A home supported blood glucose self-management software

Abstract

The application discloses a blood sugar management applet with home support function. A data acquisition interaction module is connected with an external blood glucose meter device through a Bluetooth communication protocol to collect blood sugar data of a user in real time; a home collaborative management module realizes safe sharing of blood sugar data of the user among family members, and the family members can check relevant blood sugar data, diet and exercise records through the applet end; and an intelligent analysis and reminding module is based on an LSTM neural network model, performs trend analysis on historical blood sugar data after a risk early warning trigger, generates personalized health reminding information automatically in combination with preset rules, and sends the information to the user and associated family members in a multi-modal form. The application effectively improves the blood sugar management efficiency of patients, reduces the risk of data leakage, and can significantly reduce the fasting blood sugar value and the 2-hour postprandial blood sugar value of patients compared with the prior art, reduces the cost of manual intervention, and has significant clinical application value and social and economic benefits.

Description

Technical Field

[0001] This invention relates to the field of medical and health management and information technology, specifically to a family collaborative blood glucose management method and system based on dynamic access control, which is particularly suitable for daily blood glucose monitoring and family collaborative management scenarios for diabetic patients. Background Technology 1. Background of the Invention

[0003] Diabetes is a common chronic disease requiring long-term blood glucose monitoring and management. Currently, while some blood glucose management software and devices exist, they have several shortcomings in family-based collaborative management. On one hand, existing systems have relatively simple access control, failing to satisfy the patient's intrinsic motivation for self-determination—that is, allowing the patient to decide the access permissions of different family members regarding the patient's blood glucose data and task completion status. This may lead to patient privacy leaks or family members not being able to access crucial information in a timely manner. On the other hand, the lack of an effective family collaboration mechanism makes it difficult to fully leverage the supervisory and supportive roles of family members in the patient's blood glucose management process, resulting in low efficiency and an inability to complete health management activities at the micro-system level outside the individual. Furthermore, traditional blood glucose management systems mostly rely on simple rules for alerts in data processing and analysis, failing to incorporate advanced data analysis models to accurately predict blood glucose trends, analyze historical data, and provide further precise health advice. Therefore, there is an urgent need for a blood glucose management method and system that can achieve dynamic access control, efficient family collaboration, and intelligent data analysis.

[0004] 2. Comparison of the present invention with existing technologies

[0005] Before the application date, a search of the CNIPA database using the keywords "family support AND blood sugar" and "blockchain AND medical data access" yielded three existing patents similar to this invention: CN120072361A "A Diabetes Continuous Drug Therapy Management System", CN111145850A "Blockchain-based Medical Data Query Method and Device", and CN110797099A "A Blockchain-based Medical Data Sharing Method and System". A detailed comparison is shown in the table below:

[0006] CN120072361A describes a vertical, continuous management approach that is "hospital-led and drug-centered," while this invention is a horizontal, collaborative intervention approach for blood glucose behavior that is "patient-led and family-centered." The two differ significantly in terms of role systems, permission granularity, blockchain notarization, and task gamification, but they do not constitute a direct conflict and their technical paths are complementary. CN201911028237 and CN201911337806 are both aimed at cross-institutional sharing between hospitals, with the main data being large medical documents or medical records. This invention is aimed at home blood glucose management for patients and their family members, with the main data being time-series blood glucose values ​​and behavioral data. The scenarios, data scale, and real-time requirements are completely different.

[0007] This invention differs significantly from the aforementioned existing technologies in terms of role model (introducing family member users (family member roles), permission settings (RBAC), blockchain strategy (rights confirmation + incentives), data types (time-series blood glucose values ​​and behavioral data), and clinical validation. The scenarios are complementary and do not constitute a direct conflict. Summary of the Invention

[0008] (I) Purpose of the Invention

[0009] The purpose of this invention is to provide a family collaborative blood glucose management method and system based on dynamic access control. By establishing a flexible access control system, optimizing the family collaborative process, and introducing intelligent analysis technology, it can achieve secure sharing and efficient management of patient blood glucose data, improve the efficiency of patient blood glucose management, reduce the risk of data leakage, and provide patients with personalized health management solutions.

