, ZHANG Jie, WANG Zhixiang, XU Hao, WANG Lihua
    China Digital Medicine. 2026, 21(5): 18-26. https://doi.org/10.3969/j.issn.1673-7571.2026.05.003
    Objective To improve the objectivity, consistency, and evaluation efficiency of nursing operation assessments, this paper proposes and implements an intelligent nursing operation assessment system based on multi- modal perception and multi-agent collaboration. Methods Taking clinical nursing operation videos as input, a multi- agent collaborative assessment system was constructed, and the assessment standards are converted into executable tasks. A dual-path reasoning agent was used to analyze operation videos and voice verification information respectively, and then a fusion decision-making agent was employed to complete score calculation and report generation. A verification mechanism was introduced to ensure the reliability and availability of the results. Results In various standard nursing operation scenarios, the system could stably realize automatic evaluation, which had high consistency with expert evaluation in terms of process integrity and operation standardization, and significantly reduced the time cost of manual assessment. Conclusion Organizing and applying multi-modal models in a multi-agent collaborative manner can provide an extensible, interpretable and practically applicable technical path for nursing skill assessment.
  • ZHOU Yilin, HU Wei, LIU Wei, XIE Si, LIU Chao, LIU Mailan
    China Digital Medicine. 2026, 21(5): 56-61. https://doi.org/10.3969/j.issn.1673-7571.2026.05.008
    Objective To address the challenges of semantic omissions, complex linguistic structures, and terminological heterogeneity in multi-center integrated Chinese-Western medicine electronic medical record (EMR), this study aimed to construct a DeepSeek-based automated normalization framework to efficiently process large- scale heterogeneous data and the associated complex linguistic phenomena. Methods The DeepSeek large language model was employed via its API in combination with modular prompt engineering. The framework performed explicit completion of key semantic components, ambiguity resolution and sentence restructuring, as well as multi- center clinical terminology alignment on raw heterogeneous EMR texts, thereby enabling their transformation into normalized data that were semantically clear, structurally unified, and annotation-friendly. In addition, targeted content validation was conducted in collaboration with medical experts, and the model outputs were subjected to sampling- based review and accuracy evaluation. Results Under zero-shot conditions, DeepSeek-R1 achieved 86% average accuracy in standardizing knee osteoarthritis (KOA) EHRs from three hospitals. The  "LLM-led standardization with human evaluation"  paradigm significantly reduced manual effort while improving efficiency and quality, demonstrating robust capabilities in semantic understanding and reasoning. Conclusion This framework generates high-quality standardized data, establishing a solid foundation for downstream structured annotation and knowledge extraction, and thereby supporting the construction of high-precision, disease-specific knowledge graphs for integrated Chinese- Western medicine.
  • ZHAO Xia, ZHANG Haibo, ZENG Yi, WEN Birong, HUANG Jianhuan, ZHAO Erkang, LI Xiaohua.
    China Digital Medicine. 2026, 21(3): 1-8. https://doi.org/10.3969/j.issn.1673-7571.2026.03.001
    Large Language Model (LLM) have rapidly gained widespread application in the healthcare field due
    to their powerful language and reasoning capabilities. They play a unique role, especially in applications such as
    the understanding, generation, and processing of Electronic Medical Records (EMRs) that primarily utilize natural
    language, making them a cutting-edge research area in the field of AI + healthcare. This article introduces research
    framework for EMR applications based on LLM, which consists of five components: infrastructure, data resources,
    key technologies, model capabilities, and application scenarios. It also outlines the fundamental principles of each
    component, the key technologies employed, innovative application scenarios, and the development of safety, ethics, and
    compliance systems, providing guidance and references for innovative EMR applications based on LLM.
  • WANG Tingting, LIANG Weixia, HUANG Beili
    China Digital Medicine. 2025, 20(12): 44-48、109.
    Objective To construct the refined management system of the whole-life cycle for medical consumables
    based on unique device identification (UDI), and to address the challenges of extensive management of consumables.
    Methods By upgrading the consumables information management system, the UDI code management module was
    integrated in the supply chain platform, HIS, HRP, surgical anesthesia and SPD systems to promote multi-system
    information interconnection. Qualitative and quantitative research methods were used to compare the management
    efficiency indicators before and after UDI implementation. Results The application of UDI optimized the management
    efficiency of medical consumables in acceptance, inventory, use, charge, settlement, traceability and other aspects, and
    the acceptance warehousing time and billing time of consumables were substantially reduced (P<0.001). Conclusion
    The implementation of UDI system is helpful to realize the closed-loop tracking of the whole process data of medical
    consumables, and provides key support for the refined transformation of consumables management.
  • LIN Yijun, TAN Tao, ZHANG Hua, WANG Xiaoshen
    China Digital Medicine. 2026, 21(5): 34-40. https://doi.org/10.3969/j.issn.1673-7571.2026.05.005
    Objective To develop a large language model-based agent for automated TNM staging of nasopharyngeal carcinoma and to improve the accuracy and consistency of clinical staging. Methods A multi-stage reasoning framework was proposed, comprising information extraction, feature standardization, and guideline mapping. Tool filtering and adaptive context compression were incorporated to improve inference efficiency, and an extensible tool library was established to support staging decisions. The agent operates through a thought–action–observation loop, enabling a transparent and traceable reasoning process. Results In 192 nasopharyngeal carcinoma clinical reports, the agent achieved accuracies of 92%, 74%, and 95% for T, N, and M staging, respectively. Token consumption decreased by 34.2%, 55.4%, and 63.8% under settings with 10, 30, and 50 tools, respectively, with a mean reduction of 51.1%. The agent further generated interpretable reasoning chains, supporting clinician understanding and verification of staging outcomes. Conclusion This large language model-based TNM staging agent offers an effective approach for automating clinical staging, with favorable performance in both accuracy and efficiency. Further validation in multi- center and multi-cancer settings is warranted to determine its generalizability and clinical utility.
