Chronic disease whole-process digital doctor-seeing cloud service and medical insurance real-time supervision unified large model

Abstract

The invention discloses a chronic disease whole-process digital medical treatment cloud service and medical insurance real-time supervision unified large model. The model comprises a technical architecture and a combination engine; the technical architecture comprises an infrastructure layer, a data base layer, a core technology layer, a business application layer and an ecological collaboration layer. The combination engine comprises federated learning and block chain combination, digital twinning and AI and big data combination, edge calculation and 5G and Internet of Things combination, NLP and knowledge graph and AR combination, quantum encryption and zero knowledge proof combination, and Web3.0 and DAO combination. Through cross-layer cooperation of a five-layer technical architecture and six technology combination engines, balance between data sharing and privacy protection is realized, cross-mechanism data islands are broken, data security compliance is guaranteed through technologies such as federal learning block chain quantum encryption, and the contradiction between data intercommunication and privacy protection in the prior art is solved; according to the digital twinborn AI big data combination, a full-process risk control system is constructed, and risk pre-prediction in-event intervention and post-event optimization are realized.

Description

Technical Field

[0001] This application relates to the technical field, and in particular to a unified big data model for digital medical cloud services for the entire process of chronic disease treatment and real-time medical insurance supervision. Background Technology

[0002] With the deepening of population aging and changes in lifestyle, chronic diseases have become a significant public health issue affecting public health, covering a continuously expanding population and showing a trend towards younger age groups, posing severe challenges to the supply of medical services and the sustainable operation of medical insurance funds. At the same time, policy guidance from the construction of Digital China and the reform of the medical security system is driving the digital and intelligent transformation of chronic disease management and medical insurance supervision, making the industry's demand for end-to-end, integrated solutions increasingly urgent.

[0003] In current technologies, digital management of chronic diseases is mostly concentrated in single institutions or localized scenarios. Although services such as online consultations and electronic prescriptions have emerged, there is a general problem of poor data interoperability across institutions. Data in the fields of healthcare, medical insurance, and pharmaceuticals are scattered and stored by different entities, forming data silos. Centralized data sharing models are prone to privacy leaks, while decentralized storage makes it difficult to deeply mine the value of data. As a result, chronic disease diagnosis and treatment lack complete health data support, making it difficult to implement full-course management.

[0004] In the field of medical insurance risk control, existing regulatory models mostly rely on traditional rule engines for post-event audits, which can only identify explicit violations and are insufficient to address highly concealed related fraud and latent violations. At the same time, risk control lacks the ability to dynamically predict the operational trends of the medical insurance fund, cannot provide early warnings of fund overspending risks, and struggles to accurately assess the impact of policy adjustments. This leads to increasing pressure on medical insurance fund payments year by year, and problems such as fraud and abuse still occur frequently.

[0005] In terms of full-course disease management, existing services are mostly concentrated in hospital settings, with insufficient coverage of home and community settings outside the hospital. Home health monitoring data is poorly integrated with the medical system, primary healthcare institutions have weak service capabilities and lack technical support, remote diagnosis and treatment lacks accuracy and convenience, and medication guidance lacks intuitive tools. This leads to a disconnect between outpatient and inpatient management of chronic disease patients, low medication adherence, and poor disease control.

[0006] At the level of ecosystem collaboration, existing models are mostly one-way connections dominated by centralized platforms, lacking effective collaboration mechanisms and incentives for participants such as medical institutions, insurance companies, and pharmaceutical supply chains. The ownership and usage rights of data by each party are vaguely defined, the trust costs of data sharing are high, and participants have difficulty participating in rule-making and service optimization, resulting in low efficiency in ecosystem collaboration.

[0007] In summary, we propose a unified big model for digital cloud services for chronic disease treatment and real-time medical insurance supervision throughout the entire process. Summary of the Invention

[0008] This application aims to at least partially solve one of the technical problems in the aforementioned technologies.

[0009] To achieve the above objectives, the first aspect of this application proposes a unified big model for digital medical cloud services for the entire process of chronic disease treatment and real-time medical insurance supervision, including technical architecture and combination engine;

[0010] The technical architecture includes an infrastructure layer, a data foundation layer, a core technology layer, a business application layer, and an ecosystem collaboration layer.

[0011] The infrastructure layer provides elastically scalable computing, storage, and network resources, supporting collaborative scheduling between edge nodes and the cloud, as well as redundant backup of 5G private networks and public networks. The data foundation layer enables multi-source integration, standardized governance, and secure sharing of medical insurance data, medical diagnosis and treatment data, drug supply chain data, health record data, and insurance data, supporting interconnection of federated learning nodes and two-way verification between blockchain and data lake warehouses. The core technology layer provides the core technical capabilities required for intelligent identification, risk analysis, and convenient services, enabling real-time linkage between digital twins and AI models, and cross-empowerment of NLP and knowledge graphs. The business application layer supports two core businesses: full-process service for chronic disease treatment and real-time risk control and supervision of medical insurance, realizing the integration of AR-assisted diagnosis and treatment and telemedicine, as well as closed-loop management of smart medicine boxes and prescription circulation. The ecosystem collaboration layer enables system docking and business collaboration among multiple entities such as medical insurance bureaus, medical institutions, retail pharmacies, insurance companies, and logistics companies, supporting Web3.0 ecosystem governance and DAO collaboration mechanisms, as well as interoperability of cross-regional medical insurance data alliances.

