A smart management platform, system, and method for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence.

Abstract

This invention discloses an intelligent management platform, system, and method for the entire lifecycle of osteoporosis based on national prevention and treatment guidelines and multimodal artificial intelligence, belonging to the field of medical system technology. The intelligent management platform for the entire lifecycle of osteoporosis based on national prevention and treatment guidelines and multimodal artificial intelligence includes the following logical layers: A1: Guideline knowledge graph; A2: Disease-specific data lake; A3: Internet of Things data stream; A4: Intelligent engine layer; A5: Multimodal artificial intelligence model cluster. This invention realizes the "digital implementation" of national guidelines: through the guideline specification engine, textual guidelines are transformed into automatically executable clinical pathways, fundamentally solving the problem of the disconnect between guidelines and practice, and improving the homogenization of diagnosis and treatment nationwide.

Description

Technical Field

[0001] This invention relates to the field of medical system technology, specifically to an intelligent management platform, system, and method for the entire lifecycle of osteoporosis based on national prevention and treatment guidelines and multimodal artificial intelligence. Background Technology

[0002] Osteoporosis is a major chronic disease affecting public health in my country. It is a systemic bone disease characterized by low bone mass and damage to bone microstructure, leading to increased bone fragility and susceptibility to fractures. my country has a huge number of osteoporosis patients, and its prevention and treatment face systemic challenges such as low screening rates, non-standardized diagnosis, poor treatment adherence, lack of long-term management, and uneven distribution of medical resources. Current clinical practice and management have systemic bottlenecks, such as the disconnect between guidelines and practice, low efficiency of screening and diagnosis, fragmented and disjointed management, insufficient level of intelligent decision support, loss of data value, and gaps in follow-up and rehabilitation management.

[0003] The shortcomings of existing osteoporosis management platforms are: Existing technology CN107731294A discloses an osteoporosis patient management platform and its operation method. This system can provide real-time multi-level diagnosis, referral, and consultation suggestions, helping doctors track and manage patients' conditions and provide timely alerts. It can even enable seamless, real-time teamwork between doctors and hospitals in different regions and countries. For patients, they can learn about basic disease knowledge and prevention without going to the hospital, understand their current status, clearly understand their treatment plan, stages, and precautions, and obtain timely medication, follow-up visits, and auxiliary treatment plans, track progress, and receive real-time feedback. Doctors and patients can also achieve one-on-one, many-to-one, and one-to-many diagnostic and treatment interactions. Doctors can understand patients' conditions in real time, enabling long-term and efficient management of osteoporosis patients, avoiding or reducing the occurrence of fractures, improving physicians' academic level, improving patients' quality of life, and reducing overall national healthcare expenditure.

[0004] The aforementioned technologies fail to integrate national prevention and control guidelines with practical application, resulting in issues such as low diagnostic efficiency, fragmented management, insufficient intelligent decision support, loss of data value, and gaps in follow-up and rehabilitation management. Therefore, a comprehensive intelligent management platform, system, and method for osteoporosis throughout its entire lifecycle, based on national prevention and control guidelines and multimodal artificial intelligence, is needed to address these problems. This platform should integrate guidelines and practical application, offer high screening and diagnostic efficiency, provide integrated management, offer intelligent decision support, make reasonable use of data value, and provide reasonable follow-up and rehabilitation management. Summary of the Invention

[0005] One objective of this application is to provide an intelligent management platform, system, and method for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence, which can solve the technical problems raised in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an intelligent management platform, system, and method for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence. The intelligent management platform for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence includes the following logical layers: A1: Guideline Knowledge Graph: Transform the national guidelines and normative documents for the diagnosis and treatment of primary osteoporosis into a structured, computer-executable library of regulations, clinical pathway workflows, and diagnostic logic trees; A2: Disease-Specific Data Lake: Based on the internationally accepted collaborative model for observational medical outcomes, a standardized disease-specific data model is established, aggregating desensitized clinical data from various participating institutions; A3: IoT data stream: Real-time access to time-series data generated by smart wearable devices, home health monitoring devices, and hospital imaging and testing equipment; A4: Intelligent Engine Layer: The guide specifies the engine, which dynamically parses and executes the specifications in the guide knowledge graph, driving the automatic flow and verification of the business logic throughout the entire process of screening, diagnosis, treatment and follow-up; A5: Multimodal Artificial Intelligence Model Cluster.

[0007] Preferably, the multimodal artificial intelligence model cluster includes the following categories: A51. Natural Language Processing Model: Used to automatically extract key osteoporosis-related entities and relationships from unstructured electronic medical records; A52. Computer vision model: used for automatic bone density measurement, region of interest localization, and identification and grading of vertebral fractures from dual-energy X-ray absorptiometry, quantitative CT, and X-ray images; A53. Machine Learning Prediction Model: Based on longitudinal data, construct models for predicting fracture risk, treatment efficacy response, and compliance risk; A54. Personalized Recommendation Engine: Integrating comprehensive patient data, guideline constraints, and AI model output, it generates personalized screening suggestions, diagnostic assistance, treatment plan recommendations, and follow-up plans. A55. Application Service Layer: Provides core service capabilities in the form of APIs, including patient 360-degree view service, clinical decision support service, digital therapy management service, remote follow-up service, and data statistical analysis service; A56. Interactive Access Layer: Provides front-end applications for four types of users, including patient mobile terminals, clinical physician workstations, primary care public health physician terminals, and regional management dashboards.

