Objective To design and develop an intelligent monitoring and early warning front-end software for infectious diseases based on the integrated deployment of medical institutions and in accordance with the national unified standards, so as to enhance the efficiency and accuracy of infectious disease monitoring and early warning.Methods Through the analysis and research of the existing infectious disease monitoring and early warning system,a national infectious disease intelligent monitoring and early warning front-end software was designed and developed by adopting a highly flexible system architecture and cross-platform development technology with strong adaptability, achieving efficient and seamless integration of various functional modules. Results In the actual pilot application, this software improved the timeliness and accuracy of data collection. Compared with the manual online direct reporting, the reporting rate of infectious diseases increased by 10.7%. Conclusion The intelligent monitoring and early warning front-end software for infectious diseases is an important software tool and technical means to integrate the hospital information system, which can achieve the major transformation of the infectious disease monitoring mode from the traditional passive reporting to modern active perception, providing strong support for the monitoring and control of infectious diseases in China.

This paper provides an overview of the types status, challenges and potential directions of machine learning application in clinical research. This paper focuses on the application of machine learning in disease risk prediction, disease diagnosis, inspection and examination, medical record writing and quality control, discusses the challenges faced by machine learning in clinical application, such as ethical supervision, privacy protection and data security, and puts forward the development direction of machine learning in clinical research.

As a foundational strategic resource for the nation, healthcare data plays a vital role in safeguarding public health and advancing the development of the medical and health sector through its secure and efficient utilization. Data classification and grading constitute a critical component in healthcare data security governance and the capitalization of data assets. This paper elaborates on the conceptual framework, objectives, and significance of data classification and grading. It provides an in-depth analysis of the background, necessity, unique characteristics, challenges, and regulatory trends pertaining to data classification and Grading in the healthcare industry. Key technologies and core operational workflows are systematically introduced, followed by an examination of practical implementation strategies within medical and health institutions. Furthermore, recommendations for enhancing data classification and grading capabilities are proposed. Finally, the limitations of this study are identified.

This study investigates the research progress and application practices of AI technologies in the field of cybersecurity. It traces the early exploration paths of AI technologies such as neural networks and deep learning in cybersecurity. It also expounds on AI-driven novel cybersecurity applications from four dimensions: penetration testing, code auditing, threat intelligence, and large models for vertical security domains. Subsequently, it conducts an in-depth analysis of the challenges faced by AI in cybersecurity, illustrating the security issues of AI in sensitive industries with a case study of the medical industry. It forecasts the development trends of the deep integration of AI and cybersecurity, providing references for promoting cyberspace governance capabilities.

This study aims to investigate the application of large language models (LLMs) and retrieval-augmented generation (RAG) technology in the hierarchical classification of medical data, addressing key challenges in medical data governance, including the processing difficulties of structured, semi-structured, and unstructured data, inefficiencies in sensitive data identification, and delays in dynamic knowledge updates. By harnessing the advanced capabilities of LLMs in processing large-scale datasets and the continuous algorithmic optimization afforded by deep learning, we enhance the intelligence of decision-making and predictive performance in data classification and grading engines. Furthermore, the integration of RAG as a localized knowledge base improves the accuracy and reliability of DeepSeek's applications in medical data scenarios through enriched retrieval of domain-specific knowledge. Through iterative advancements in model architecture, retrieval techniques, and self-learning mechanisms, the proposed framework achieves an 80% improvement in classification efficiency and elevates recognition accuracy to over 90%, significantly optimizing the precision and operational efficacy of medical data classification and grading. The proposed methodology offers a highly interpretable and dynamically adaptive technical solution for medical data governance, demonstrating substantial potential for practical implementation in critical areas such as data asset cataloging and secure data sharing.

With the in-depth development of the digital era, a new generation of information technology is stimulating the explosive growth of data in the healthcare and medical sector, which covers many fields such as disease diagnosis, treatment and prevention, health management, public health, and elderly care services. The market potential of healthcare and medical data is huge, but there are still many problems in the circulation and transaction of its assets. By studying and sorting out relevant literature, this paper summarizes the trends in the field of healthcare and medical data asset circulation and transaction in China. Based on relevant theories and practices of data assets, this paper proposes suggestions such as optimizing the confirmation model of healthcare and medical data rights, improving the pricing mechanism, strengthening privacy protection and data security technology, building a data open sharing system with the participation of multiple parties including the government, medical institutions, research institutes and enterprises, achieving the maximum utilization of data assets and the enhancement of social value, so as to promote data sharing and further drive the standardized and efficient development of circulation and transaction in the healthcare and medical data market.