[0010] (II) Technical Solution

[0011] The family-based collaborative blood glucose management method based on dynamic access control described in this invention specifically includes the following steps:

[0012] 1. Establish family relationships

[0013] In response to the main user's (patient role) operation, a family group is created on the server side, and at least one family member user is associated with the family group based on the member contact information entered by the main user (patient role). During the association process, the identity information of the family member user needs to be verified. The identity information verification method is mobile phone number verification. If the verification fails, the association operation is rejected and an error message is sent to the main user (patient role).

[0014] 2. Configure dynamic permission rules

[0015] Each family member in the family group is assigned an independent set of access permissions for blood glucose data. This set includes at least one of the following: permission to view blood glucose data (all data, no abnormal blood glucose data, no all blood glucose records); permission to view behavioral data (can view plan completion data, no plan completion data); and permission based on preset blood glucose threshold ranges or time rules. This permission set is dynamically generated based on the settings of the primary user (patient role) and can be adjusted in real time by the primary user (patient role). Furthermore, only the primary user (patient role) can share text / images / videos. Family users can comment on and like the primary user's (patient role's) status. Comments and likes between family users are not visible to each other, ensuring the product remains centered on the primary user (patient role) and preventing the shifting of discussion focus.

[0016] 3. Blood glucose data synchronization and permission filtering

[0017] The system receives blood glucose data from the blood glucose monitoring device associated with the primary user (patient role) or manually entered data. Based on the real-time permission set configured for each family member user in step 2, the blood glucose data is dynamically filtered, and then encrypted using the AES-256 encryption algorithm. Each family member user has a unique decryption key, which is bound to their permission set. Data can only be decrypted and viewed using the corresponding decryption key when permission verification is successful. For data that passes permission verification, the system provides multi-dimensional data visualization: it supports generating historical blood glucose fluctuation curves by time dimension (day / week / month) and annotating event nodes such as diet, exercise, and medication; abnormal blood glucose data is highlighted with colors and animation effects, and a brief analysis of the reasons for the abnormality is provided.

[0018] 4. Targeted push and collaboration

[0019] Invitations from the primary user (patient role) to family member users (family member roles) are sent to the corresponding family member user's terminal via a collaborative task push interface. Family members can complete collaborative tasks shared by the patient, including dietary supervision tasks, exercise accompaniment tasks, and knowledge learning tasks. Specifically, the dietary supervision task involves the family member user receiving a recipe shared by the primary user (patient role) (this recipe is pushed to the primary user (patient role) by the system), uploading a photo of the food with a time watermark, identifying the food type using OCR technology, comparing it with the system's preset dietary plan, and completing the shared dietary requirements by uploading a photo or recording a text message to count as a successful check-in. The exercise accompaniment task involves the family member user receiving a daily exercise course shared by the primary user (patient role) (this exercise course is recommended to the primary user (patient role) by the system), synchronizing exercise data (such as steps and heart rate) to the system via wearable devices / WeChat Sports, completing the same exercise task, and checking in to verify the authenticity of the exercise. The knowledge learning task involves the family member user receiving a recipe shared by the primary user (patient role)... The courses shared by the main user (patient role) are mainly in the form of short videos. After watching, family members need to answer questions (the questions are the same as those for the main user (patient role)). The learning effect is verified through simple tests (such as multiple choice questions). After verifying the completion of tasks (daily tasks pushed by the system for the main member, and tasks selected by family members for the main member based on their own daily tasks), there is also a task incentive and feedback module: the task loop is closed according to "completed tasks / to be completed" after the main member completes the daily plan; if the main member achieves the target blood sugar three times in a row, the system adjusts the blood sugar target value, generates points for achieving the target, and distributes them to all family members. Points can be redeemed for health management-related rewards, such as virtual level titles or physical prizes; when a family member completes a collaborative task and passes the verification, the system accumulates points for that member, and the points can be redeemed for rewards in the same way; the system regularly generates a family task completion ranking and pushes it to all family members to promote healthy competition and collaborative enthusiasm among family members; the priority of tasks and the corresponding point weights are reset according to the user's task completion status.