[0008] Preferably, the osteoporosis full-cycle intelligent management platform based on national prevention and control guidelines and multimodal artificial intelligence includes the following subsystems: B1. Guideline-driven intelligent screening and risk stratification subsystem: It integrates internationally and domestically authoritative risk assessment models such as FRAX®, OSTA, and IOR, supports online and offline multi-channel self-assessment and peer assessment, connects to wearable devices to obtain fall-related risk data, realizes dynamic risk perception, automatically stratifies the population based on preset thresholds in guidelines, and pushes high-risk population management tasks to the primary healthcare system. B2. Multi-source integrated intelligent diagnosis and auxiliary reporting subsystem: Through standardized medical data interfaces, it automatically collects and structures key diagnostic data from hospital information systems, testing systems, and image archiving systems, activates AI image analysis models, performs real-time analysis on uploaded images, automatically generates a structured report draft containing key indicators, automatically integrates questionnaire, physical examination, test and image data, generates a structured intelligent diagnostic report that meets national guidelines and diagnostic standards with one click, and provides AI prompts for differential diagnosis; B3. Personalized full-cycle intervention and treatment management subsystem; B4. Continuous closed-loop follow-up and efficacy monitoring subsystem; B5. Data-driven value assessment and system optimization subsystem.

[0009] Preferably, the personalized full-cycle intervention and treatment management subsystem includes the following modules: B31. Digital Therapy Generator: Based on the guidelines for non-pharmacological interventions and combined with the individual patient's situation, it automatically generates a digital intervention package that includes nutritional prescriptions, personalized exercise programs, and suggestions on light exposure and health behaviors. B32. Guideline Compliance Prescription Reviewer: When doctors prescribe medications, the system compares the prescriptions with national guidelines in real time, providing proactive alerts and priority recommendations regarding drug selection, dosage, treatment duration, contraindications, and drug interactions. B33. Intelligent Medication Adherence Management Module: By connecting to smart pillboxes, pushing medication reminders, tracking electronic medication records, and implementing incentive mechanisms, this module manages patients' medication behavior throughout the entire process.

[0010] Preferably, the continuous closed-loop follow-up and efficacy monitoring subsystem includes the following modules: B41. Automated Follow-up Engine: Automatically generates personalized long-term follow-up timelines and task lists based on diagnostic results and initial treatment plans. B42. Remote Treatment Monitoring and Early Warning Center: Continuously collects patient self-reported outcomes, wearable device data, and follow-up visit data, uses AI treatment prediction models for evaluation, and sends early warnings to the medical team for suspected treatment failures, new fractures, and adverse reaction risk events. B43. Tiered Collaborative Follow-up Network: Supports the intelligent transfer and collaboration of follow-up tasks between specialists in higher-level hospitals and family doctors in primary care settings, achieving a closed loop of "emergency cases treated at the center, chronic cases treated in the community".

[0011] Preferably, the data-driven value assessment and system optimization subsystem includes the following modules: B51. Real-world research platform: Based on disease-specific data lakes, it provides visual cohort construction and statistical analysis tools to support comparative studies of treatment effects and pharmacoeconomic evaluations. B52. Quality Control and Decision-Making Dashboard: This dashboard dynamically displays core performance indicators for health administrators, including osteoporosis screening rate, standardized diagnosis rate, treatment rate, follow-up completion rate, and fracture incidence rate, supporting precise management and policy optimization.

[0012] Preferably, the method for intelligent management of osteoporosis throughout its entire lifecycle based on national prevention and control guidelines and multimodal artificial intelligence includes the following steps: S1: Risk screening and intelligent stratification: Risk assessment is initiated through multiple channels, and the system automatically completes the initial screening and stratification of population risks based on guidelines, thereby identifying management objectives; S2: Standardized Collaborative Diagnosis: Guides medium- and high-risk individuals into a standardized diagnostic process. The system automatically aggregates multi-source data, calls AI models to assist in analysis, generates standardized diagnostic reports, and supports doctors in making efficient diagnoses. S3: Personalized treatment plan decision-making and initiation: With the guidance of the system and the support of AI decision-making, doctors work with patients to develop personalized treatment plans, and the system simultaneously generates a detailed management plan for the patient. S4: Full-Scenario Execution and Collaborative Management: Patients execute treatment plans outside the hospital, with the system providing guidance through digital therapy and monitoring through the Internet of Things. Primary care physicians and senior physicians share information and collaborate on management tasks through the platform. S5: Dynamic Follow-up and Precise Optimization: The system automatically executes the follow-up plan, continuously collects multi-dimensional data, and the AI ​​model dynamically evaluates the efficacy and risks, providing data insights for doctors to adjust treatment plans and forming a management closed loop of "assessment-intervention-reassessment". S6: Data Aggregation and Value Feedback: The anonymized data generated throughout the process is continuously fed into the National Disease Data Lake to train more accurate AI models, conduct high-level real-world research, and evaluate the effectiveness of public health policies, ultimately feeding back into the continuous iteration of national guidelines and the optimization and upgrading of the prevention and control system.

[0013] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention realizes the "digital implementation" of national guidelines: through the guideline specification engine, text guidelines are transformed into automatically executable clinical pathways, fundamentally solving the problem of the disconnect between guidelines and practice, and improving the level of homogenization of diagnosis and treatment nationwide.

[0014] 2. This invention constructs an "AI-enhanced closed loop" covering the entire lifecycle: multimodal AI models are deeply integrated into all aspects of screening, diagnosis, treatment, and follow-up, providing medical staff with unprecedented intelligent assistance capabilities and significantly improving work efficiency and decision-making quality.

[0015] 3. This invention establishes a new management model of "integration of medical treatment and prevention, and linkage between different levels": the platform breaks down the data and business barriers between medical institutions at all levels, realizes continuous care and hierarchical diagnosis and treatment of osteoporosis, and truly implements the service concept of "patient-centered".