Based on the trend of refined data security management, this study will start from the data governance requirements, integrate the security compliance requirements, and explore the application of data security classification in the level of refined data management. According to the relevant standards of the medical industry, the practice process of classification and classification was explored through business research, asset clarification, data dictionary improvement and standard formulation. After data classification management, hospitals can reasonably allocate medical resources through fine study of specific categories of data; after hierarchical management based on data importance and sensitivity, hospitals can better identify and evaluate the risks of various data, which helps to formulate targeted risk management strategies for different levels and ensure the improvement of compliance of hospitals in laws and industry standards. Classification and grading not only optimize data storage and access, improve management efficiency, but also significantly reduce the risk of data leakage. It not only provides support for the rational allocation of hospital medical resources and scientific research innovation, but also achieves the purpose of comprehensively improving data security management.

The development of big data and artificial intelligence technology has driven an increase in the asset and application value of healthcare data, and the risks of data leakage, tampering and illegal utilization have also increased. China's data security governance has formed a top-level design at the policy level, but data classification and grading in the healthcare field based on security governance is still in the exploratory stage. In this paper, based on the data security governance work of a grade A class three hospital, we explore the theoretical basis, technical points and practical path of data classification and grading, construct a hospital data classification standard that contains 6 categories, 25 subcategories and 44 subcategories, and data grading rules based on level 1~5, complete the data asset sorting of about 120 sets of information systems in the hospital, and establish a variety of algorithmic combinations to realize the classification and grading automation, and the algorithmic model embedded in the data. The classification is automated, and the algorithm model is embedded in the data security monitoring service platform, which effectively improves the capacity and efficiency of hospital data security governance by real-time monitoring of the systematic distribution and flow of different types and levels of data.

Establish acceptance standards for hospital informatization construction projects, standardize the acceptance work of informatization construction projects in secondary and above hospitals, and improve the quality of informatization project construction and application effects. Through research methods such as literature research, standard and policy research, model construction, field investigation, and expert consultation, a set of acceptance standards for hospital informatization construction projects has been formulated. This study breaks down hospital informatization projects into three major dimensions: infrastructure, software construction, and services. Different acceptance quality assessment models are constructed for their distinctive features, and dynamic acceptance process norms are also provided. This research outcome not only provides a reusable acceptance methodology for hospital informatization construction, but also promotes the standardized development of the medical industry from the top-level design perspective. By strengthening standard constraints and process supervision, it can effectively avoid construction risks, improve resource utilization efficiency, and provide key support for the construction of smart hospitals, demonstrating significant industry demonstration value.

Objective To construct a doctor-patient dialogue abstract generation system based on large language model, which can automatically extract and summarize key medical information from doctor-patient dialogues without any annotated data. Methods The doctor-patient dialogue data were pre-labeled with a larger-scale large language model, and then the large language model with small parameter was trained based on the pseudo-parallel data. In the reasoning stage, contextual learning method was introduced, which provided a few examples integrated with command engineering, enabling the large language model to understand the doctor-patient dialogue more accurately and generate the final summary. Results The system constructed in this study significantly outperformed the existing unsupervised summarization techniques and large language models in retaining key medical information. Conclusion The ability inheritance of the small-parameter model to the large-parameter model can be realized by using the knowledge distillation method to pre-label the doctor-patient dialogue, so as to reduce the dependence on training data and improve the generality and portability of the model.