[0012] The combined engines include federated learning and blockchain, digital twins and AI and big data, edge computing and 5G and IoT, NLP and knowledge graphs and AR, quantum encryption and zero-knowledge proofs, and Web3.0 and DAO.

[0013] Federated learning and blockchain are combined at the data foundation layer for cross-regional data collaboration modeling and privacy protection; digital twins, AI, and big data are combined at the core technology layer for medical insurance fund risk prediction and policy simulation; edge computing, 5G, and IoT are combined between the infrastructure layer and the business application layer for full-scenario data collection and real-time processing; NLP, knowledge graphs, and AR are combined at the core technology layer and the business application layer for intelligent diagnosis and medication guidance; quantum encryption and zero-knowledge proofs are combined at the data foundation layer and the ecosystem collaboration layer for secure transmission and privacy sharing of highly sensitive data; and Web3.0 and DAO are combined at the ecosystem collaboration layer to build a multi-party collaborative governance ecosystem.

[0014] In addition, the unified big data model for digital medical cloud service and real-time medical insurance supervision of chronic diseases proposed in this application may also have the following additional technical features:

[0015] As a further description of the above technical solution:

[0016] The combination of federated learning and blockchain includes a federated learning framework and a consortium blockchain. The federated learning framework sets up provincial aggregation nodes and municipal participating nodes. Each participating node trains model parameters locally and only uploads updated parameter values. The consortium blockchain records each node's modeling permissions, parameter transmission trajectory, and model iteration versions.

[0017] As a further description of the above technical solution:

[0018] The combination of digital twins with AI and big data includes a digital twin of the medical insurance fund, a deep learning model, and real-time data streams. It integrates medical insurance fund data, population data, chronic disease diagnosis and treatment data, and drug consumption data to construct a digital twin of the medical insurance fund. It uses time series models to predict the peak of short-term expenditures and the sustainability of medium and long-term fund expenditures, and uses graph neural networks to uncover hidden fraud patterns. Furthermore, the digital twin of the medical insurance fund supports multi-scenario simulation and deduction by inputting cost control policy parameters.

[0019] As a further description of the above technical solution:

[0020] The combination of NLP, knowledge graph, and AR includes a medical NLP engine, a chronic disease diagnosis and treatment knowledge graph, and AR visualization tools. The chronic disease diagnosis and treatment knowledge graph covers information on diseases, drugs, and treatment rules, and is updated in real time with medical insurance policies and clinical guidelines. The medical NLP engine parses unstructured diagnosis and treatment data and uses a multi-model voting mechanism to achieve intelligent prescription review. The AR visualization tools are used for remote consultation to assist in annotation and provide medication guidance for patients.

[0021] As a further description of the above technical solution:

[0022] The combination of edge computing with 5G and IoT includes edge gateways, 5G private networks, and multimodal IoT devices. Edge gateways are deployed in a hierarchical manner in community health service centers and smart screens at home, processing health monitoring data collected by multimodal IoT devices locally in real time and triggering early warnings. The 5G private network adopts a medical-specific slicing network to ensure that the data transmission latency for remote consultations is ≤50ms.

[0023] As a further description of the above technical solution:

[0024] It also includes a drug circulation traceability module, which uses blockchain, IoT and RFID technologies. The smallest sales unit of a drug is affixed with an RFID tag to record the entire circulation data. The IoT tracks the logistics status in real time, and the blockchain stores traceability information, realizing a unique binding between drugs, prescriptions and patients, with traceability accuracy down to the single-item level.

[0025] As a further description of the above technical solution:

[0026] The Web3.0 and DAO combination includes decentralized applications and distributed autonomous organizations. Decentralized applications support access from multiple nodes such as patients, doctors, medical institutions, and insurance companies. Distributed autonomous organizations allocate voting rights according to the contribution value of each party, and participate in rule-making, service evaluation, and risk reporting through token incentives. They also combine zero-knowledge proof technology to achieve data privacy sharing.

[0027] As a further description of the above technical solution:

[0028] AR visualization tools are set up on the doctor's, patient's, and pharmacist's ends. On the doctor's end, AR overlays anatomical structure annotations to assist in remote consultation and diagnosis. On the patient's end, AR scans drugs to obtain medication information. On the pharmacist's end, AR presents the risk of drug interactions.

[0029] As a further description of the above technical solution:

[0030] The combination of quantum encryption and zero-knowledge proof includes a quantum key distribution device and a zero-knowledge proof algorithm. The core node is equipped with a quantum key distribution device to encrypt the transmission of sensitive data. The zero-knowledge proof algorithm verifies the authenticity of key information only when data is authorized for query, without exposing complete private data.