[0016] 4. This invention opens up a new path for "data-driven value-based healthcare": through a disease-specific data lake and value assessment subsystem, the medical process is transformed into measurable, analyzable, and optimizable data assets, providing a scientific basis for medical insurance payment reforms such as bundled payments, drug and medical device evaluation, and public health decision-making; it lays the foundation for a "sustainably evolving smart ecosystem": its platform-based and modular design gives it strong scalability, allowing for rapid integration of new AI algorithms, IoT devices, or medical services, forming a national-level digital health infrastructure that can continuously learn and evolve. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the overall architecture of the platform of this invention; Figure 2 This is a diagram showing the interaction relationships between the core subsystems of this invention. Figure 3 This is a flowchart of the intelligent screening and diagnosis process driven by the guidelines of this invention; Figure 4 A schematic diagram illustrating the full-cycle management closed loop and data value creation achieved by this invention; Figure 5 This is a flowchart of the risk dynamic prediction method based on multi-source data fusion in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1, please refer to Figure 1 , Figure 2 , Figure 3 , Figure 4 and Figure 5 An intelligent management platform, system, and method for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence. The intelligent management platform for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence includes the following logical layers: A1: Guideline Knowledge Graph: Transform the national guidelines and normative documents for the diagnosis and treatment of primary osteoporosis into a structured, computer-executable library of regulations, clinical pathway workflows, and diagnostic logic trees; A2: Disease-Specific Data Lake: Based on the internationally accepted collaborative model for observational medical outcomes, a standardized disease-specific data model is established, aggregating desensitized clinical data from various participating institutions; A3: IoT data stream: Real-time access to time-series data generated by smart wearable devices, home health monitoring devices, and hospital imaging and testing equipment; A4: Intelligent Engine Layer: The guide specifies the engine, which dynamically parses and executes the specifications in the guide knowledge graph, driving the automatic flow and verification of the business logic throughout the entire process of screening, diagnosis, treatment and follow-up; A5: Multimodal Artificial Intelligence Model Cluster; Multimodal artificial intelligence model clusters include the following categories: A51. Natural Language Processing Model: Used to automatically extract key osteoporosis-related entities and relationships from unstructured electronic medical records; A52. Computer vision model: used for automatic bone density measurement, region of interest localization, and identification and grading of vertebral fractures from dual-energy X-ray absorptiometry, quantitative CT, and X-ray images; A53. Machine Learning Prediction Model: Based on longitudinal data, construct models for predicting fracture risk, treatment efficacy response, and compliance risk; A54. Personalized Recommendation Engine: Integrating comprehensive patient data, guideline constraints, and AI model output, it generates personalized screening suggestions, diagnostic assistance, treatment plan recommendations, and follow-up plans. A55. Application Service Layer: Provides core service capabilities in the form of APIs, including patient 360-degree view service, clinical decision support service, digital therapy management service, remote follow-up service, and data statistical analysis service; A56. Interactive Access Layer: Provides front-end applications for four types of users, including patient mobile terminals such as mini-programs and apps, clinical physician workstations such as web and desktop, and primary care public health physician terminals such as mobile apps and regional management dashboards such as large screens and web. The intelligent management platform for the entire lifecycle of osteoporosis, based on national prevention and control guidelines and multimodal artificial intelligence, includes the following subsystems: B1. Guideline-driven intelligent screening and risk stratification subsystem: It integrates internationally and domestically authoritative risk assessment models such as FRAX®, OSTA, and IOR, supports online and offline multi-channel self-assessment and peer assessment, connects to wearable devices to obtain fall-related risk data such as gait and balance, realizes dynamic risk perception, automatically stratifies the population into low, medium, and high risk levels according to preset thresholds in guidelines, and pushes high-risk population management tasks to primary healthcare systems. B2. Multi-source integrated intelligent diagnosis and auxiliary reporting subsystem: Through standardized medical data interfaces, it automatically collects and structures key diagnostic data from hospital information systems, testing systems, and image archiving systems, activates AI image analysis models, performs real-time analysis on uploaded images, automatically generates a structured report draft containing key indicators, automatically integrates questionnaire, physical examination, test and image data, generates a structured intelligent diagnostic report that meets national guidelines and diagnostic standards with one click, and provides AI prompts for differential diagnosis; B3. Personalized full-cycle intervention and treatment management subsystem; B4. Continuous closed-loop follow-up and efficacy monitoring subsystem; B5. Data-driven value assessment and system optimization subsystem; The personalized full-cycle intervention and treatment management subsystem includes the following modules: B31. Digital Therapy Generator: Based on the guidelines for non-pharmacological interventions and combined with the individual patient's situation, it automatically generates a digital intervention package that includes a nutritional prescription, a personalized exercise program with fall prevention training, and suggestions on light exposure and health behaviors. B32. Guideline Compliance Prescription Reviewer: When doctors prescribe medications, the system compares the prescriptions with national guidelines in real time, providing proactive alerts and priority recommendations regarding drug selection, dosage, treatment duration, contraindications, and drug interactions. B33. Intelligent Medication Adherence Management Module: Through connection to smart pillboxes, push medication reminders, electronic medication history tracking, and incentive mechanisms, it manages patients' medication behavior throughout the entire process. The continuous closed-loop follow-up and efficacy monitoring subsystem includes the following modules: B41. Automated follow-up engine: Based on the diagnosis results and initial treatment plan, automatically generate a personalized long-term follow-up timeline and task list, such as bone density re-examination time and laboratory test items. B42. Remote Treatment Monitoring and Early Warning Center: Continuously collects patient self-reported outcomes, wearable device data, and follow-up visit data, uses AI treatment prediction models for evaluation, and sends early warnings to the medical team for suspected treatment failures, new fractures, and adverse reaction risk events. B43. Tiered Collaborative Follow-up Network: Supports the intelligent transfer and collaboration of follow-up tasks between specialists in higher-level hospitals and family doctors in primary care, realizing a closed loop of "emergency cases treated in the center, chronic cases treated in the community"; The data-driven value assessment and system optimization subsystem includes the following modules: B51. Real-world research platform: Based on disease-specific data lakes, it provides visual cohort construction and statistical analysis tools to support comparative studies of treatment effects and pharmacoeconomic evaluations. B52. Quality Control and Decision-Making Dashboard: Provides health administrators with dynamic displays of core performance indicators for regions and institutions, such as osteoporosis screening rate, diagnostic standardization rate, treatment rate, follow-up completion rate, and fracture incidence rate, supporting precise management and policy optimization. The intelligent management method for the entire life cycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence includes the following steps: S1: Risk screening and intelligent stratification: Risk assessment is initiated through multiple channels, and the system automatically completes the initial screening and stratification of population risks based on guidelines, thereby identifying management objectives; S2: Standardized Collaborative Diagnosis: Guides medium- and high-risk individuals into a standardized diagnostic process. The system automatically aggregates multi-source data, calls AI models to assist in analysis, generates standardized diagnostic reports, and supports doctors in making efficient diagnoses. S3: Personalized treatment plan decision-making and initiation: With the guidance of the system and the support of AI decision-making, doctors work with patients to develop personalized treatment plans, including medication and digital therapy. The system simultaneously generates a detailed management plan for the patient. S4: Full-Scenario Execution and Collaborative Management: Patients execute treatment plans outside the hospital, with the system providing guidance through digital therapy and monitoring through the Internet of Things. Primary care physicians and senior physicians share information and collaborate on management tasks through the platform. S5: Dynamic Follow-up and Precise Optimization: The system automatically executes the follow-up plan, continuously collects multi-dimensional data, and the AI ​​model dynamically evaluates the efficacy and risks, providing data insights for doctors to adjust treatment plans and forming a management closed loop of "assessment-intervention-reassessment". S6: Data Aggregation and Value Feedback: The de-identified data generated throughout the process is continuously fed into the National Disease Data Lake to train more accurate AI models, conduct high-level real-world research, and evaluate the effectiveness of public health policies, ultimately feeding back into the continuous iteration of national guidelines and the optimization and upgrading of the prevention and control system; Example 2, please refer to Figure 1 An explanation of the interaction diagram of the core subsystems of an intelligent management platform, system, and method for the whole-cycle management of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence: Detailed Explanation of Bottom-Up Architecture Layering: First layer: Data and knowledge layer, component: Guideline knowledge graph library: stores structured national prevention and control guidelines, clinical pathways and diagnostic logic in the form of a graph database; Disease-specific data lake: A pool of structured clinical data aggregated based on a general data model, supporting efficient analysis; External data interfaces: a standardized set of interfaces connecting regional health information platforms, hospital information systems including HIS, LIS and PACS, IoT devices including wearable devices and smart pillboxes, and public health databases; Functions and flow: This layer is the platform's "memory and perception" system. It continuously extracts, transforms, and loads multi-source heterogeneous data from external systems into the data lake, and provides unified and standardized data services to the upper layer. The guide knowledge graph is the "code" that drives all intelligent logic. The second layer: the brain and core intelligent engine layer, components: guidelines and regulations, engine: core processing unit, dynamically executes the regulations in the knowledge graph, and drives business process automation; Multimodal AI model cluster: NLP engine: processes text medical records and extracts key information; CV engine: analyzes medical images, identifies fractures and measures bone density; ML prediction engine: performs risk prediction and efficacy assessment; personalized recommendation engine: integrates patient data, guidelines and AI insights to generate personalized suggestions; microservice governance and API gateway: manages all service interfaces in a unified manner to ensure secure and stable service calls. Functions and flow: This layer is the "intelligent hub" of the platform. It receives raw data from the lower layer, calls the AI ​​model for processing, and the prescribed engine makes logical judgments according to the guidelines. Finally, the recommendation engine forms decision suggestions, and the processing results are provided to the application service layer through the API gateway. The third layer: the application service layer, which is a set of business capabilities. Components: presented in the form of a "service mesh" are a series of independent microservices, mainly including: patient 360° view service, clinical decision support service, digital therapy management service, remote follow-up and early warning service, and scientific research analysis and quality control service. Functions and Flow: This layer is the platform's "business arm," which encapsulates the capabilities of the intelligent engine layer into specific, reusable business functions. For example, when a doctor needs diagnostic assistance, the "Clinical Decision Support Service" will combine and call the lower-level specified engine, CV engine, and recommendation engine to package and return complete suggestions. The fourth layer: the interactive access layer, i.e., the user interface, includes components that showcase four main terminal application portals. Patient mobile devices: Mini-programs or apps that provide functions such as risk assessment, health education, digital therapy implementation, and follow-up feedback; Clinical physician workstation: integrated into the hospital information system or a standalone web application, providing a full-process workbench for intelligent outpatient services, assisted diagnosis, treatment plan development, and patient management; For primary care and public health physicians: a mobile app focusing on high-risk population screening, follow-up tasks, and collaborative referrals; Management and Research Dashboard: A web-based data dashboard that provides regional health managers and researchers with panoramic data visualization, quality indicator monitoring, and research analysis tools; Functions and Flow: This layer is the "face" of the platform. Different users access specific services of the application service layer through the corresponding portals. All interactive operations and data requests eventually go down to the intelligent engine and data layer for processing, and the results are then returned to the interface for display, forming a complete closed loop.