Objective Relying on the platform of TCM evidence-based medicine research of Guang'anmen Hospital, China Academy of Chinese Medical Sciences, a standardized and scientific TCM special disease database of pulmonary nodules was established, so as to provide an important approach for clinical medicine to focus on the dominant diseases of TCM and to explore high-quality evidence. Methods Taking the population with pulmonary nodules as the research object, the clinical data and information of them were collected and summarized. Through AI technologies such as natural language processing, the original data were desensitized, encrypted and cleaned to achieve high degree of standardization of disease-specific data. Thirteen structured disease-specific data sets of pulmonary nodules were constructed. Results A specialized TCM clinical database for lung nodules has been established.This database contains the medical data of 43,976 patients from January 2017 to June 2023, and realizes a series of functions such as data retrieval, database updates, TCM clinical data analysis and visualization, as well as follow-up investigations. It can accurately analyze the target population and be effectively applied to the research of therapeutic effect evaluation, providing strong data support for clinical research and TCM practice. Conclusion The pulmonary nodule disease-specific database has established the first longitudinal time distribution model of pulmonary nodules incorporating TCM data, which opens up an important path for in-depth epidemiological research of pulmonary nodules and the discovery of high-quality evidence. At the same time, this disease-specific database also can provide valuable data support and reference for the construction of a standardized, evidence-based medical research system with characteristics of both Chinese and Western medicine.

To develop a medical data asset management platform, improve the value and utilization efficiency of medical data through data standardization, capitalization and productization, and support the needs of medical treatment, scientific research, management and other aspects. The platform design includes data discovery layer, data resource layer, data asset layer, data product layer, data service layer, and general support layer. It adopts data standardization, privacy computing, blockchain evidence storage, and AI large model to realize the whole lifecycle management of data. Results The application of the platform significantly enhanced the data management efficiency and data quality of medical institutions, promoted data integration and sharing, optimized the allocation of medical resources, and accelerated medical research innovation. The construction and application of the platform have significantly enhanced the data governance ability and informatization level of medical institutions, effectively addressed issues such as data silos, inconsistent data standards, and insufficient data security and privacy protection, which can provide solid support for improving the quality of medical services and research level.

As a powerful natural language processing model, ChatGPT demonstrates broad application prospects in the field of clinical medicine. This article summarizes the current clinical application status of ChatGPT in assisting clinical decision support, nursing practice optimization, medical intelligent Q&A, medical documents generation, etc. It also explores the limitations of generative pre-trained transformer (GPT) models represented by ChatGPT in data timeliness, understanding of medical terminology, medical risk assessment, legal risk and liability determination, and provides corresponding application standard suggestions. Finally, the potential future developments of ChatGPT in personalized healthcare services, drug research, and telemedicine are prospected, aiming to promote the deep integration of artificial intelligence technology with clinical medicine.

Medical insurance plays an important role in chronic disease management. By sorting out and summarizing the relevant policies on the application of digital technology in chronic disease management in China, discussing the application of digital therapy in the field of chronic disease at home and abroad, and its medical insurance payment policy, medical insurance development suggestions for chronic disease digital therapy are proposed, aiming to provide perfect support for the digital development of chronic disease management in China.

Realize the intelligent early warning of security events, and solve the problems of low inspection efficiency, insufficient video data application and analysis, passive event detection and lagging disposal in the traditional security work of hospitals. Based on the AI visual analysis center as the foundation of the entire system, fully utilizing technologies such as full target structured analysis of video scenes, AI parsing of video, AR panoramic technology, intelligent algorithm warehouse, etc., to meet the three major functional areas of key personnel prevention and control, intelligent security control, and refined comprehensive management of hospital security. The system has achieved precise targeting of targets, foreseeing the occurrence of security incidents, overall control of security situations, and scientific acquisition of security data in security management. The system can improve the efficiency of hospital safety management, assist in the refinement and intelligent upgrading of hospital management, and promote the high-quality development of hospitals.

Objective Information technology infrastructure refers to the network, computing, and storage resources that provide basic support for the operation of information systems, mainly including cabinets, information networks, servers and storage devices, security facilities, etc. With the rapid development of medical science and information science, especially the deep integration and application of new generation information technologies such as artificial intelligence and big data with medicine, new requirements have been put forward for information technology infrastructure. Methods Based on emerging information technology applications such as artificial intelligence and big data, oriented the construction of smart hospitals, high-level interoperability, deep application of AI and artificial intelligence, a new generation of hospital information technology infrastructure integration system covering modular data centers, information networks, servers, storage devices, and security facilities is proposed. Results The overall architecture and technical principles and methods of each component of the new generation hospital information technology infrastructure were proposed, and detailed descriptions were made on AI server GPU, storage devices, data classification and grading management, and IT application innovation. Conclusion Providing a new generation of hospital information technology infrastructure architecture and methods for the transformation and development of hospital digitization. 