  • HUANG Tingting, XIE Junjie, ZHOU Yilin, HU Wei, LIU Wei
    China Digital Medicine. 2026, 21(5): 99-107. https://doi.org/10.3969/j.issn.1673-7571.2026.05.014
    Objective To construct a knowledge graph for the safe use of Traditional Chinese Medicine (TCM) and improve rational clinical medication. Methods Data on Chinese patent medicines were integrated from drug instructions and the Chinese Pharmacopoeia, and an Excel data table was established. This table includes information such as Chinese patent medicines, ingredients, indications, main efficacy, and dosage and administration, and the extracted information was preprocessed. The Label Studio platform was used to create intelligent annotation examples, the API interface of Doubao was called for knowledge extraction, and finally the Neo4j graph database was used for knowledge storage and visual display. Results A TCM safe medication knowledge graph based on Neo4j was constructed, which included 1 702 types of Chinese patent medicines, generating 8 592 entity nodes and 46 409 entity relationships (involving Chinese patent medicines, symptoms, syndromes, etc.). Information query and content visualization could be realized based on the constructed knowledge graph. Conclusion The constructed TCM safe medication knowledge graph can provide auxiliary decision support for clinicians, improve the safety of TCM medication, and serve as a reference for the construction of large-scale Chinese patent medicines knowledge graphs in the future.
  • LIU Huimin, HU Yancen, LIU Zhihua, JI Ping, XU Chang
    China Digital Medicine. 2026, 21(2): 1-9. https://doi.org/10.3969/j.issn.1673-7571.2026.02.001
    The emergence of Multimodal Large Language Models (MLLMs) offers opportunities in the medical
    field, yet the vast quantities of sensitive data they depend on present significant challenges in data governance and
    patient ethics. This research comprehensively reviews literature from Chinese and English databases including CNKI,
    PubMed, and Web of Science from January 1, 2020, to August 31, 2025. It focuses on the core challenges faced
    by MLLMs in medicine, including data fusion, quality control, transparency, and hallucinations, as well as ethical
    dilemmas such as patient informed consent, privacy protection, and algorithmic bias, and analyzes corresponding
    coping strategies.Research indicates that standardizing and quality-controlling heterogeneous medical data from
    multiple sources presents significant challenges, with data silos being a widespread issue. Multimodal fusion training
    tends to amplify algorithmic biases. The “black-box” decision-making and “hallucination” issues of models undermine
    their credibility. Insufficient patient involvement in governance and ambiguous responsibility delineation pose critical
    challenges, while cross-modal associations heighten privacy leakage risks. To address these challenges, it is imperative
    to establish trustworthy data infrastructure, implement multi-stakeholder collaborative governance and accountability
    mechanisms, promote explainable AI, and integrate privacy-enhancing technologies with dynamic informed consent
    mechanisms to safeguard patient rights. This will foster a healthy ecosystem for MLLMs centered on patients, supported
    by technology, and guided by ethical principles.
  • MA Xinyan, CHEN Yu, ZHENG Lin, LI Qiang.
    China Digital Medicine. 2026, 21(5): 62-67. https://doi.org/10.3969/j.issn.1673-7571.2026.05.009
    Objective This study aims to construct a DRG (Diagnosis-Related Groups) full-process management system based on knowledge graphs to address the challenges of refined management in medical institutions posed by DRG payment reform. Methods By integrating clinical pathways with DRG information, a multidimensional medical knowledge graph was constructed using knowledge graph technology. The system includes modules such as pre-admission intelligent decision support, real-time monitoring and intervention during treatment, and closed-loop feedback optimization after discharge, enabling full-process management of DRG. Results After implementation of the system, the admission rate of clinical pathways increased by 24.07%, the optimization rate of pathway versions grew by 17.55%, the variation rate of pathways decreased by 6.49%, the DRG surplus rate rose by 8.03%, and the average length of hospital stay was shortened by 1.02 days. These outcomes indicate that the system effectively integrates medical resources, standardizes treatment processes, improves medical efficiency and quality, and reduces medical costs. Conclusion A DRG full-process management system based on knowledge graphs demonstrates significant effects in enhancing medical efficiency, quality, and economic benefits. It provides strong support for refined management in hospitals under DRG payment reform and has good potential for promotion.
  • LIU Yuhan, CHENG Yaping, YU Siwei, LU Long
    China Digital Medicine. 2025, 20(12): 1-8.
    This study systematically reviews the fundamental concepts and core values of digital therapeutics, and
    comprehensively analyzes its current development landscape both globally and within China, focusing on regulatory
    frameworks, product typologies, and application domains. Furthermore, it identifies the major challenges that constrain
    DTx development, including insufficient clinical validation, limited patient adherence, incomplete reimbursement
    mechanisms, and data security vulnerabilities. Based on existing literature, the paper proposes strategic countermeasures
    to address these barriers and discusses future development trends in technological integration, policy support, and
    market expansion. The review aims to offer theoretical foundations and practical references for the clinical application
    and health policy formulation of digital therapeutics.
  • BAI Ling, ZHANG Shihong, JU Wensheng.
    China Digital Medicine. 2026, 21(4): 1-6. https://doi.org/10.3969/j.issn.1673-7571.2026.04.001
    Combining domestic research cases and practical experience, this study employs literature analysis,
    on-site surveys, and key informant interviews to analyze the current state of medical health data trading. It outlines
    critical stages in the entire process, identifies existing issues, and proposes practical strategies for advancement, aiming
    to promote compliant and efficient circulation of medical health data. Currently, China's medical health data trading
    is gradually forming a development pattern dominated by exchange trading, focused on application scenarios, and
    prioritizing compliance. Pilot regions such as Beijing, Shanghai, Shenzhen, and Guiyang have taken the lead in practical
    exploration, accumulating valuable operational experience in core areas like valuation pricing, data rights confirmation,
    and matching supply-demand pairs, laying the foundation for industry standardization. However, challenges persist,
    including inconsistent data standard systems, imperfect market-based pricing mechanisms, prominent risks of data
    security and privacy breaches, and a shortage of professional, interdisciplinary talent, all of which hinder industry
    efficiency improvements. To address these shortcomings, the study proposes optimization pathways across multiple
    dimensions—enhancing industry standardization, refining value assessment and pricing mechanisms, regulating revenue
    distribution rules, strengthening data security technical safeguards, and cultivating professional talent pools. These
    measures provide actionable practical references and development guidance for hospitals to conduct standardized data
    trading and optimize medical data resources.