[0020] Example 3, please refer to Figure 2 This document describes an intelligent management platform, system, and method for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence. It includes an explanation of the interaction diagram of the core subsystems, detailed explanations of subsystem positioning and interactions, arranged from left to right according to business logic. I. Guideline-Driven Intelligent Screening and Risk Stratification Subsystem: Function icon: Displayed as a composite icon containing a questionnaire icon, a wearable device icon, and a risk level dashboard; Inputs: patient and resident self-reported questionnaire data, wearable device dynamic data, and basic demographic information; Processing: The subsystem call guide specifies the preset screening thresholds and risk assessment models in the engine to automatically calculate and stratify the input data; Output and Interaction: Patients and public health personnel: Real-time display of individual risk levels and recommendations; Intelligent Diagnosis and Assisted Reporting Subsystem: Automatically pushes "List of High-Risk and Medium-Risk Groups" and preliminary assessment reports, triggering standardized diagnostic processes; II. Multi-source fusion intelligent diagnosis and auxiliary reporting subsystem: Function icon: Displayed as a composite icon connecting medical devices, an AI brain icon, and a structured report document; Inputs: patient list from the previous subsystem, and multi-source clinical data from the hospital information system, such as images, lab reports, and medical record texts; Processing: Automatically schedule and collect device data, call multimodal AI model clusters such as CV model to analyze images and NLP model to parse text for auxiliary analysis; summarize all information, and the engine verifies the integrity of the diagnostic logic according to the guidelines, such as whether it meets the diagnostic criteria of bone mineral density T-score ≤ -2.5 + fracture history; Output and Interaction: Clinical Physician Workstation: Generates and presents structured intelligent diagnostic reports, including AI-annotated images, key indicators, and diagnostic conclusions, for physicians to review and confirm; The full-cycle intervention and treatment management subsystem automatically transmits the files and diagnostic details of diagnosed patients, providing a basis for developing treatment plans; III. Personalized Full-Cycle Intervention and Treatment Management Subsystem Function icons: The digital therapy module is displayed as a combination of a pill and a movement icon, as well as a composite icon of a prescription review warning light; Inputs: Confirmed patient records, treatment pathway database from national guidelines, and drug knowledge base; Processing: Digital therapy generator: Generates personalized treatment plans by combining data from a library of non-pharmacological interventions in guidelines based on the patient's condition; Prescription reviewer: Calls the guidelines-based engine in real time to conduct compliance reviews and recommendations when doctors issue prescriptions; Output and Interaction: Patient mobile app: Personalized digital therapy is pushed, including exercise videos and nutrition plans, along with medication reminders; For clinicians: Provides prescription decision support; Closed-loop follow-up and efficacy monitoring subsystem: synchronizes the initial treatment plan and management goals as the follow-up baseline; IV. Continuous Closed-Loop Follow-up and Efficacy Monitoring Subsystem Function icon: Displayed as a composite icon of a cyclic arrow calendar, a remote monitoring dashboard, and a warning triangle; Inputs: Initial treatment plan, daily patient monitoring data, with data from wearable devices, smart pillboxes, and patient-reported outcomes; Processing: Automated follow-up engine: Automatically generates follow-up task calendars based on guidelines and protocols; AI predictive model: Analyzes continuously input monitoring data, assesses efficacy trends, and predicts fracture or adherence risks; Output and Interaction: For both patients and public health physicians: Push follow-up task reminders; provide health feedback to patients; Clinician side: Sends efficacy evaluation reports and risk warning signals; Value assessment and system optimization subsystem: delivers anonymized structured follow-up data for macro-level analysis; Feedback to the treatment management subsystem: When an alert is triggered, a reverse notification can be sent to initiate the review and adjustment process of the treatment plan; Data-driven value assessment and system optimization subsystem Function icons: Displayed as composite icons representing the database, data analysis charts, and decision-making dashboard; Input: De-identified and standardized end-to-end data from all the aforementioned subsystems; Processing: Utilize statistical analysis and machine learning models to mine population data, calculate key performance indicators, and conduct real-world research; Output and Interaction: Management Cockpit: Display area / organizational level quality control dashboard; Research platform: Provides cohort analysis tools to support studies such as efficacy comparisons; Core feedback loop: Research findings and performance analysis conclusions are fed back to the guideline engine in the form of "evidence packages," providing real-world evidence for the dynamic updates of future guidelines and the optimization of clinical pathways, thereby enabling the system to self-evolve.