[0031] Advantages of this invention:

[0032] Based on the unified big data model for digital medical cloud services and real-time medical insurance supervision throughout the entire chronic disease process proposed in this application, a balance between data sharing and privacy protection is achieved through cross-layer collaboration of a five-layer technical architecture and six major technology engines. This breaks down data silos across institutions and ensures data security and compliance through technologies such as federated learning, blockchain, and quantum encryption, resolving the contradiction between data interoperability and privacy protection in existing technologies. Relying on a combination of digital twins, AI, and big data, a full-process risk control system is constructed to achieve pre-event risk prediction, in-event intervention, and post-event optimization. It accurately identifies hidden and related fraud and can also simulate policy impacts, overcoming the passivity and limitations of existing risk control methods. Furthermore, it integrates edge computing, 5G, and the Internet of Things. AR technology enables comprehensive chronic disease management across hospitals, communities, and homes, improving the accuracy of remote diagnosis and treatment, the real-time nature of home monitoring, and the intuitiveness of medication guidance. It addresses the disconnect between outpatient and inpatient management, thereby improving medication adherence. By combining Web3.0 and DAO, a decentralized co-governance system is established, clarifying the data rights and responsibilities of all parties, setting up effective incentive mechanisms, enhancing the collaborative enthusiasm and rule adaptability of participants, and constructing a closed loop linking medical, pharmaceutical, and medical insurance services. The standardized design of technical modules adapts to the differentiated needs of multiple cities, resulting in low deployment costs and high implementation efficiency. It balances system stability and performance, effectively supporting city-level large-scale applications and meeting the needs of integrated services and supervision.

[0033] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0034] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0035] Figure 1 This is a hierarchical diagram of a unified big model for digital medical cloud service and real-time medical insurance supervision of the entire chronic disease process according to an embodiment of this application;

[0036] Figure 2 This is a schematic diagram of the engine of a unified big model for digital medical cloud service for the entire process of chronic disease treatment and real-time medical insurance supervision, according to an embodiment of this application. Detailed Implementation

[0037] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0038] The following description, in conjunction with the accompanying drawings, describes the unified large-scale model of digital cloud service for chronic disease treatment and real-time medical insurance supervision throughout the entire process of this application.

[0039] like Figure 1-2 As shown, the unified big model for digital medical cloud service and real-time medical insurance supervision of the entire process of chronic disease treatment in Embodiment 1 of this application may include a technical architecture and a combined engine.

[0040] The infrastructure layer provides elastically scalable computing, storage, and network resources, supporting collaborative scheduling between edge nodes and the cloud, as well as redundant backups of 5G private networks and public networks. Through hybrid deployment and redundant design, it balances resource elasticity and service stability, adapting to large-scale deployments in multiple cities and peak business processing demands, reducing system downtime risk. The data foundation layer achieves multi-source integration, standardized governance, and secure sharing of medical insurance data, medical diagnosis and treatment data, pharmaceutical supply chain data, health record data, and insurance data. It supports interconnection of federated learning nodes and two-way verification between blockchain and data lake warehouses. Data standardization and cross-technology verification break down data silos while ensuring data authenticity and privacy, enabling in-depth value mining of multi-source data and providing high-quality data support for subsequent services and risk control. The core technology layer provides the core technical capabilities required for intelligent identification, risk analysis, and convenient services, enabling real-time linkage between digital twins and AI models, and cross-empowerment of NLP and knowledge graphs. Through cross-layer technology linkage, it strengthens intelligent decision-making. The system enhances decision-making capabilities, improving diagnostic accuracy and risk identification efficiency. The business application layer supports two core businesses: full-process chronic disease management and real-time medical insurance risk control and supervision. It integrates AR-assisted diagnosis and remote medical care, and implements closed-loop management of smart medicine boxes and prescription circulation. Deep integration of business scenarios and technical tools constructs a service loop, covering the entire medical process and improving service continuity and convenience. The ecosystem collaboration layer enables system integration and business collaboration among multiple entities, including medical insurance bureaus, medical institutions, retail pharmacies, insurance companies, and logistics companies. It supports Web3.0 ecosystem governance, DAO collaboration mechanisms, and interoperability of cross-regional medical insurance data alliances. Through decentralized architecture and collaboration mechanisms, it reduces trust costs in multi-party cooperation, strengthens ecosystem participation stickiness, and builds a collaborative governance system. The combined engines include federated learning and blockchain, digital twins and AI and big data, edge computing and 5G and IoT, NLP and knowledge graphs and AR, quantum encryption and zero-knowledge proofs, and Web3.The system combines 0 with DAO; federated learning and blockchain are integrated at the data foundation layer for cross-regional data collaboration modeling and privacy protection. Local modeling and parameter sharing avoid the transfer of raw data, while blockchain notarization ensures modeling compliance, balancing data privacy and cross-institutional collaboration value. Digital twins, AI, and big data are integrated at the core technology layer for medical insurance fund risk prediction and policy simulation. Digital twins recreate the fund's operational logic, and AI models uncover data patterns to achieve early risk warnings and accurate policy assessments. Edge computing, 5G, and IoT are integrated between the infrastructure layer and the business application layer for full-scenario data collection and real-time processing. Localized computing at edge nodes reduces transmission latency, and 5G ensures high-concurrency data transmission, covering multiple scenarios outside hospitals and improving data processing capabilities. Real-time performance is ensured; NLP, knowledge graphs, and AR are combined at the core technology layer and business application layer for intelligent diagnosis and medication guidance. NLP parses unstructured data, knowledge graphs match treatment rules, and AR visualizes information, improving diagnosis efficiency and medication adherence. Quantum encryption and zero-knowledge proofs are combined at the data foundation layer and ecosystem collaboration layer for secure transmission and privacy sharing of highly sensitive data. Quantum encryption ensures transmission security, and zero-knowledge proofs ensure data is usable but not visible, mitigating the privacy leakage risk of sharing highly sensitive data. Web3.0 and DAO are combined at the ecosystem collaboration layer to build a multi-party collaborative governance ecosystem. The decentralized architecture clarifies data rights and responsibilities, and incentive mechanisms motivate participation, improving ecosystem collaboration efficiency and rule adaptability.