  • ZHANG Li, LIU Li, GONG Yanting, YU Zhihao, ZHUANG Yan, LIU Minchao.
    China Digital Medicine. 2026, 21(3): 9-15. https://doi.org/10.3969/j.issn.1673-7571.2026.03.002
    Objective To Build a generative electronic medical record system based on large language models
    (LLMs) and explore its technical implementation path and clinical application value. Methods In the scenario
    of structured medical records, design a standardized template and prompt system using the Qwen3-32B model to
    automatically generate the first draft of medical records. Conduct a comparative analysis between the generated
    electronic medical records and traditionally hand - written medical records and evaluate them in terms of medical record
    completeness, accuracy, writing time, etc. Results LLM-generated electronic medical records received widespread
    recognition for their accuracy, logical coherence, user satisfaction, and efficiency, while there is room for improvement
    in terms of completeness. Conclusion The application of LLM-based generative electronic medical records can
    rapidly generate medical documents, improve work efficiency, and enhance record quality. However, the model's
    flexibility and adaptability may be limited, and its application scenarios still have certain constraints.
  • HUANG Weiqiang, ZHU Mingang, LIAO Qinghong, GU Yunfei.
    China Digital Medicine. 2026, 21(1): 13-19. https://doi.org/10.3969/j.issn.1673-7571.2026.01.003
    With the deepening of the construction of close-knit medical communities, information integration has
    become the key to breaking down the barriers between urban and rural medical services, implementing hierarchical
    medical treatment, and ensuring the homogeneity of health services for residents. The traditional decentralized
    information model has problems such as numerous system silos, fragmented data standards, poor business coordination,
    and discontinuity in the connection of upper and lower-level medical treatments, which are difficult to meet the needs
    of integrated management and coordinated development of "people, finance, materials, and information" in medical
    communities. This study, by reviewing the information integration construction process of the First People's Hospital of
    Jiashan County Medical Community, deeply analyzes the "all-domain planning, all-domain digital intelligence, and all
    domain integration" information architecture and practical path based on new-generation information technology, and
    evaluates the construction effect through quantitative data. The research shows that through phased promotion of system
    integration, construction of a unified information foundation, and reconstruction of all-domain business applications,
    Jiashan County has successfully broken down the data barriers within the medical community, achieved efficient
    sharing of medical resources, and significantly enhanced the service capabilities at the grassroots level, at the same time,
    clarify the collaborative mechanism with the regional health information platform and the distribution models of funds,
    medical insurance and benefits. Providing a referenceable "Jiashan Model" for the high-quality development of close
    knit medical communities in counties.
  • ZHU Wenting, HE Lei, ZHANG Wenbo, SHEN Yuchen, FU Yi, XING Lumin
    China Digital Medicine. 2026, 21(5): 27-33. https://doi.org/10.3969/j.issn.1673-7571.2026.05.004
    Objective To address the inefficiency faced by clinicians in selecting appropriate referral destinations due to high workload, this study aims to design and validate a referral institution recommendation system based on multi-agent collaboration, integrating a two-tower model and a large language model. Methods Utilizing a multi-agent architecture, the system's functions were encapsulated into four agents: perception & preprocessing, two-tower retrieval, deep reasoning, and resource allocation. The core of the system synergized the efficient vector retrieval capability of a two-tower model with the deep semantic reasoning of the DeepSeek-V3 large language model, enabling a "coarse- to-fine" collaborative recommendation process for candidate institutions. Leveraging 13 245 patient records from a tertiary hospital, a final dataset of 1 155 real-world cases with explicit "discharge by physician's advice for transfer" records was constructed after rigorous screening for model training and validation. Recommendation performance was evaluated using metrics including Hit@K, MRR, and NDCG. System response time and robustness were also tested. Results On the test set comprising 10 medical consortium institutions, the system achieved Hit@1, Hit@3, and Hit@5 of 62.9%, 85.3%, and 93.1%, respectively, with an MRR of 0.714 and an NDCG@3 of 0.653. The average end-to-end response time was 53.2 ms. When 30% of structured test indicators or unstructured text fields were randomly masked, Hit@3 decreased marginally by only 0.9 and 2.9 percentage points, demonstrating strong robustness. Conclusion The proposed system delivers accurate, interpretable, and rapid recommendations for referral institutions, effectively narrowing clinicians' decision-making scope and offering a viable technological approach to alleviate the burden associated with the referral process.
  • WU Guanpeng, CUI Xiaodi, LIU Wenqiang
    China Digital Medicine. 2026, 21(5): 108-114. https://doi.org/10.3969/j.issn.1673-7571.2026.05.015
    Shandong provincial third hospital, in alignment with the national guidelines outlined in the "Notice on Accelerating the Large-Scale Deployment and Application of Internet Protocol Version 6 (IPv6)", has implemented an IPv6-based hospital network planning and cybersecurity protection system. Through a comprehensive analysis of the hospital’s operational requirements and existing network infrastructure, the institution has designed a standardized three- layer IPv6 network architecture. By rationally allocating the IPv6 address space provided by internet service providers (ISPs), the hospital has completed IPv6 network planning and deployment. It has laid a network foundation for the informatization development of the hospital, including the construction of internet hospitals, cloud-based medical films, teleconsultation, and examination and test order entry. To mitigate cybersecurity risks such as network attacks, malware infiltration, and phishing threats, the hospital has developed a multi-layered cybersecurity framework. This system integrates intrusion prevention, antivirus protection, web application security, threat intelligence analysis, and other advanced safeguards. These initiatives aim to drive innovation in healthcare IT, ensure stable network operations, and achieve comprehensive cybersecurity protection. This study provides a useful reference for IPv6 network construction, hospital informatization development, and security protection in the medical industry.
  • HOU Jie, LIU Ting, DUAN Congzhe.
    China Digital Medicine. 2025, 20(12): 49-54.