Objective Computer network system is the key information infrastructure of smart hospital, which needs 7×24 hours uninterrupted operation to support the all-weather business needs of the hospital. Methods Based on the open network communication protocol, network virtualization and other technologies were adopted to transform the hospital network system and architecture, and a unified intelligent network operation and maintenance management platform was established to links users, terminals, network elements and other resource management mechanisms, to automatically monitor the operating status, independently defend against attacks, and proactively discover and quickly dispose the faults. Results The combined application of network technology optimized the traditional operation and maintenance mode, and improved the controllability and protection level of the network system. Conclusion Intelligent network operation and maintenance management can reduce the risk of failure and operation and maintenance cost, improve the quality and efficiency of operation and maintenance, and ensure the continuous, stable, efficient and safe operation of the hospital network system.

Objective To achieve desensitization of electronic medical documents during sharing, and to protect patient privacy. Methods A lexical analyzer for medical data integrating multiple machine learning models was constructed to sort out Chinese word segmentation, part-of-speech tagging, and named entity recognition corpora in the field of medical and healthcare. Sensitive information in electronic medical documents was identified by using natural language processing technologies such as Hidden Markov Models and Conditional Random Fields and built-in sensitive information signature library, and dynamic desensitization was realized through result set streaming processing technology. Results The algorithm model has a good effect on the processing of routine sensitive personal information, with an average time of detection and desensitization of sensitive personal information was milliseconds. Conclusion The method of natural language processing with sensitive information signature library can realize the recognition and real-time desensitization of sensitive information in unstructured electronic medical documents.

Objective To address the challenges modern scholars face in comprehending and utilizing the medical wisdom embedded in ancient books of traditional Chinese medicine due to language evolution and cultural disparities, this study aims to construct a large-scale parallel corpus for the translation of ancient Chinese medical literature from past to present. Method By selecting representative ancient texts, conducting digital processing of the manuscripts, and precisely aligning the original texts with their modern translations, a large-scale parallel corpus of traditional Chinese medicine literature has been established. Results A large-scale parallel corpus for ancient and modern translations  of ancient books of traditional Chinese medicine has been successfully established, and experimental evidence has demonstrated the effectiveness of specialized corpora in enhancing translation outcomes.Conclusion This corpus holds significant application value in the fields of Chinese medicine research and natural language processing, contributing to an in-depth exploration of the wisdom found in ancient book of traditional Chinese medicine and promoting the modernization of Chinese medicine.

Objective To propose a data-driven predictive maintenance method for hospital medical equipment failure records, so as to optimize equipment management strategies. Methods Effective fault records of medical equipment in a hospital were collected and cleaned. Weibull distribution was used to model the fault interval data of each equipment, and the Maximum Likelihood Estimate method was used to fit the model parameters. The Kolmogorov-Smirnov (K-S) test was then employed to screen out models that fit the actual distribution. Finally, the Mean Time Between Failures (MTBF) and the failure probability of each device were calculated based on the failure interval distribution model. Results A dynamic database for equipment failure prediction was established, which included features such as failure probabilities of each device. By the end of 2023, 99.63% of all medical equipment in the hospital had failure intervals consistent with the Weibull distributions. Conclusion The model based on Weibull distribution can effectively reflect the distribution characteristics of failure interval of each medical equipment, and are of great significance for making reasonable predictive maintenance strategies for medical equipment.

To investigate methods and technologies based on datasets and scenarazation for achieving effective practice and efficient application of health data classification and grading. We analyzed the challenges and requirements in health data management and explored methods, technologies and applications for health data classification and grading based on datasets and scenarazation. Practical implementation in the classification and grading of electronic medical records and electronic health records in public health specialty hospital demonstrated that the proposed methods and technologies based on datasets and scenarization are scientific and effective. The health data classification and grading methods and technologies based on datasets and scenarazation provide a compliant and efficient foundational framework for data security, data circulation, and the capitalization of data elements.

Objective To explore and summarize the design scheme of regional cloud imaging platform. Methods Based on the relevant construction standards and norms, the demand analysis was carried out for the pain points in the construction of regional cloud imaging platform. The new infrastructure planning concept based on cloud computing, big data, AI, blockchain and other technologies as the core was adhered to, and the regional cloud imaging platform was built based on the practical experience in the construction of Jiangsu Province's health cloud imaging platform. Results Based on the basic principles of unified collection, unified storage, unified access, unified quality control, unified mutual recognition and unified application, the platform adopted a new generation of information technology to realize the overall standardization of regional imaging data and value-added imaging services. Conclusion The technical base of regional cloud imaging platform is constructed by cloud architecture, which supports the construction of a regional unified medical image access and sharing system, and provides advanced practical experience for the construction mode of regional cloud imaging platform.