[0041] like Figure 1-2 As shown:

[0042] The combination of federated learning and blockchain includes a federated learning framework and a consortium blockchain. The federated learning framework sets up provincial aggregation nodes and city-level participating nodes. Each participating node trains model parameters locally and only uploads updated parameter values. The consortium blockchain records the modeling permissions of each node, the parameter transmission trajectory, and model iteration versions. Provincial nodes aggregate parameters to form a global model. The consortium blockchain traces the entire modeling process, avoiding privacy leaks caused by cross-regional data transfer, while ensuring the compliance and traceability of model training, and improving the efficiency and security of cross-regional collaborative modeling.

[0043] like Figure 1-2 As shown:

[0044] The combination of digital twins with AI and big data includes a digital twin of the medical insurance fund, a deep learning model, and real-time data streams. It integrates medical insurance fund data, population data, chronic disease diagnosis and treatment data, and drug consumption data to construct a digital twin of the medical insurance fund. Through time series models, it predicts the short-term peak expenditure and medium- to long-term sustainability of the fund. Through graph neural networks, it uncovers hidden fraud patterns. The digital twin of the medical insurance fund supports multi-scenario simulation and extrapolation by inputting cost control policy parameters. The digital twin maps the real state of fund operation, and the deep learning model captures the time series patterns and correlation characteristics of data. It breaks through the limitations of traditional ex-post audits, realizes accurate prediction of fund risks and identification of hidden fraud, and provides data support for policy formulation, reducing the trial and error costs of policy adjustments.

[0045] like Figure 1-2 As shown:

[0046] The combination of NLP, knowledge graphs, and AR includes a medical NLP engine, a chronic disease diagnosis and treatment knowledge graph, and AR visualization tools. The chronic disease diagnosis and treatment knowledge graph covers information on diseases, drugs, and treatment rules, and is updated in real time with medical insurance policies and clinical guidelines. The medical NLP engine parses unstructured diagnosis and treatment data and uses a multi-model voting mechanism to achieve intelligent prescription review. The AR visualization tool is used for remote consultation to assist in annotation and provide medication guidance for patients. NLP transforms unstructured data into structured information, the knowledge graph provides support for accurate diagnosis and treatment rules, multi-model voting reduces review errors, and AR intuitively presents professional information, improving the accuracy of prescription review and diagnosis and treatment, reducing the cost of doctor-patient communication, and enhancing the effectiveness of medication guidance.

[0047] like Figure 1-2 As shown:

[0048] The combination of edge computing with 5G and IoT includes edge gateways, 5G private networks, and multimodal IoT devices. Edge gateways are deployed hierarchically in community health service centers and smart screens at home, processing health monitoring data collected by multimodal IoT devices locally in real time and triggering early warnings. The 5G private network adopts a medical-specific slicing network to ensure that the data transmission latency for remote consultations is ≤50ms. Edge nodes process data locally to reduce transmission links, and the 5G dedicated slice ensures bandwidth and low latency for critical services, enabling real-time collection and feedback of health data in out-of-hospital scenarios such as homes and communities, improving the remote diagnosis and treatment experience, and supporting lightweight deployment in primary healthcare institutions.

[0049] like Figure 1-2 As shown:

[0050] It also includes a drug circulation traceability module, which uses blockchain, IoT, and RFID technologies. The smallest sales unit of a drug is affixed with an RFID tag to record the entire circulation data. The IoT tracks the logistics status in real time, and the blockchain stores traceability information, realizing a unique binding between drugs, prescriptions, and patients, with traceability accuracy down to the single-item level. RFID identifies individual drugs, the IoT tracks the circulation trajectory, and the blockchain ensures that traceability information cannot be tampered with, realizing full traceability of the drug circulation process, preventing drug substitution and the circulation of counterfeit drugs, and ensuring medication safety.