    Objective To explore the application of large language models (LLMs) in the connotative quality
    management of medical records and analyze the facilitative effect of emerging technologies on medical quality
    improvement. Methods Using the "Quality Control Indicators for Medical Record Management (2021 Edition)"as
    the target indicators to be evaluated in this study. LLMs were deeply integrated with the clinical decision support
    system (CDSS), and structured desensitized medical data was utilized to train the model's proficiency in medical
    record text comprehension. Through functions such as real-time in-process reminders and natural language interaction,
    physicians were assisted in promptly detecting and rectifying connotative quality defects during the documentation
    process. The clinical application accuracy of LLMs was verified by comparing the connotative quality control defects
    identified by the model with manual adjudication results. Results All eight established connotative quality control
    rules achieved an accuracy rate exceeding 90% during the experiment, and three target indicators were improved, which
    was consistent with the experimental hypotheses. Conclusion LLMs exhibit superior capabilities over traditional
    artificial intelligence (AI) quality control systems in medical record quality management. They contribute to enhancing
    the connotative quality of medical records, strengthening physicians' awareness of documentation standardization,
    significantly alleviating the workload of clinical staff, quality control personnel, and medical record administrators,
    promoting the further optimization of the hospital's medical record quality management system, and aligning with the
    overall goal of high-quality hospital development.
  • WANG Anhong, DONG Jiafan, WANG Dongyang, DENG Liang, LI Pingnan, WEN Jinbiao, PENG Hong, WANG Jiwei.
    China Digital Medicine. 2026, 21(5): 114-120. https://doi.org/10.3969/j.issn.1673-7571.2026.05.016
    Objective To address the code management challenges brought about by the increasing complexity of hospital software systems, and to solve the problems of missing version control, low efficiency in collaborative development, and potential security risks in code storage under the traditional model, a hospital software code management platform that meets the requirements of information innovation should be constructed to achieve standardized management of source code. Methods By analyzing the current state of hospital code management and the adaptation needs for information innovation, a platform architecture centered on containerization technology is proposed. The platform integrates four major functional modules: code hosting, version control, collaborative development, and security auditing. It utilizes domestically produced information innovation ecosystem components to achieve software and hardware adaptation, and ensures the security of the code throughout its lifecycle through a multi-layered security protection mechanism. Results After the platform was deployed in a hospital, the code version conflict rate decreased by 68%, collaborative development efficiency improved by 55%, and the compliance rate of core business system code reached 100%. The security protection system successfully prevented the risks of source code leakage and tampering, demonstrating its technical feasibility and practicality. Conclusion The code management platform based on the information innovation environment effectively addresses the issues of standardization, security, and collaboration in code management during the hospital's informatization process, providing an independently controllable solution for the medical industry. This platform not only significantly enhances the efficiency of code management and the ability to ensure information security, but also offers reusable practical experience for the industry's information innovation transformation and sustainable development.
  • XU Yinghui, CHEN Yangyang.
    China Digital Medicine. 2026, 21(5): 41-48.
    The resource scheduling problem of hospital information systems (HIS) exhibits NP-hard complexity with multiple constraints and objectives, where traditional scheduling methods struggle to meet real-time and optimization requirements. To address the limitations of classical ant colony optimization (ACO), such as slow convergence and proneness to local optima, an improved ACO with an adaptive pheromone management mechanism is proposed. This algorithm integrates a dynamic evaporation coefficient adjustment strategy driven by resource utilization, constructs a multi-objective Pareto optimal search mechanism, and incorporates multi-dimensional heuristic information tailored to hospital business characteristics. Implemented based on a distributed architecture, it achieves standardized integration with HIS. Experimental results demonstrate that compared with traditional methods, the improved algorithm increases convergence speed by 37.9% and resource utilization by 20.5%, while reducing patient waiting time by 33.6%, verifying its effectiveness and practicality.
  • XIE Shirong, YE Qingyun, CHEN Cengceng, RUAN Guochen, YE Zhengqiang, ZHOU Xingtao.
    China Digital Medicine. 2026, 21(4): 88-91. https://doi.org/10.3969/j.issn.1673-7571.2026.04.016
    Objective To explore the application value of large language models in addressing challenges associated with electronic health records (EHRs), such as time-consuming data entry, low quality control coverage, and poor data interoperability. Methods An intelligent system covering the entire process of EHR generation and quality control was constructed based on multi-source data fusion and medical large language models. A practical implementation and evaluation were conducted within the ophthalmology department of the Eye & ENT Hospital, Fudan University. Results Following system implementation, the average documentation time per record decreased from 2.7 to 1.1 minutes. AI-driven quality control achieved 100% real-time, full-coverage inspection. The rate of clinical semantic errors was reduced from 18.68% to 5.19%. Metrics including the rate of top-grade records, standardized diagnosis usage, and 24- hour filing compliance all reached 100%. Conclusion The system significantly enhances the efficiency and quality of EHR documentation, providing an effective pathway for the intelligent transformation of specialised hospitals.
  • LI Qin, MO Lidan, HU Shi, PANG Jing, LIU Ying
    China Digital Medicine. 2026, 21(5): 93-99. https://doi.org/10.3969/j.issn.1673-7571.2026.05.013
    The prevalence of hypertension among young and middle-aged individuals continues to rise, with their low medication adherence becoming a critical bottleneck in blood pressure control. Digital therapeutics, leveraging smart technology, offer innovative pathways for addressing adherence interruptions in traditional management models through dynamic monitoring, personalized interventions, and remote support. However, digital therapeutics still face challenges such as patient privacy protection, variations in technology compatibility, and unclear long-term intervention effects. In the future, it is necessary to integrate artificial intelligence technology to deepen personalized interventions and promote sustainable development through standardized regulation. This article aims to review the research progress of digital therapeutics in the medication adherence of young and middle-aged hypertensive patients, in order to facilitate the application of digital therapeutics among this demographic and provide a theoretical basis for effective and convenient management models to enhance medication adherence.
  • LI Ming, ZUO Xiuran, XU Chang, WANG Yaru, HOU Liping, CHEN Xiao, WEI Shuang.