[0020] 5. Specific implementation of the LSTM neural network model

[0021] 1) Training data and preprocessing

[0022] The training data comes from diabetes.csv, which contains historical blood glucose data of 5,000+ diabetic patients (collected every 30 minutes, covering a 1-year period), and is synchronously linked to diet records (100,000+ records, including food type, calories, and intake time), exercise data (steps, heart rate, exercise duration), and medication records (medication type, dosage, and time).

[0023] Except for equipment error values ​​(data with a deviation from the historical mean > 3 SD), linear interpolation was used to fill in missing values ​​for data cleaning.

[0024] Features such as blood glucose level (mmol / L) and number of steps (steps / day) are standardized to the [0,1] interval to avoid the impact of dimensional differences on model convergence;

[0025] Using a 7-day time window (336 time steps), a training sample was constructed consisting of "input sequence (data from the first 6 days) - output label (blood glucose value on day 7)".

[0026] 2) Model structure and parameters

[0027] Input layer (336-dimensional features), hidden layer (2 LSTM layers, 64 neurons per layer), dropout layer (dropout rate 0.2, to prevent overfitting), fully connected layer (16 neurons), output layer (1-dimensional, predicting blood glucose levels for the next 24 hours).

[0028] The core parameters are:

[0029] Learning rate: Initially 0.001, dynamically adjusted using the Adam optimizer;

[0030] Number of iterations: 50 rounds (early stopping mechanism: training is terminated if the validation set error does not decrease for 5 consecutive rounds);

[0031] Loss function: mean squared error (MSE), target convergence to <0.5 (mmol / L)².

[0032] 3) Prediction and Optimization Logic

[0033] Input your blood glucose data and behavioral events (diet / exercise) from the last 72 hours. The system will generate a trend curve for the next 36 hours using a sliding window (updated every hour) and output a 95% confidence interval (e.g., predicted value 7.2 mmol / L ± 0.3 mmol / L).

[0034] The ranking results are generated by calculating the contribution of key influencing factors (such as "2-hour postprandial blood glucose peak" to "high-sugar diet (weight 65%), lack of exercise (weight 25%), and medication delay (weight 10%)") and calculating the contribution of "high-sugar diet (weight 65%), lack of exercise (weight 25%), and medication delay (weight 10%)".

[0035] 6. Core Algorithm of Blockchain Smart Contracts

[0036] 1) Simplified code logic for permission change contracts

[0037] / / Consortium blockchain smart contracts (based on Hyperledger Fabric)

[0038] contractPermissionContract{

[0039] / / Define the permission structure: including change timestamp, primary user (patient role) ID, family member user (family member role ID), and comparison of old and new permissions.

[0040] structPermissionChange{

[0041] uint256timestamp;

[0042] stringownerID; / / Blockchain address of the primary user (patient role)

[0043] stringmemberID; / / Family member user (family member role blockchain address)

[0044] stringoldPermission; / / The set of permissions before the change (e.g., "Permission 5")

[0045] string newPermission; / / The changed permission set (e.g., "Permission 6")

[0046] }

[0047] / / Storage permission change record (Key: Unique credential ID, Value: Change details)

[0048] mapping(string=>PermissionChange)publicchangeRecords;

[0049] / / Permission change function (only callable by the primary user (patient role))

[0050] functionupdatePermission(stringmemory_memberID,stringmemory_oldPerm,stringmemory_newPerm)public{

[0051] / / Verify whether the caller is the primary user (patient role) (via blockchain address verification)

[0052] require(msg.sender==ownerAddress,"Non-primary user (patient role) has no permission to operate");

[0053] / / Generate a unique credential ID (hash value: primary user (patient role) ID + family member user (family member role) ID + timestamp)

[0054] stringmemoryrecordID=keccak256(abi.encodePacked(ownerID,_memberID,block.timestamp));

[0055] / / Store change records

[0056] changeRecords[recordID]=PermissionChange({

[0057] timestamp:block.timestamp,

[0058] ownerID:ownerID,

[0059] memberID:_memberID,

[0060] oldPermission:_oldPerm,

[0061] newPermission:_newPerm

[0062] });