[0051] like Figure 1-2 As shown:

[0052] The combination of Web3.0 and DAO includes decentralized applications and distributed autonomous organizations. Decentralized applications support access from multiple nodes such as patients, doctors, medical institutions, and insurance companies. Distributed autonomous organizations allocate voting rights according to the contribution value of each party, and participate in rule-making, service evaluation, and risk reporting through token incentives. Combined with zero-knowledge proof technology, data privacy sharing is achieved. The decentralized architecture ensures clear data rights and responsibilities, the contribution value mechanism and token incentives mobilize participation, and zero-knowledge proofs balance data sharing and privacy protection. This breaks the one-way control model of centralized platforms, enhances the initiative and collaborative efficiency of ecosystem participants, and builds a sustainable ecosystem governance system.

[0053] like Figure 1-2 As shown:

[0054] AR visualization tools are installed on the doctor's, patient's, and pharmacist's ends. On the doctor's end, AR overlays anatomical structure annotations to assist in remote consultation and diagnosis. On the patient's end, AR scans drugs to obtain medication information. On the pharmacist's end, AR presents the risks of drug interactions. AR technology visualizes and contextualizes abstract medical information, making remote diagnosis more accurate for doctors, medication understanding clearer for patients, and prescription review more efficient for pharmacists, comprehensively improving the safety and convenience of the entire process of diagnosis, treatment, and medication.

[0055] like Figure 1-2 As shown:

[0056] The combination of quantum encryption and zero-knowledge proof includes a quantum key distribution device and a zero-knowledge proof algorithm. The core node is equipped with a quantum key distribution device to encrypt the transmission of sensitive data. The zero-knowledge proof algorithm verifies only the authenticity of key information during authorized data queries without exposing complete private data. The unbreakable nature of the quantum key ensures transmission security, while the zero-knowledge proof separates data verification from privacy protection, meeting the secure sharing needs of highly sensitive medical data and medical insurance payment data, while balancing data usage efficiency and privacy compliance requirements.

[0057] Example 2, taking a certain city as an example, has more than 2 million patients with chronic diseases. There are problems such as data incompatibility across hospitals, high pressure on medical insurance fund overspending, and insufficient coverage of outpatient management. Based on this model, the Medical Insurance Bureau of City A, together with the Health Commission, 3 tertiary hospitals, 20 community health service centers, 50 designated pharmacies and 3 insurance companies, has built an integrated service and supervision system covering 6 key chronic diseases such as hypertension and diabetes. The details are explained below.

[0058] Technical Architecture:

[0059] The infrastructure layer uses the Kubernetes container orchestration model to achieve elastic resource scheduling, and the scheduling algorithm is a greedy algorithm based on load balancing (the formula is...). ,in Assigning priority to node scheduling For the remaining computing resources of the node, (Based on the current load of the node), the 5G private network and public network redundancy backup adopt a primary-backup switching model. The switching trigger condition is network latency > 50ms or packet loss rate > 1%.

[0060] In the data foundation layer, data standardization adopts an ontology-based mapping model, with mapping rules defined through XMLSchema. Data cleaning employs a missing value imputation model (continuous data is imputed with the mean, and discrete data with the mode). Federated learning nodes utilize... The algorithm framework and parameter aggregation formula are as follows: ( These are global model parameters. The number of participating nodes, For the first Nodes (Round local model parameters).

[0061] The core technology layer of the digital twin of the medical insurance fund adopts a system dynamics model, and the core equation is: , for Time Fund Balance Initial balance for Time Fund Income for For time-based fund expenditures, the medical NLP engine uses the BERT-BiLSTM-CRF model to parse medical record text, achieving an entity recognition accuracy of ≥93%. The chronic disease diagnosis and treatment knowledge graph is constructed using RDF triples, covering 1200 diseases, 3000 drugs, and 10000+ diagnosis and treatment rules. Relational reasoning uses the TransE model. ( For the head entity, For the relationship, (The tail entity).

[0062] In the business application layer, intelligent prescription review adopts a multi-model fusion voting mechanism, the formula of which is: ( For the probability of the final review result, , , As weight and =1, The results are from the logistic regression model. The results are from the random forest model. (Based on the results of a deep learning model), the unified payment platform adopts the aggregated payment and settlement formula as follows: The AR-assisted diagnosis and treatment tool is integrated into applications for doctors, patients, and pharmacists. The doctor's side is compatible with the Microsoft HoloLens 2 head-mounted AR device for remote consultation and 3D annotation of lesions. The patient's side is compatible with iPhone 15 series, Huawei Mate 60 series smart terminals, and iPad Pro 2024 tablets, enabling visualization of medication guidance through a dedicated APP. The pharmacist's side is compatible with Magic Leap 2 AR glasses and Lenovo Legion YogaPad Pro tablets, assisting in viewing drug interaction visualization maps and realizing functions such as remote diagnosis annotation and visualization of medication guidance.