    China Digital Medicine. 2026, 21(2): 10-16. https://doi.org/10.3969/j.issn.1673-7571.2026.02.002
    Objective To develop a regional tuberculosis AI diagnostic model based on multimodal data fusion,
    aiming to enhance the efficiency of early screening and precision diagnosis for tuberculosis and provide technical
    support for regional tuberculosis treatment and prevention. Methods Multi-source heterogeneous data from the
    Wuhan Population Health Platform were integrated and fused with multimodal data such as medical images, clinical
    texts, laboratory tests, and epidemiological information to establish a regional tuberculosis multimodal database. A
    weighted scoring method was used to quantify the risk of pulmonary tuberculosis in patients. And a comprehensive 
    evaluation system was established based on precision, recall, and F1-score. A multimodal AI diagnostic model was
    constructed. The model underwent phased training, testing, optimization, external validation, and regional large-scale
    validation. Results During the external validation involving 385 patients, the model constructed based on multimodal
    data achieved an F1-score of 0.91, with both recall and precision exceeding 0.9, demonstrating value for preliminary clinical
    screening and auxiliary diagnosis. The model enabled imaging-based diagnosis and treatment recommendation, supporting
    pilot implementation in hospitals and scalable regional deployment. Conclusion The multimodal AI model effectively
    integrates multi-source diagnostic and treatment data, achieving a balance between high sensitivity and high accuracy in
    tuberculosis diagnosis, thereby providing a technical pathway for regional tuberculosis prevention and control.
  • WU Zhiping, LIN Hengmei, SU Shaosheng
    China Digital Medicine. 2026, 21(5): 49-55. https://doi.org/10.3969/j.issn.1673-7571.2026.05.007
    Objective To construct an intelligent guidance platform based on microservices architecture technology, aiming to enhance the service efficiency and quality of hospital physical examination services and promote the development of a smart healthcare service system. Methods The platform was developed using a distributed microservices architecture, domestic database technologies, and an intelligent guidance module incorporating a multi- attribute decision-making model algorithm. Results The platform was successfully deployed and put into operation. Through its multi-channel information dissemination function, it provides users with real-time, precise intelligent guidance for physical examination pathways. The platform significantly improved operational efficiency of physical examination center, reducing average user examination time by 25%. Conclusion The intelligent guidance platform effectively facilitates refined management of physical examination services. Its successful implementation offers a replicable reference model for the intelligent development of smart outpatient services, demonstrating potential for broader application.
  • CHEN Xiaoyun, ZHU Yunya
    China Digital Medicine. 2026, 21(1): 1-6. https://doi.org/10.3969/j.issn.1673-7571.2026.01.001
    Sort out the development trajectory, relevant policies, and theoretical foundations of digital management
    in medical consortium, and clarify the conceptual connotations. Analyze the main problems currently existing in the
    information integration construction of integrated medical consortium. Explain that the primary content of digital
    management in medical consortium includes innovative models, methods, and pathways in patient management, service
    processes, personnel organization, logistical support, operational management, quality control, and other aspects
    empowered by digital intelligent technologies. Key development strategies include developing top-level designs to
    promote regional coordinated development, continuously exploring digital organizational transformations for integrated
    and efficient operations, conducting evaluations of digital management capabilities in medical consortium to foster
    development, effectively utilizing performance evaluation tools for consortium coordination, and facilitating the
    realization of tiered healthcare delivery.
  • GAO Huanhuan, WANG Xi.
    China Digital Medicine. 2026, 21(5): 68-74. https://doi.org/10.3969/j.issn.1673-7571.2026.05.010
    Objective To optimize the functionality of the "Junzihao-1" outpatient system, reduce the workload of medical staff, minimize unnecessary patient waiting time, and improve the operational efficiency of the outpatient system. Methods Conducting analysis of existing outpatient workflows to identify optimizable stages, and leveraging Web service technology to upgrade the outpatient system and optimize processes. Results In a minimal code-intrusive manner, this project successfully completed the upgrade of the outpatient system and optimization of workflow processes. By implementing batch processing for repetitive tasks and automation of manual operations, we achieved efficient data circulation and significantly enhanced the operational efficiency of outpatient services. Conclusion Leveraging the low-coupling characteristics of Web service technology. The upgraded features enable independent development, maintenance, and deployment, demonstrating excellent scalability and maintainability. This study demonstrates the pivotal role of Web service technology in hospital outpatient system workflows, laying the technical groundwork for the institution's digital transformation, intelligent upgrades, and high-quality development.
  • LYU Hang, XIAO Qingying, LI Peng, YU Guangjun.
    China Digital Medicine. 2026, 21(2): 32-37. https://doi.org/10.3969/j.issn.1673-7571.2026.02.005
    Objective To resolve the structural dilemma of the "quality-efficiency-flexibility" impossible trinity
    in urban primary healthcare, construct a digital service platform integrating multimodal large models and multi-agent
    collaboration, and enhance the service capabilities of family doctor teams as well as residents' health engagement.
    Methods By leveraging the Transformer-XL cross-modal alignment encoder and federated learning technology, we
    developed the Hua Tuo GPT-AI Family Doctor System. Utilizing a "three-end collaboration" architecture, the system
    achieves closed-loop management of data collection, intelligent analysis, precise intervention, and feedback. Using 1.817
    million contracted residents in Longgang District, Shenzhen, as samples, the system's effectiveness was validated across
    three dimensions: quality and efficiency of medical services, resident engagement and healthcare-seeking behavior, and
    health economics and resource allocation efficiency. Results The health record completeness rate reached 100%, with
    a multimodal medical data interpretation accuracy rate of ≥ 92%. The family doctor contracting rate achieved 90.59%,
    and the annual outpatient follow-up rate increased from 64.99% to 72.66%. The average number of residents managed
    per family doctor team was approximately 1 009. The effective referral rate at the primary level reached 97%, and the 
    average referral time was shortened by approximately 40%. Conclusion This system reshapes the ecosystem of urban
    primary healthcare services, driving the transition from extensive management to refined governance. It provides a replicable
    technical pathway and practical paradigm for the large-scale implementation of digital medicine in primary healthcare.
  • CUI Hai, WANG Meifu.
    China Digital Medicine. 2026, 21(1): 7-13. https://doi.org/10.3969/j.issn.1673-7571.2026.01.002
    Objective To explore an effective path for constructing a regional health service system empowered
    by digital technology, this study focuses on the construction model and practical experience of the closely knit medical
    community in Dongtai City, with an information platform as the core. Methods Using the case study method, the
    construction process of the closely knit medical community in Dongtai City was systematically reviewed, with a focus
    on analyzing its construction plan, framework, and specific applications centered on building an information integration
    platform. The specific practical logic of integrating data resources of medical and health institutions in the region,
    achieving information interconnection and business collaboration through this information platform was explored.