As an important component of data elements, medical data elements have become a significant element influencing the high-quality development of medical and healthcare industry. Through literature research and case analysis, this paper reviews the development process of healthcare and medical data elements, proposes the composition and functional positioning of the “healthcare and medical data elements circulation framework system”, elaborates on the challenges faced by healthcare and medical data elements in the three-level transformation process of “data resource data asset-data product”, and puts forward countermeasures and suggestions, aiming to provide useful references for the high-quality development of the medical and healthcare industry in the new era.

Objective To conduct systematic research on the classification and grading of medical and health data, respond to the mandatory requirements of the "Data Security Law of the People's Republic of China", improve the efficiency and security of data management, promote the exploration of data elements, and facilitate the development of medical informatization. Methods Based on national standards such as "Information Security TechnologyGuidelines for Health and Medical Data Security" (GB/T 39725 － 2020) and "Data Security Technology Rules for Data Classification and Grading" (GB/T 43697 － 2024), a data classification and grading system suitable for the medical and health field was constructed. In line with the actual situation of the Second Affiliated Hospital of the Chinese University of Hong Kong (Shenzhen), the data was divided into two major categories: individual identity-related data and medical data, and further classified into five levels. Additionally, the research team developed a classification and grading management tool that integrates functions such as data cleaning, classification, grading, and security protection, to efficiently and accurately process large-scale data. Results Through empirical research, the effectiveness and feasibility of the data classification and grading system and management tools were verified. Conclusion Data classification and grading can not only enhance the availability of data, but also can effectively reduce the risks of data leakage and abuse. The standardized and normalized management of data can promote the interconnection and interoperability among different systems, break down information island, and drive the prosperous development of the data element market.

A study was conducted on the application status of clinical pathways in a grade A class three hospital from 2014 to 2023. Through multi-level and multi-dimensional data analysis and utilization, targeted improvement strategies for clinical pathway implementation were proposed. A comprehensive comparative analysis of clinical pathway application data from 2014 to 2023 was performed. Patient data meeting the criteria for the "age-related cataract" clinical pathway were extracted. A t-test was used to compare hospitalization days and costs between patients who completed the pathway and those who did not. Patients who completed the clinical pathway exhibited significantly shorter hospitalization days and lower hospitalization costs compared to non-completers, with statistically significant differences (P＜0.05). However, existing issues include insufficient awareness of clinical pathways among medical staff, poor integration of information systems, underutilization of information platforms, and a lack of effective supervision mechanisms. Clinical pathways play a significant role in improving medical quality and reducing healthcare costs. To address current challenges, it is essential to strengthen policy advocacy, optimize information systems, enhance platform construction, and refine incentive mechanisms. These measures will promote the advancement of medical service quality in hospitals under the digital medicine framework. 

Objective To predict the mortality risk of patients with acute liver failure and explore the risk factors affecting patient survival. Methods Patients with acute liver failure treated in West China Hospital from January 1,2020 to August 31, 2023 were selected. Feature selection was performed using univariate and multivariate logistic regression, Lasso, and Random Forest. Then, a mortality risk prediction model was constructed using machine learning technologies such as Random Forest and XGBoost, and the model performance was evaluated using clinical decision curve analysis and SHAP analysis. Results A total of 667 patients were ultimately included in this study, with a mortality rate of acute liver failure of 53.98%. In terms of prediction models, XGBoost demonstrated the best performance on both the validation set and the training set, with an AUC value of 0.801 on the validation set. In the clinical decision curve analysis, XGBoost also showed optimal benefits. Conclusion The mortality risk prediction model for acute liver failure can help clinicians identify high-risk patients at an early stage, so as to reduce the mortality.