[0063] The calculation model for the contribution value of DAO organizations in the ecological synergy layer is as follows: ( Total contribution value , , As weight, To contribute to the amount of data, To improve service participation, (Contributing to rule optimization), cross-subject data authorization adopts an attribute-based access control (ABAC) model, with the authorization rules being: .

[0064] Combined Engine:

[0065] The combination of federated learning and blockchain aims to identify medical insurance fraud in City A. Each district and county node trains an XGBoos model based on local data. The objective function is:

[0066] Only the model parameters are uploaded to the city-level aggregation node to form a global fraud identification model. The blockchain adopts the Hyperledger Fabric consortium chain, the block generation adopts the PBFT consensus algorithm, the block generation time is set to 10 seconds, the block hash calculation adopts the SHA-256 algorithm, and the entire modeling process and parameter flow trajectory are recorded.

[0067] Fund risk prediction using a combination of digital twins, AI, and big data: An LSTM model is used to predict peak short-term fund expenditures; the core formula is... ,in In hidden state, In cellular state, For the Gate of Oblivion As the input gate, a graph neural network (GNN) is used to uncover hidden fraud patterns such as prescriptions from multiple hospitals and abnormal connections between doctors and pharmacies. The node embedding formula is as follows: ,in For nodes Embedded vector, For nodes The set of neighboring nodes is used to simulate the impact of different cost control policy parameters (such as expanding the scope of centralized drug procurement and adjusting the reimbursement ratio) on fund operation, patient burden and institutional income by inputting them into the digital twin.

[0068] In the combination of edge computing, 5G, and IoT, edge nodes are set up as edge servers in community health service centers to process health monitoring data from 500-1000 patients with chronic diseases within their jurisdiction, with a response time of ≤200ms. Smart screens in homes integrate edge computing modules, connecting to devices such as smart blood pressure monitors and blood glucose meters to achieve local data caching and anomaly alerts. In data collection and processing, multimodal IoT devices collect patients' health data such as blood pressure, blood glucose, and heart rate. Edge nodes use a sliding window model (window size set to 5s) to process data, and anomaly detection uses the 3σ criterion. ,in The mean of the data. The standard deviation is used to determine an anomaly and trigger an early warning if this condition is met. The 5G private network ensures the transmission of high-definition audio and video for remote consultations, with a latency of ≤50ms and a bandwidth of ≥100Mbps.

[0069] NLP combined with knowledge graphs and AR: The medical NLP engine parses medical record text, the chronic disease diagnosis and treatment knowledge graph covers 1,200 diseases and 3,000 drug rules, and the AR tool is deployed on the doctor's, patient's and pharmacist's ends to realize remote annotation and medication guidance.

[0070] In a medical NLP engine combining quantum encryption and zero-knowledge proofs, patient medical record text is parsed to extract core diagnostic and treatment information. Knowledge graphs are used to match disease types and medication rules, assisting doctors in prescribing accurate prescriptions. An AR visualization tool provides doctors with remote consultation and lesion annotation capabilities, with a positioning error ≤5mm (adjustable positioning error threshold). Entity similarity calculations in the knowledge graph utilize the cosine similarity formula. In medication guidance, patients can scan drug packaging with AR to intuitively obtain information such as dosage, administration time, and contraindications. Pharmacists can use AR tools to view drug interaction risks and improve prescription review efficiency (AR application scenarios can be expanded).

[0071] In the combination of quantum encryption and zero-knowledge proof, quantum encryption deploys quantum key distribution (QKD) devices at core nodes in the medical insurance bureau and top-tier hospitals, with a key generation rate ≥1kbps, to encrypt and protect highly sensitive medical record data and medical insurance payment data. Zero-knowledge proof employs the zk-SNARKs algorithm. When insurance companies query patients' chronic disease treatment data, they only verify key information such as the type of chronic disease and medication compliance, without exposing the complete medical record. The verification formula is as follows: ,in For public keys, Information to be verified. To verify the document, a 1 is returned to indicate that the verification passed.

[0072] In the Web3.0 and DAO combination, decentralized applications are developed based on Ethereum smart contracts, using Solidity to write contracts. This enables functions such as on-chaining of patient health data, record-keeping of doctor treatment behavior, and product customization for insurance companies. The token incentive formula for the DAO organization is as follows: ,in For the number of tokens, As contribution value, As an incentive, the tokens can be redeemed for health management services or cash withdrawals.

[0073] The drug circulation traceability module utilizes technology deployed by affixing RFID tags to the smallest sales unit of the drug to record data across the entire supply chain, including production, distribution, delivery, and use. The Internet of Things (IoT) tracks the real-time status of drug logistics, and blockchain technology stores traceability information. Traceability verification ensures a unique link between the drug, prescription, and patient. The traceability verification formula is as follows: ,in This is the final traceability verification value. The data flow at each stage is traceable with an accuracy down to the single-item level.