    Results The closely knit medical community in Dongtai City successfully integrated regional medical data resources
    and achieved inter institutional information interconnection and business collaboration by building an information
    platform for the closely knit medical community; It has played a key role in optimizing service processes, improving
    medical quality and efficiency, supporting scientific management evaluation, and building a new health service system
    based on modern information technology. Conclusion The closely integrated medical community construction
    model with informationization as the core in Dongtai City is feasible and effective. Its practical path of integrating
    resources and collaborating businesses through information platforms can provide important practical references for the
    construction of informationization platforms for medical communities in other regions.
  • IANG Jiahui, YAO Pan, HUANG Lei.
    China Digital Medicine. 2025, 20(12): 19-26.
    Insomnia, as the most common sleep disorder, faces challenges in traditional psychological treatment
    due to resource constraints, high costs, and limited accessibility, making it difficult to meet clinical demands. Digital
    therapeutics for insomnia (DTI) has consequently emerged, leveraging technologies such as mobile applications and
    online platforms to transform diverse application formats or theoretical treatment methods into flexible and accessible
    digital intervention solutions. DTI has a significant effect on improving insomnia and is cost-effective. However,
    the field still faces a series of challenges, including limited applicability across specific populations, suboptimal
    patient adherence, insufficient data security and regulatory frameworks, a lack of long-term efficacy evidence, and a
    homogeneity in intervention models. Future efforts should prioritize the creation of diversified intervention pathways,
    establish stepped-care models, and strengthen policy support and standardization efforts. These measures are essential to
    achieving sustainable clinical utility and commercial value for DTI.
  • LIU Jin, WANG Yi, WANG Qin, WANG Yuxiang, HE Lan, JI Lijun, WANG Fei, XIONG Minchao.
    China Digital Medicine. 2026, 21(5): 75-82. https://doi.org/10.3969/j.issn.1673-7571.2026.05.011
    Objective To construct a low-dose CT lung cancer screening model based on Artificial Intelligence (AI) and radiomics, and to verify the model. Methods A total of 460 patients with pulmonary nodule in Ezhou Central Hospital from January 2020 to June 2025 were collected and divided into training set (322 cases) and validation set (138 cases) by a ratio of 7:3. According to the pathological results, 322 patients in the training set were divided into benign group (217 cases of benign nodules) and malignant group (105 cases of malignant nodules). The clinical data of the two groups were compared. LASSO regression was used to screen the radiomics features, and multivariate Logistic regression was used to analyze the influencing factors of lung cancer. The diagnostic models of Logistic Regression (LR), Support Vector Machine (SVM), eXtreme Gradient Boosting (XGBoost), and k-Nearest Neighbors (KNN) were constructed in the training set and validation set, and the Receiver Operating Characteristic (ROC) Curve was used to evaluate the efficacy of the models. Results There were statistically significant differences in pleural indentation sign, vascular convergence sign, bronchial sign, spiculation sign, Cyto-keratin 19 fragment antigen 21-1 (CYFRA21-1) and Carcinoembryonic Antigen (CEA) between malignant group and benign group (P<0.05). LASSO regression was used to screen 7 radiomic features and construct a radiomic score. Pleural indentation sign (OR=4.843, 95%CI: 2.685~8.737), vascular convergence sign (OR=7.586, 95%CI: 3.658~ 15.732), bronchial sign (OR=4.007, 95%CI: 2.257~7.115), spiculation sign (OR=3.014, 95%CI: 1.659~5.477), CYFRA21-1 (OR=1.480, 95% CI: 1.136~ 1.927), CEA (OR=1.399, 95% CI: 1.237~ 1.582) and radiomic score (OR=2.076, 95% CI: 1.584~2.721) were all independent influencing factors for lung cancer (P<0.05). The results of training set and validation set showed that the AUC value of XGBoost model was the largest, which was significantly larger than that of LR, SVM and KNN models. Conclusion There are many influencing factors of lung cancer. Based on the above factors, a lung cancer screening model is constructed, and the XGBoost model has good diagnostic efficacy.
  • XIANG Peng, FANG Qichuan, LEI Jianbo
    China Digital Medicine. 2025, 20(12): 9-18.
    Objective To analyze the characteristics of approved Digital Therapeutics (DTx) products in the US,
    Germany, and Belgium, providing references for the development of China's DTx industry. Methods Data were
    sourced from the US FDA, German DiGA directory, and Belgian mHealthBELGIUM platform. Devices meeting the
    DTx definition were identified based on inclusion and exclusion criteria, and their characteristics were analyzed using
    descriptive statistics. Results By the end of 2024, a total of 272 DTx devices were included (US: 192; Germany: 55;
    Belgium: 25). Following a period of fluctuating growth, the approval rates in all three countries showed a slowdown
    in the past year. US products primarily focused on chronic disease management, largely leveraging mobile health 
    technologies. German products were concentrated on treating mental and behavioral disorders, predominantly
    using cognitive behavioral therapy. Belgian products emphasized health status monitoring in oncology and chronic
    diseases. Conclusion China should draw on these international experiences to improve its regulatory and payment
    systems, strengthen real-world evidence and clinical value verification, and promote the standardized and sustainable
    development of the DTx industry.
  • ZHANG Xinlong, WANG Jianlin, GUANG Qi, LIU Hao, DING Xueqian, YANG Zi, LIANG Zhixing.
    China Digital Medicine. 2026, 21(1): 110-115. https://doi.org/10.3969/j.issn.1673-7571.2026.01.018
    Objective To address the insufficient real-time performance and low recovery efficiency from logical
    failures in hospital core database disaster recovery systems, a disaster recovery and backup integrated architecture based
    on continuous data protection (CDP) is designed to overcome the limitations of traditional dual-system hot standby
    solutions in meeting medical business continuity requirements. Methods A “dual-active–backup–drill” integrated
    architecture is proposed. By leveraging CDP real-time I/O capture technology, a hybrid FC SAN/iSCSI networking
    mode, and heterogeneous database adaptation strategies, a disaster recovery system is constructed that supports zero
    data loss, minute-level recovery, and automated disaster recovery drills. Results Zero data loss (RPO = 0) was
    achieved for the hospital core database. Physical failure recovery time was reduced to less than 10 minutes (RTO <
    10 min), logical failure recovery time to less than 1 minute (RTO < 1 min), and the disaster recovery drill cycle was
    shortened from 3 days to within 10 minutes. Conclusion In the context of hospital core databases, CDP technology
    significantly improves data security and business continuity. The proposed “dual-active–backup–drill” integrated
    architecture provides a replicable paradigm for disaster recovery construction in similar systems.