[0021] Example 4, please refer to Figure 3 This paper presents an intelligent management platform, system, and method for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence. It explains the guideline-driven intelligent screening and diagnosis flowchart, focusing on the first phase: multi-channel risk triggering and preliminary assessment. Starting point: The process is triggered by multiple channels, including community health education, physical examination report notifications, and patients' proactive inquiries; Step 1: Data Entry and Initial Risk Screening Implementers: Patients or public health personnel; Action: Enter basic demographic information and key risk factors such as age, gender, fracture history, and hormone use history via mobile device or workstation; System intelligent action: The system automatically calls the embedded risk assessment models in the "Guidelines" such as OSTA and FRAX® core parameters to complete the preliminary calculation; Decision point 1: Based on the preset thresholds in the guidelines, the system automatically determines the risk level; Low risk: The process concludes with general health education provided; Medium or high risk: Proceed to the next stage, namely standardized risk assessment; Phase Two: Standardized Risk Assessment and Tiered Management Step 2: In-depth risk assessment: For individuals at medium or high risk, the system guides them to complete a more comprehensive questionnaire recommended by the guidelines, such as the full FRAX® questionnaire, which includes bone mineral density-dependent risk factors; Optional integration: If a wearable device is already connected, the system can automatically obtain recent fall-related data such as gait stability index; System Intelligent Action: The system integrates all information and uses a guideline-based engine to perform precise risk stratification; Decision point 2: Based on the hierarchical results, the system automatically executes different management paths; Medium-risk: Marked as "key concern group", an annual follow-up plan is generated and stored in the community health record; High risk: Automatically triggers and generates an electronic request form for "bone density test recommendation"; at the same time, the system pushes the pending tasks to the relevant primary care physicians' workstations; Phase Three: Collaborative Inspection Arrangements and Data Collection Step 3: Check application and appointment: Implementer: Primary care physician confirms and submits the application; Intelligent system operation: Based on the regulations and guidelines of the medical consortium, the system intelligently recommends the most suitable examination type, such as prioritizing DXA, and recommending QCT or quantitative ultrasound when resources are insufficient, while also recommending the performing institution; patients can complete the examination appointment through the platform; Step 4: Automated collection of multi-source data: The patient went to the appointment for the examination; The system's core functions are as follows: After the examination is completed, the IoT data interface automatically uploads the raw images and data generated by the bone densitometer, CT, and other equipment to the platform in a standardized manner; at the same time, the system automatically retrieves the relevant bone turnover biochemical marker test results from the hospital's LIS. Phase Four: Artificial Intelligence-Assisted Analysis and Diagnosis Synthesis Step 5: Parallel Analysis of Multimodal AI The computer vision model automatically starts: it performs real-time analysis on the uploaded images and automatically completes: ① region of interest localization; ② bone mineral density calculation and T or Z value determination; ③ vertebral body morphology analysis, identification and grading of compression fractures according to the Genant method. The natural language processing model is launched simultaneously: it automatically extracts key text information related to the differential diagnosis of osteoporosis from the patient's historical electronic medical records, such as history of rheumatoid arthritis, history of hyperthyroidism, and medication records. Step 6: Intelligent Diagnostic Report Generation and Review: As the core intelligent action of the system: the guidelines stipulate that when the engine is activated, it performs logical verification and integration of the following information: AI-extracted image results such as T-scores and fracture information; AI-extracted test results such as PINP and β-CTX; AI-extracted key medical record texts; and a complete list of risk factors. Based on the diagnostic criteria in the guidelines, such as a T-score ≤ -2.5 or a T-score between -1.0 and -2.5 combined with a history of fragility fracture, the engine automatically generates structured diagnostic conclusions, such as: primary osteoporosis, severe or low bone mass. The system can generate a richly illustrated "Intelligent Diagnostic Assistance Report" with one click, including a data overview, AI image annotations, diagnostic basis, and differential diagnosis tips. Step 7: Doctor Review and Confirmation: Performer: Clinicians, including radiologists, orthopedic endocrinologists; Action: Doctors review the report online, focusing on the AI's annotations and conclusions; Doctors can make minor adjustments or confirm the report. Final output: The report becomes effective after being electronically signed by the doctor, is automatically archived in the patient's exclusive full-cycle health record, and is shared with the relevant doctors in real time; Example 5, please refer to Figure 4 Based on national prevention and control guidelines and multimodal artificial intelligence, this paper explains the full-cycle intelligent management platform, system and method for osteoporosis, and the diagram of the realized full-cycle management closed loop and data value creation. I. Inner loop: Individual full-cycle management business closed loop, i.e. execution layer. This loop describes the management journey of a single patient or user and is the source of value generation. Starting point and driving core: At the center of the loop are the "National Guideline Knowledge Graph" and the "Multimodal AI Model", which are the "intelligent heart" driving the entire business closed loop operation; The four key nodes, in clockwise order, are: Intelligent Screening and Diagnosis: Based on guidelines and AI, risk identification and accurate diagnosis are achieved, generating structured diagnostic data; Personalized Intervention and Treatment: Based on guidelines and AI, treatment plans are developed and implemented, generating treatment adherence data and digital therapy interaction data; Continuous Follow-up and Monitoring: Based on guidelines and AI, long-term efficacy tracking is conducted, generating dynamic efficacy data and risk warning signals; Dynamic Assessment and Optimization: Based on AI analysis of follow-up data, clinicians are supported in dynamically adjusting treatment plans, generating plan adjustment decision data. Closed-loop flow: The arrows show that these four stages are connected end to end, forming a patient-centered, continuously operating personalized management closed loop; the data generated in all stages flows to the "Osteoporosis Specialty Data Lake" below the chart in real time and in a standardized manner. II. Outer Loop: Data Value Enhancement and System Empowerment Closed Loop, i.e., the strategic layer. This loop describes how data is refined, transformed, and ultimately feeds back into the system, serving as the engine for value amplification. The second half of the data aggregation and refinement process involves continuously feeding full-cycle, multi-dimensional data from the inner ring and various levels of medical institutions into the "Osteoporosis Specialty Data Lake, such as OMOP and CDM." Through advanced analysis and mining methods, including statistical modeling, machine learning, and real-world research, three categories of high-value "knowledge products" are extracted from the data lake: real-world evidence, such as "a comparison of the long-term fracture prevention effects of drug A versus drug B in the elderly population" or "cost-effectiveness analysis of new digital therapies"; optimized predictive models, such as more accurate AI models for fracture risk and treatment response trained on larger-scale, higher-quality data; and insights into system effectiveness, such as regional screening gap analysis and quality and efficiency assessment reports of different treatment pathways. Value feedback and empowerment, upper loop: The aforementioned "knowledge products" flow upwards, injecting into and innovating the core components of the three major systems, thus completing the closure of the outer loop: Dynamically updated guidelines and paths: Real-world evidence provides the most direct evidence-based basis for the regular updating and refinement of national prevention and control guidelines, making the guidelines more aligned with clinical practice; Iterative intelligent engine: Optimized predictive models directly update the multimodal AI model cluster in the platform, making the system more intelligent and driving precise decision-making and policy; System effectiveness insights directly support the "management or research cockpit," providing health administrative departments with precise decision support for adjusting medical insurance payment policies, such as promoting bundled payments, allocating public health resources, and supervising medical quality; Closed-loop empowerment: Updated guidelines, more intelligent AI models, and more precise policies, in turn, serve as "better regulations and tools," injecting back into and strengthening the "intelligent heart" of the inner loop, thereby driving a new round of higher-quality individual full-cycle management; In this way, the inner and outer loops are interlocked, forming a perpetual value creation flywheel; Key design elements and flow labels: The "Data-Information-Knowledge-Wisdom" value chain is clearly illustrated by vertical arrows on the side of the diagram, showing how raw data is transformed into decision-making wisdom step by step along the outer ring. Double-ring interactive arrows: Thick arrows clearly mark the convergence flow from "inner ring business data" to "outer ring data lake", as well as the empowerment feedback flow from "outer ring update results" to "inner ring intelligent core"; Stakeholder labeling: Next to the key outputs on the outer ring of the diagram, label the main beneficiaries, such as "researchers" who obtain real-world evidence, "health policymakers" who obtain systemic insights, or "industry innovators" who obtain demand and validation scenarios; The underlying support cloud icon: The bottom of the chart displays "secure, compliant, and interconnected cloud computing and network infrastructure" in the form of clouds, which is the technical foundation for the operation of the entire dual-ring system.