[0074] The multimodal identity verification in the biometric fraud prevention module integrates facial recognition, voiceprint recognition, and medical behavior feature analysis technologies to construct a weighted fusion model. ,in To assess overall similarity, , , The similarity scores are calculated based on facial features, voiceprints, and behavioral characteristics, with a threshold of 0.85. Information from the public security population database is imported, and liveness detection is used to prevent impersonation and fraud. The accuracy rate of identity verification is ≥99.99%.

[0075] Through the above implementation plan, when a patient needs medical attention:

[0076] Patients can log in to the system using facial recognition and voiceprint verification via a smart screen at home or a mobile application, thus gaining real-name authentication.

[0077] Patients can choose to have an online follow-up consultation or visit a community health service center in person. The system will match the patient with a corresponding doctor based on their health record.

[0078] Doctors access patients' historical medical records, prescription records, and health monitoring data to issue electronic prescriptions using AR-assisted diagnosis and treatment tools. These electronic prescriptions are then submitted for review after the doctor's electronic signature.

[0079] The online prescription review team combines NLP engines and knowledge graphs to review prescriptions for compliance and reasonableness. Once approved, the prescriptions are pushed to the prescription circulation platform.

[0080] Patients can choose to pick up their medication at a community pharmacy or have it delivered to their home, and complete a one-stop settlement of medical insurance, commercial insurance, and personal expenses through a unified payment platform.

[0081] The smart pillbox reminds patients to take their medication on time, edge nodes monitor patients' health data in real time, triggering community doctors to intervene when abnormalities occur, and regularly pushing AR health guidance.

[0082] Through the above implementation and deployment, the risk control process for medical insurance is as follows:

[0083] Digital twins predict fund expenditure trends, federated learning models identify high-risk groups and institutions, and issue early warnings. When patients seek medical treatment, biometric modules verify their identity. When doctors issue prescriptions, intelligent review systems intercept illegal prescriptions in real time. During medical insurance settlement, the system verifies the compliance of expenses.

[0084] We conduct a full-coverage audit of medical insurance settlement data, use GNN models to uncover hidden fraudulent activities, generate risk control reports, and carry out offline audits for abnormal cases.

[0085] In summary, based on the unified big data model for digital medical cloud services and real-time medical insurance supervision of chronic diseases in the embodiments of this application, after the model was launched in a certain city for 6 months, the proportion of online consultations for chronic disease patients reached 40%, the consultation time was shortened from 3 hours to 15 minutes, the number of consultations at primary medical institutions increased by 35%, the hierarchical medical system achieved significant results, the illegal expenditure of medical insurance funds decreased by 22%, the identification rate of hidden fraud increased to 96%, the fund expenditure prediction error was ≤4%, the risk of fund overspending was warned 3 months in advance, the patient medication adherence increased from 65% to 90%, the medical treatment satisfaction rate reached 92%, the drug delivery time was ≤24 hours, the service coverage rate for patients with mobility difficulties reached 100%, the activity of DAO organization ecosystem participants increased by 70%, the claim efficiency of chronic disease insurance products of insurance companies increased by 50%, the accuracy rate of drug circulation traceability reached 100%, and no incidents of counterfeit drug circulation or drug substitution occurred.

[0086] In the description of this specification, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0087] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0088] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A unified large-scale model for digital medical cloud services and real-time medical insurance monitoring throughout the entire process of chronic disease treatment, characterized by: Including technical architecture and engine integration; The technical architecture includes an infrastructure layer, a data foundation layer, a core technology layer, a business application layer, and an ecosystem collaboration layer. The infrastructure layer provides elastically scalable computing, storage, and network resources, supporting collaborative scheduling between edge nodes and the cloud, as well as redundant backup of 5G private networks and public networks. The data foundation layer enables multi-source integration, standardized governance, and secure sharing of medical insurance data, medical diagnosis and treatment data, drug supply chain data, health record data, and insurance data, supporting interconnection of federated learning nodes and two-way verification between blockchain and data lake warehouses. The core technology layer provides the core technical capabilities required for intelligent identification, risk analysis, and convenient services, enabling real-time linkage between digital twins and AI models, and cross-empowerment of NLP and knowledge graphs. The business application layer supports two core businesses: full-process service for chronic disease treatment and real-time risk control and supervision of medical insurance, realizing the integration of AR-assisted diagnosis and treatment and telemedicine, as well as closed-loop management of smart medicine boxes and prescription circulation. The ecosystem collaboration layer enables system docking and business collaboration among multiple entities such as medical insurance bureaus, medical institutions, retail pharmacies, insurance companies, and logistics companies, supporting Web3.0 ecosystem governance and DAO collaboration mechanisms, as well as interoperability of cross-regional medical insurance data alliances. The combined engines include federated learning and blockchain, digital twins and AI and big data, edge computing and 5G and IoT, NLP and knowledge graphs and AR, quantum encryption and zero-knowledge proofs, and Web3.0 and DAO. Federated learning and blockchain are combined at the data foundation layer for cross-regional data collaboration modeling and privacy protection; digital twins, AI, and big data are combined at the core technology layer for medical insurance fund risk prediction and policy simulation; edge computing, 5G, and IoT are combined between the infrastructure layer and the business application layer for full-scenario data collection and real-time processing; NLP, knowledge graphs, and AR are combined at the core technology layer and the business application layer for intelligent diagnosis and medication guidance; quantum encryption and zero-knowledge proofs are combined at the data foundation layer and the ecosystem collaboration layer for secure transmission and privacy sharing of highly sensitive data; and Web3.0 and DAO are combined at the ecosystem collaboration layer to build a multi-party collaborative governance ecosystem.