  • ZHENG Rongwei, SU Liangbo, HUANG Rong, DONG Jiafan, WANG Dongyang, CHEN Gang, WANG Jiwei.
    China Digital Medicine. 2026, 21(1): 83-88. https://doi.org/10.3969/j.issn.1673-7571.2026.01.014
    Objective In order to improve the independent and controllable ability of hospital information
    technology application, all business system databases, including core businesses related to medical and health, will
    be replaced with domestic databases. Methods Establish an environment, import all database objects and test data,
    complete performance testing, then apply a synchronization tool to synchronize the source and target databases. Finally,
    a shutdown cutover is performed to replace with the domestic databases. Results After a 110-minute downtime, the
    original 89 business system was fully migrated to the DM8 database, completing the localization and replacement of the
    database. Conclusion Since the localization replacement of the database was completed, it has been operating stably
    for over 15 months. Practice has proven that the localization of databases for core business systems in the medical
    industry is entirely feasible.
  • HU Wanfei, WU Xusheng, WU Sizhu, LIU Xiayang, DUAN Yongheng, QIAN Qing
    China Digital Medicine. 2026, 21(2): 59-66. https://doi.org/10.3969/j.issn.1673-7571.2026.02.008
    Federated learning can achieve the collaborative computation of multi-source data on the premise of
    protecting data privacy, providing a new way to solve the problem of medical data sharing. This paper reviews the latest
    progress of federated learning in medical data sharing and points out that its core role is to achieve the secure sharing
    of medical data through mechanisms such as local training and parameter update, encrypted sharing and fuzzy sharing.
    In the fields of disease prediction, medical image analysis and drug research and development, federated learning has
    shown great potential in enhancing model prediction performance, promoting personalized medicine and accelerating
    drug discovery. However, the application of federated learning in the medical field still faces challenges such as privacy
    and security issues, standardization and interoperability problems, data quality problems, difficulties in combining
    expert knowledge, incentive mechanisms and data sharing willingness, and efficiency issues. In the future, federated
    learning is expected to play a more important role in promoting medical data sharing and promoting the realization of
    more accurate, efficient and safe medical services.
  • QUE Jianren, LING Sikai.
    China Digital Medicine. 2025, 20(12): 55-61.
    Objective It effectively integrated and utilized medical imaging data across various hospital campuses,
    established a unified management standard for such data, so as to build a secure and efficient medical imaging data
    integration and intelligent diagnosis platform. Methods Based on a privatized platform architecture, distributed storage
    technology was adopted to achieve unified management of petabyte-level imaging data across multiple campuses. The
    DICOM standard and FHIR protocol were utilized for standardized and fusion of multimodal data (CT, MRI, ultrasound,
    endoscopy, nuclear medicine, etc.). Blockchain technology was implemented to ensure data security and privacy
    protection. Deep learning-based intelligent diagnosis modules (3D ResNet and Transformer models) were deployed to
    provide lesion detection, segmentation, and assisted diagnosis functions. Results The platform successfully integrated
    imaging data from hospitals within the Ruijin Medical Consortium, enabling second-level retrieval of cross-campus
    imaging data. The intelligent diagnosis module achieved over 96% accuracy in detecting typical conditions such as
    pulmonary nodules, stroke, and fractures, with performance exceeding traditional methods by over 15%. Conclusion
    The platform enables efficient fusion and intelligent application of multi-campus medical imaging data, significantly 
    enhancing clinical diagnostic efficiency and scientific research capabilities. It offers a scalable solution for regional
    medical resource sharing and smart hospital construction.
  • ZHENG Wanting, TONG Yuanyuan, CHAI Jiaqi, ZHU Ling
    China Digital Medicine. 2026, 21(2): 85-94. https://doi.org/10.3969/j.issn.1673-7571.2026.02.012
    Objective To address the strong subjectivity in Traditional Chinese Medicine (TCM) syndrome
    differentiation and the lack of dynamic responsiveness in traditional formula recommendation, this study proposes a
    deep reinforcement learning decision model that integrates syndrome prediction confidence, designing a self-adaptive
    and optimized personalized TCM diagnosis and treatment system. Methods Based on 5 970 clinical cases of four
    diagnostic methods data, a Deep Neural Network (DNN) syndrome prediction model was constructed, using the SMOTE
    (Synthetic Minority Over-sampling Technique) algorithm to address syndrome class imbalance. The syndrome feature
    vector and its prediction confidence were combined to form the state space for a Deep Q-Network (DQN). The decision
    strategy was dynamically optimized through a reward function, incentivizing exploration in low-confidence states and
    enabling precise formula recommendation in high-confidence states. Results The DNN syndrome identification model
    achieved an accuracy of 91.1% (macro-F1 90.9%), significantly outperforming traditional methods. The confidence
    driven DQN system achieved an intelligent formula selection rate of 68.9%, with the comprehensive score rising to 0.658. 
    Conclusion The "Syndrome Confidence-DQN" collaborative decision-making paradigm overcomes the limitations of
    traditional static formula recommendation. It provides a quantifiable, adaptive, and dynamic decision-making tool for
    intelligent TCM diagnosis and treatment, promoting the transformation of clinical practice from experience-based rules
    to data-driven intelligence.