[0022] Example 6, please refer to Figure 5 This document describes an intelligent management platform, system, and method for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence. It also includes an explanation of the flowchart for a dynamic risk prediction method based on multi-source data fusion, highlighting the core stages of the flowchart. Phase A, Multi-source Data Acquisition: Clinical diagnosis and treatment data: Structured clinical information is automatically acquired through standardized medical interfaces; Personal behavioral and environmental data: Real-time collection of daily health-related behaviors through IoT devices; Genetic and static characteristics: Includes traditional risk factors and innovative quantitative data on TCM syndromes; Phase B, Data Standardization and Feature Engineering: Standardization: Use medical terminology standards to ensure data consistency; Time-series feature construction: Create a personal health timeline and extract key dynamic change indicators; Multimodal fusion: Integrating different types of data into a unified feature vector to prepare for model analysis; Phase C, Establishment of Personal Digital Twin Baseline, Initial Assessment of Key Steps: By combining the initial standard medical test results with multi-source features, a personalized initial risk profile is created; an initial version of the "personal skeletal digital twin" is established as a benchmark for subsequent dynamic comparisons. Phase D, Dynamic Risk Assessment Calculation, Core Calculation Process: Prescription Engine: Quickly identifies extremely high-risk situations by applying clear prescriptions based on clinical guidelines; Machine Learning Model: Analyzes complex feature patterns to predict fracture probability and bone loss trends; Comprehensive Score: Integrates prescriptions and model outputs to generate a comprehensive risk assessment and interpretation of key factors; Phase E, Output and Feedback: Multi-dimensional visualization: providing customized risk display interfaces for different users; Intelligent early warning: triggering graded early warnings and personalized suggestions based on risk levels; Data closed loop: collecting user feedback and clinical outcome data to form a basis for continuous improvement; Stage F, the model continues to evolve: Federated learning: Enabling cross-institutional model collaborative optimization while protecting data privacy; Individual adaptation: As personal data accumulates, the accuracy of model predictions continues to improve; Process innovation is reflected in: I. True dynamism: Risk assessment is not a one-time calculation, but a continuous updating process over time; it captures dynamic indicators such as the slope of bone density changes and changes in behavioral patterns through time-series characteristics; II. Integration of Traditional Chinese and Western Medicine: Incorporating quantitative data of TCM syndromes into the feature engineering stage; This provides a basis for risk assessment in subsequent integrated traditional Chinese and Western medicine interventions; III. Dual-engine architecture: The defining engine ensures the bottom line of clinical safety and handles clearly defined high-risk situations; the machine learning engine uncovers complex risk patterns and provides accurate probability predictions. IV. Closed-loop evolutionary system: A complete closed loop is formed from data acquisition to model update; Federated learning enables continuous optimization of the model while protecting privacy.