Objective To enhance the level of paperless medical records application in case hospitals, shorten the submission cycle of medical records, reduce the quality control time of nursing medical records, and lower the rejection rate of medical records. Methods Taking the paperless medical records submitted by the Department of Intensive Care Medicine as the research object, the construction of paperless system was analyzed, and the electronic medical record system, surgical anesthesia system and examination report system were reformed and optimized to improve the paperless rate of case hospitals. Results After optimization, the paperless rate of Intensive Care Medicine Department increased from (90.10±1.23)% to (95.60±1.41)%, and the paperless rate of the whole hospital increased from (89.02±2.03)% to (92.58±2.81)%, and the differences were statistically significant (P<0.05). Conclusion Through the reformation of electronic medical record system, operation anesthesia system and examination report system, the paperless rate was inproved, the time of nursing quality control records was reduced, and the return rate of medical records decreased, the level of paperless application in the hospital can be improved.

随着数字化转型的深入，医养康产业的融合发展成为推动社会养老、医疗健康和康复服务创新的重要途径。本研究提出基于数据资产要素的医养康融合模式，通过数据资产化提升产业效能，构建覆盖“汇聚 - 治理 - 加工 - 流通 - 交易 - 应用”全生命周期的技术体系。通过“四维一体”的数据治理框架，实现数据标准化与权属界定。结合联邦学习、差分隐私等技术保障数据安全共享，并基于区块链与智能合约设计多层级医养康数据交易平台，支撑跨机构数据流通与价值转化。通过典型应用场景，展示了数据资产如何为医养康服务提供技术支撑，提升健康管理、养老服务与康复治疗的效率与质量。展望了医养康数据资产要素驱动下生态系统面临的技术挑战，并对未来的发展方向做了简要探讨，以期推动医养康行业的数字化转型。

With the continuous development and application of new technologies such as big data, cloud computing,and artificial intelligence, data elements have become a key force to promote the transformation of the medical and health field and the transformation of data intelligence. The demand for converting medical data into data assets is increasingly urgent, but it is also facing severe data security challenges. Therefore, enhancing the security support capability of data assets has become an important task in the process of high-quality development of hospitals. In accordance with relevant national laws and regulations and industry standards, and with reference to the data security capability evaluation model, a tertiary Grade-A hospital built a management-technology-operation integrated data security governance and management system through a series of measures such as optimizing organizational structure, improving management system, asset sorting and classification, risk assessment, strengthening technical protection,improving operation management and control, and personnel training. This system enables the hospital to manage the security of data throughout its entire life cycle, enhances its data security protection capability, and provides a strong guarantee for the hospital to apply and share data assets and explore the value of data elements in compliance with laws and regulations.

Objective To solve the problems of difficulty, long cycle, low screening efficiency, and information asymmetry in patient recruitment in clinical trials, and to construct a one-stop intelligent recruitment system for clinical trial patients. Methods Based on the analysis of clinical trial requirements, a complete recruitment plan was designed, including OA process initiation, precise push of hospital service numbers, participant mini program registration, researcher data viewing, and a closed-loop process for patient enrollment. Result  Within six months of the system's launch, it successfully completed 7 recruitment tasks, accurately pushed more than 6 000 patients, and accumulated 546 registrations. The average recruitment time was shortened to 10 hours, and the enrollment rate reached 20%. Conclusion The system effectively shortens the recruitment process, improves recruitment efficiency and patient enrollment rate, and helps participants better understand recruitment information.

To develop an intelligent Traditional Chinese Medicine (TCM) constitution recommendation system based on multimodality, and explore the effective way to combine multimodal information with constitution and health preservation recommendation, in order to improve the accuracy of Traditional Chinese Medicine (TCM) constitution identification and personalized health management. Improve constitution questionnaire and collect basic information such as user's age, height, weight, etc, record their symptomatic manifestations and past (disease) history in detail, and analyze and process them together with tongue and face images to optimize the existing constitution prediction rules. User's constitution, BMI, tooth marks on the tongue, cracks on the tongue, tongue color, thickness of tongue coating, tongue coating color, and face color can be accurately displayed. Based on the above analyses,  further personalized health preservation program covering some dimensions such as massage health preservation, exercise health preservation, food health preservation, medicinal diet health preservation, health preservation gongfu, and music health preservation (with MP3 playback function) are recommended. The construction of intelligent TCM constitution recommendation system based on multimodality realizes the qualitative analysis of constitution cold and heat and completes the accurate quantification of constitution characteristics through the comprehensive analysis of user's multidimensional information, and in the recommendation decision-making link, the system combine user's real-time health status and historical data characteristics through the dynamic reasoning mechanism to realize the adaptive generation and dynamic optimization of personalized health preservation recommendation.The system can provide effective technical support and solutions for the construction of Traditional Chinese Medicine (TCM) intelligent health management system.