  • LAI Yishan, HUANG Yihao, ZENG Xingxin, ZHANG Zhigang, XU Yisha, MENG Chenxing
    China Digital Medicine. 2026, 21(5): 82-92. https://doi.org/10.3969/j.issn.1673-7571.2026.05.012
    Objective To evaluate the clinical value of an improved machine learning model for predicting spontaneous hemorrhagic transformation (HT) risk after acute ischemic stroke (AIS) and provide evidence for clinical decision-making. Methods A multicenter retrospective study included 260 AIS patients admitted to three hospitals from January 2023 to December 2024. Structured electronic medical records were used to extract multidimensional data: demographic characteristics, comorbidities, risk factors, laboratory indicators, and neuroimaging features. The dataset was split into training (n= 208) and test (n = 52) sets at an 8:2 ratio. Feature subsets, base learners, and hyperparameter combinations were optimized using the Improved Blood-Sucking Leech Optimizer (IBSLO) algorithm. Feature rationality was validated via Least Absolute Shrinkage and Selection Operator (LASSO) regression and SHapley Additive exPlanations (SHAP) interpretability analysis. Results The IBSLO algorithm significantly outperformed other algorithms in both global optimization performance and convergence efficiency. The Auto Machine Learning (AutoML) model guided by IBSLO achieved an Area Under the Receiver Operating Characteristic Curve (ROC-AUC) of 0.903 8 and an Area Under the Precision-Recall Curve (PR-AUC) of 0.809 1 on the training set, and 0.877 5 (ROC- AUC) and 0.817 5 (PR-AUC) on the test set, significantly surpassing traditional machine learning models. Decision Curve Analysis (DCA) indicated that applying the AutoML model to predict HT risk could yield greater clinical net benefit compared to traditional methods. SHAP analysis revealed the top three core predictors as age, atrial fibrillation, and low-density lipoprotein (LDL). Conclusion The improved machine learning model enables early prediction of post-AIS spontaneous HT risk and supports therapeutic optimization.
  • ANG Caixing, RONG Li, WU Zhen, DUAN Mengqi.
    China Digital Medicine. 2026, 21(4): 33-38. https://doi.org/10.3969/j.issn.1673-7571.2026.04.006
    Objective This study aimed to compare the applicability of reasoning models and general models in nursing decision support system, providing evidence-based guidance for the construction of clinical nursing decision- making frameworks. Methods A unified framework for the nursing decision support system was developed, allowing the switching between reasoning model (DeepSeek-R1) and general model (GLM-3-Turbo) to complete three tasks: nursing plan generation, real-case analysis, and postgraduate entrance examination question answering. System operational efficiency was quantitatively recorded, and the quality of the decision-making content was evaluated by two deputy-level nursing experts using a ten-point scale. Results The reasoning model demonstrated the higher content quality in nursing decision support (scores for the three tasks: 9.22±0.76, 8.97±0.83, 8.70±1.17), but the runtime efficiency is low (response time: 55.40±21.39 seconds, median token consumption: 1 716 (1 367, 2 934); the general model has acceptable accuracy (three tasks: 8.35±1.31, 7.60±0.89, 6.90±1.41), and its runtime efficiency is much higher (response time: 7.87±3. 14 seconds, median token consumption: 687 (556, 2 234). Conclusion Priority should be given to the reasoning model to ensure high accuracy in complex nursing decision-making tasks. The general model, combined with a knowledge base, can serve as an efficiency-enhancing tool for routine nursing decision-making tasks. 
  • MA Liming, HUANG Shaobin, JIANG Zhuobin, HUANG Guoxing, ZHANG Xinghua.
    China Digital Medicine. 2025, 20(12): 93-98.
    Objective This paper provides a practical reference for constructing a network security system for
    medical institutions facing challenges in multi-campus and diverse information systems environments. Methods
    The system adheres to the “Information Security Technology-Baseline for Classified Protection of Cybersecurity”
    standard, employing a refined partitioned and zoned network architecture, combines high performance with highly
    available critical path design, integrates multi-level mature security protection technology and normalized network
    security operation, and constructs a unified security assurance system for multiple hospital areas. Results The
    system's effectiveness has been validated through multiple attack-defense drills, which promotes internal management
    and technical upgrading. Conclusion The successful establishment of an integrated architecture for network and
    information security across multi-campus hospitals enables centralized management of security incidents, enhances
    business continuity and security, and provides reference for the construction of cybersecurity systems in healthcare
    institutions.
  • LIU Yi, WANG Wenrui.
    China Digital Medicine. 2026, 21(4): 98-105. https://doi.org/10.3969/j.issn.1673-7571.2026.04.018
    Objective To delineate the technological development trajectory of Clinical Decision Support Systems (CDSS) in China using bibliometric methods, and to analyze application hotspots and evolutionary characteristics across diverse clinical domains, thereby providing references for optimizing system design and practical implementation. Methods Bibliometric analysis was conducted using CiteSpace software on core journal articles related to CDSS published between 2009 and 2024, retrieved from the China National Knowledge Infrastructure (CNKI), Wanfang Database, and VIP Database. Keyword co-occurrence, clustering, and burst detection analyses were performed. Results A total of 720 articles were included. The publication volume of CDSS research exhibited a three-phase growth pattern. High-frequency keywords included artificial intelligence, nursing, and electronic health records. Research focused on three key dimensions: advancements in underlying technologies, exploration in the management of multiple disease types and the full disease course, and intelligent nursing practices. Conclusion Future efforts should prioritize overcoming bottlenecks in multi-source heterogeneous data integration, dynamic knowledge base updating, and cross- modal technology collaboration. Additionally, cultural adaptability and ethical compliance must be addressed to advance CDSS from decision support to comprehensive intelligent transformation across the entire clinical workflow.
  • ZHANG Zhelai, TANG Jingyun, ZHU Jie.
    China Digital Medicine. 2025, 20(12): 110-115.
    Objective Symptom surveillance is of great significance for the early warning of infectious diseases
    and public health emergencies. The existing symptom surveillance systems still rely on rule-based natural language
    processing technology, which is difficult to obtain high precision symptom recognition rate and affects the reliability
    and final effect of symptom surveillance. Methods A symptom monitoring system was designed and constructed
    based on the Suzhou Universal Health Information Platform. This system applied the large language model technology
    to the task of symptom entity recognition, specifically to identify symptoms in outpatient and emergency electronic
    medical records by fine-tuned BERT pre-training model. Results The introduction of large language model technology
    greatly improved the recall rate and accuracy rate of symptom recognition. Conclusion This system realized the high
    precision symptom monitoring of T+1 day in outpatient and emergency departments of public medical institutions in the
    city