[0023] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention, and no reference numerals in the claims should be construed as limiting the rights involved.

Claims

1. An intelligent management platform for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence, characterized by: The osteoporosis full-cycle intelligent management platform based on national prevention and control guidelines and multimodal artificial intelligence includes the following logical layers: A1: Guideline Knowledge Graph: Transform the national guidelines and normative documents for the diagnosis and treatment of primary osteoporosis into a structured, computer-executable library of regulations, clinical pathway workflows, and diagnostic logic trees; A2: Disease-Specific Data Lake: Based on the internationally accepted collaborative model for observational medical outcomes, a standardized disease-specific data model is established, aggregating desensitized clinical data from various participating institutions; A3: IoT data stream: Real-time access to time-series data generated by smart wearable devices, home health monitoring devices, and hospital imaging and testing equipment; A4: Intelligent Engine Layer: The guide specifies the engine, which dynamically parses and executes the specifications in the guide knowledge graph, driving the automatic flow and verification of the business logic throughout the entire process of screening, diagnosis, treatment and follow-up; A5: Multimodal Artificial Intelligence Model Cluster.

2. The osteoporosis full-cycle intelligent management platform based on national prevention and control guidelines and multimodal artificial intelligence as described in claim 1, characterized in that: The multimodal artificial intelligence model cluster includes the following categories: A51. Natural Language Processing Model: Used to automatically extract key osteoporosis-related entities and relationships from unstructured electronic medical records; A52. Computer vision model: used for automatic bone density measurement, region of interest localization, and identification and grading of vertebral fractures from dual-energy X-ray absorptiometry, quantitative CT, and X-ray images; A53. Machine Learning Prediction Model: Based on longitudinal data, construct models for predicting fracture risk, treatment efficacy response, and compliance risk; A54. Personalized Recommendation Engine: Integrating comprehensive patient data, guideline constraints, and AI model output, it generates personalized screening suggestions, diagnostic assistance, treatment plan recommendations, and follow-up plans. A55. Application Service Layer: Provides core service capabilities in the form of APIs, including patient 360-degree view service, clinical decision support service, digital therapy management service, remote follow-up service, and data statistical analysis service; A56. Interactive Access Layer: Provides front-end applications for four types of users, including patient mobile terminals, clinical physician workstations, primary care public health physician terminals, and regional management dashboards.