This article introduces 5 key technologies (representation, fusion, transformation, alignment, and collaborative learning), 6 application steps (determining application scenarios, selecting modal data, optimizing algorithm models, preparing data, training and validating models, deploying and implementing), and 2 basic environments (computing power infrastructure, algorithm model tools) for medical multimodal data machine learning, providing readers with reference for research and application of medical multimodal data machine learning.

Objective To analyze the current situation of TCM data assets governance, explore its development path, realize technological implementation, so as to promote the modernization and digital transformation of traditional Chinese medicine. Methods Cutting-edge technologies such as cloud computing, big data, artificial intelligence, blockchain and privacy computing were integrated to build a secure, trustworthy and controllable data assets governance platform, and the multi-source and multi-modal Chinese medicine data were standardized, tagged, classified and graded, and intelligently applied. Results The TCM data assets management platform was established, and the large model of TCM diagnosis and treatment, digitization of ancient books, optimization of smart pharmacy and development of data products were realized, which significantly improved the service efficiency and data value of traditional Chinese medicine. Conclusion The management of TCM data assets is an important foundation to promote the modernization and internationalization of TCM. In the future, we should further strengthen technology research and application practice to maximize the utilization of TCM data assets.

The article grounded in the zero trust security model, proposes and implements a multi-layered defense architecture tailored to the security requirements of hospital research data platforms. This architecture ensures secure access control through a bastion host, employs sandboxing technology to guarantee data access isolation, and relies on distributed firewalls and a dynamic privilege management mechanism to safeguard data during remote logins by domain users. Additionally, the adoption of centralized domain control management strategies and internet behavior auditing effectively reduces the risk of sensitive data leakage. The research demonstrates that the proposed solution significantly enhances the security protection capabilities and compliance management level of hospital research data platforms, providing practical reference for the digital transformation of the healthcare industry.

Objective Deepen the reform of the healthcare system by building an information platform for the sharing and mutual recognition of medical test results. Improve the efficiency of medical resource utilization, optimize the patient's medical treatment process, eliminate duplicate examinations and charges, save patient treatment time, effectively reduce the burden of medical treatment on the public, and save medical insurance costs. Methods By building a shared mutual recognition engine and utilizing technologies such as cloud computing, big data, machine learning, intelligent analysis, and data mining, closed-loop management of inspection and testing results can be achieved. Results Establish a provincial-level inspection and testing result sharing and mutual recognition information platform, realize cross level and cross regional mutual recognition and sharing of medical institution inspection and testing results, and apply technical modules such as appointment loading, screen display, mutual recognition reference, monitoring and warning, SMS notification, etc, to improve mutual recognition efficiency, strengthen mutual recognition quality control, carry out mutual recognition assessment. Conclusion The construction and promotion of the provincial inspection and testing results sharing and mutual recognition information platform have effectively promoted the rational utilization of health resources, further optimized the patient's medical treatment process, improved the quality of medical services, significantly reduced the patient's medical expenses and saved money on health insurance. 

Objective To assess the stroke risk in heavy smokers and provide reference for optimizing public health prevention strategies. Methods Based on relevant data of NHANES from 2017-2020, feature selection was performed by Lasso regression, followed by modeling with seven machine learning algorithms including Random Forest, XGBoost, LightGBM and stacking algorithm. Model performance was evaluated using 10-fold cross-validation,and decision curve analysis (DCA) and clinical impact curve (CIC) analysis were conducted. SHAP values were used to enhance model interpretability. Results The stacking model performed best on the test set, with an AUC of 0.7645, effectively distinguishing between individuals with high and low stroke risks. The AUC value on the training set was 0.753 3, confirming the model's stability during training. DCA and CIC analyses demonstrated that the model provided significant net benefits at multiple clinical decision thresholds. SHAP value analysis showed the contribution of key variables such as history of heart disease and hepatitis B vaccination to the prediction. Conclusion Machine learning can effectively predict stroke risk in heavy smokers, providing scientific basis for personalized prevention strategies. The study demonstrates the potential of data-driven models in disease prevention.