2. The unified large-scale model for digital medical cloud service and real-time medical insurance supervision of the entire chronic disease process as described in claim 1, is characterized in that, The combination of federated learning and blockchain includes a federated learning framework and a consortium blockchain. The federated learning framework sets up provincial aggregation nodes and municipal participating nodes. Each participating node trains model parameters locally and only uploads updated parameter values. The consortium blockchain records each node's modeling permissions, parameter transmission trajectory, and model iteration versions.

3. The unified large-scale model for digital medical cloud service and real-time medical insurance supervision of the entire chronic disease process as described in claim 1, is characterized in that, The combination of digital twins with AI and big data includes a digital twin of the medical insurance fund, a deep learning model, and real-time data streams. It integrates medical insurance fund data, population data, chronic disease diagnosis and treatment data, and drug consumption data to construct a digital twin of the medical insurance fund. It uses time series models to predict the peak of short-term expenditures and the sustainability of medium and long-term fund expenditures, and uses graph neural networks to uncover hidden fraud patterns. Furthermore, the digital twin of the medical insurance fund supports multi-scenario simulation and deduction by inputting cost control policy parameters.

4. The unified large-scale model for digital medical cloud service and real-time medical insurance supervision of the entire chronic disease process as described in claim 1, characterized in that, The combination of NLP, knowledge graph, and AR includes a medical NLP engine, a chronic disease diagnosis and treatment knowledge graph, and AR visualization tools. The chronic disease diagnosis and treatment knowledge graph covers information on diseases, drugs, and treatment rules, and is updated in real time with medical insurance policies and clinical guidelines. The medical NLP engine parses unstructured diagnosis and treatment data and uses a multi-model voting mechanism to achieve intelligent prescription review. The AR visualization tools are used for remote consultation to assist in annotation and provide medication guidance for patients.

5. The unified large-scale model for digital medical cloud service and real-time medical insurance supervision of the entire chronic disease process as described in claim 1, characterized in that, The combination of edge computing with 5G and IoT includes edge gateways, 5G private networks, and multimodal IoT devices. Edge gateways are deployed in a hierarchical manner in community health service centers and smart screens at home, processing health monitoring data collected by multimodal IoT devices locally in real time and triggering early warnings. The 5G private network adopts a medical-specific slicing network to ensure that the data transmission latency for remote consultations is ≤50ms.

6. The unified large-scale model for digital medical cloud service and real-time medical insurance supervision of the entire chronic disease process as described in claim 1, characterized in that, It also includes a drug circulation traceability module, which uses blockchain, IoT and RFID technologies. The smallest sales unit of a drug is affixed with an RFID tag to record the entire circulation data. The IoT tracks the logistics status in real time, and the blockchain stores traceability information, realizing a unique binding between drugs, prescriptions and patients, with traceability accuracy down to the single-item level.

7. The unified large-scale model for digital medical cloud service and real-time medical insurance supervision of the entire chronic disease process as described in claim 1, characterized in that, The Web3.0 and DAO combination includes decentralized applications and distributed autonomous organizations. Decentralized applications support access from multiple nodes such as patients, doctors, medical institutions, and insurance companies. Distributed autonomous organizations allocate voting rights according to the contribution value of each party, and participate in rule-making, service evaluation, and risk reporting through token incentives. They also combine zero-knowledge proof technology to achieve data privacy sharing.

8. The unified large-scale model for digital medical cloud service and real-time medical insurance supervision of the entire chronic disease process as described in claim 4, is characterized in that, AR visualization tools are set up on the doctor's, patient's, and pharmacist's ends. On the doctor's end, AR overlays anatomical structure annotations to assist in remote consultation and diagnosis. On the patient's end, AR scans drugs to obtain medication information. On the pharmacist's end, AR presents the risk of drug interactions.

9. The unified large-scale model for digital medical cloud service and real-time medical insurance supervision of the entire chronic disease process as described in claim 4, is characterized in that, The combination of quantum encryption and zero-knowledge proof includes a quantum key distribution device and a zero-knowledge proof algorithm. The core node is equipped with a quantum key distribution device to encrypt the transmission of sensitive data. The zero-knowledge proof algorithm verifies the authenticity of key information only when data is authorized for query, without exposing complete private data.

Images