3. An intelligent management system for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence, applicable to the intelligent management platform for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence as described in any one of claims 1-2, characterized in that: The osteoporosis full-cycle intelligent management platform based on national prevention and control guidelines and multimodal artificial intelligence includes the following subsystems: B1. Guideline-driven intelligent screening and risk stratification subsystem: It integrates internationally and domestically authoritative risk assessment models such as FRAX®, OSTA, and IOR, supports online and offline multi-channel self-assessment and peer assessment, connects to wearable devices to obtain fall-related risk data, realizes dynamic risk perception, automatically stratifies the population based on preset thresholds in guidelines, and pushes high-risk population management tasks to the primary healthcare system. B2. Multi-source integrated intelligent diagnosis and auxiliary reporting subsystem: Through standardized medical data interfaces, it automatically collects and structures key diagnostic data from hospital information systems, testing systems, and image archiving systems, activates AI image analysis models, performs real-time analysis on uploaded images, automatically generates a structured report draft containing key indicators, automatically integrates questionnaire, physical examination, test and image data, generates a structured intelligent diagnostic report that meets national guidelines and diagnostic standards with one click, and provides AI prompts for differential diagnosis; B3. Personalized full-cycle intervention and treatment management subsystem; B4. Continuous closed-loop follow-up and efficacy monitoring subsystem; B5. Data-driven value assessment and system optimization subsystem.

4. The osteoporosis full-cycle intelligent management system based on national prevention and control guidelines and multimodal artificial intelligence as described in claim 3, characterized in that: The personalized full-cycle intervention and treatment management subsystem includes the following modules: B31. Digital Therapy Generator: Based on the guidelines for non-pharmacological interventions and combined with the individual patient's situation, it automatically generates a digital intervention package that includes nutritional prescriptions, personalized exercise programs, and suggestions on light exposure and health behaviors. B32. Guideline Compliance Prescription Reviewer: When doctors prescribe medications, the system compares the prescriptions with national guidelines in real time, providing proactive alerts and priority recommendations regarding drug selection, dosage, treatment duration, contraindications, and drug interactions. B33. Intelligent Medication Adherence Management Module: By connecting to smart pillboxes, pushing medication reminders, tracking electronic medication records, and implementing incentive mechanisms, this module manages patients' medication behavior throughout the entire process.

5. The osteoporosis full-cycle intelligent management system based on national prevention and control guidelines and multimodal artificial intelligence as described in claim 3, characterized in that: The continuous closed-loop follow-up and efficacy monitoring subsystem includes the following modules: B41. Automated Follow-up Engine: Automatically generates personalized long-term follow-up timelines and task lists based on diagnostic results and initial treatment plans. B42. Remote Treatment Monitoring and Early Warning Center: Continuously collects patient self-reported outcomes, wearable device data, and follow-up visit data, uses AI treatment prediction models for evaluation, and sends early warnings to the medical team for suspected treatment failures, new fractures, and adverse reaction risk events. B43. Tiered Collaborative Follow-up Network: Supports the intelligent transfer and collaboration of follow-up tasks between specialists in higher-level hospitals and family doctors in primary care settings, realizing a closed loop of "emergency cases treated at the center, chronic cases treated in the community".

6. The osteoporosis full-cycle intelligent management system based on national prevention and control guidelines and multimodal artificial intelligence as described in claim 3, characterized in that: The data-driven value assessment and system optimization subsystem includes the following modules: B51. Real-world research platform: Based on disease-specific data lakes, it provides visual cohort construction and statistical analysis tools to support comparative studies of treatment effects and pharmacoeconomic evaluations. B52. Quality Control and Decision-Making Dashboard: This dashboard dynamically displays core performance indicators for health administrators, including osteoporosis screening rate, standardized diagnosis rate, treatment rate, follow-up completion rate, and fracture incidence rate, supporting precise management and policy optimization.

7. A method for intelligent management of osteoporosis throughout its entire lifecycle based on national prevention and control guidelines and multimodal artificial intelligence, applicable to the intelligent management system for the entire lifecycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence as described in any one of claims 3-6, characterized in that: The intelligent management method for the entire life cycle of osteoporosis based on national prevention and control guidelines and multimodal artificial intelligence includes the following steps: S1: Risk screening and intelligent stratification: Risk assessment is initiated through multiple channels, and the system automatically completes the initial screening and stratification of population risks based on guidelines, thereby identifying management objectives; S2: Standardized Collaborative Diagnosis: Guides medium- and high-risk individuals into a standardized diagnostic process. The system automatically aggregates multi-source data, calls AI models to assist in analysis, generates standardized diagnostic reports, and supports doctors in making efficient diagnoses. S3: Personalized treatment plan decision-making and initiation: With the guidance of the system and the support of AI decision-making, doctors work with patients to develop personalized treatment plans, and the system simultaneously generates a detailed management plan for the patient. S4: Full-Scenario Execution and Collaborative Management: Patients execute treatment plans outside the hospital, with the system providing guidance through digital therapy and monitoring through the Internet of Things. Primary care physicians and senior physicians share information and collaborate on management tasks through the platform. S5: Dynamic Follow-up and Precise Optimization: The system automatically executes the follow-up plan, continuously collects multi-dimensional data, and uses AI models to dynamically evaluate efficacy and risks, providing doctors with data insights to adjust treatment plans and forming a management closed loop of "assessment-intervention-reassessment". S6: Data Aggregation and Value Feedback: The anonymized data generated throughout the process is continuously fed into the National Disease Data Lake to train more accurate AI models, conduct high-level real-world research, and evaluate the effectiveness of public health policies, ultimately feeding back into the continuous iteration of national guidelines and the optimization and upgrading of the prevention and control system.

Images