[0063] / / Trigger event (notify consortium blockchain nodes to synchronize records)

[0064] emitPermissionUpdated(recordID,ownerID,_memberID);

[0065] }

[0066] }

[0067] Furthermore, this invention also relates to a family collaborative blood glucose management system based on the above method, including a data acquisition and interaction module, a family collaborative management module, an intelligent analysis and reminder module, and a security permission management module. The data acquisition and interaction module connects to an external blood glucose meter via Bluetooth communication protocol, collects user blood glucose data in real time, and uploads the data to a cloud server. The family collaborative management module uses the AES-256 encryption algorithm to achieve secure sharing of user blood glucose data among family members. Family members can view relevant blood glucose data, diet, and exercise records through a mini-program. The intelligent analysis and reminder module, based on an LSTM neural network model, performs trend analysis on historical blood glucose data and, combined with preset rules, automatically generates personalized health reminder information, which is sent to the user and associated family members in multiple modalities such as phone calls, SMS messages, and mini-program customer service push notifications. When performing targeted push notifications, if the push content involves blood glucose abnormality warnings, the push modality is dynamically selected according to the following rules: when the blood glucose value exceeds a preset emergency threshold, a warning pop-up is simultaneously triggered within the mini-program to the main user (patient role), and an SMS message containing blood glucose data is sent to family member users with "all data permissions"; when the blood glucose value exceeds the target range three times consecutively, based on the LSTM neural network model, a warning pop-up is sent to the main user (patient role). The system provides data access control for family members (family member roles) and related data permissions. It pushes images and text analysis to the mini-program, including support for generating blood glucose fluctuation curves and future trend predictions by time dimension (day / week / month), correlation analysis of triggering factors (based on blood glucose management behavior completion), further suggestions, and automatically generates personalized health reminders. The security permission management module is based on the RBAC (Role-Based Access Control) model. When a user registers for the first time, they select a role (patient or family member user (family member role)) and authorize a set of permissions for the family member user (family member role). Different roles and permissions have different data viewing and operation permissions. When the main user (patient role) adjusts the permission set of any family member, the system automatically generates a smart contract containing a timestamp of the permission change and a comparison of the permission content before and after the change. This smart contract is deployed to the consortium blockchain network and bound to the unique identifiers (such as blockchain addresses) of the main user (patient role) and family members, ensuring that all permission change records are immutable and traceable, and that key permission operations are solidified on the blockchain through smart contracts.

[0068] (III) Beneficial Effects

[0069] 1. Efficient family collaborative management

[0070] By establishing family relationships and configuring dynamic permission rules, differentiated access and operation of patient blood glucose data and blood glucose management behavior data by family members are realized, which fully mobilizes the enthusiasm of family members to participate in patient blood glucose management and improves the efficiency of family collaborative management.

[0071] 2. Robust data security protection

[0072] The AES-256 encryption algorithm is used to encrypt blood glucose data. Combined with dynamic access control based on the RBAC model and blockchain technology, key access operations are solidified and stored, which effectively reduces the risk of data leakage and protects the privacy and security of patient data.

[0073] 3. Precise intelligent analysis and personalized services

[0074] Based on an LSTM neural network model, historical blood glucose data is analyzed for trends. Combined with multi-source data, accurate blood glucose prediction is achieved, and personalized health reminders and suggestions are automatically generated. Simultaneously, multi-dimensional data visualization and task-based incentive feedback mechanisms provide patients and their family members with more intuitive and targeted health management services, helping patients better control their blood glucose and reduce the risk of complications. This demonstrates significant clinical application value and socio-economic benefits.

[0075] 4. Effective clinical results

[0076] This invention has been verified through actual clinical trials. In 94 elderly patients with type 2 diabetes, after 4 weeks of intervention, the experimental group showed significantly better blood glucose management behavior, fasting blood glucose level, postprandial blood glucose level, and perceived family support than the control group and the blank group (p<0.05). For details, please refer to Part (V) of the instruction manual for “Experimental Verification and Clinical Efficacy Evaluation”. Attached Figure Description

[0077] (one) Figure 1 Blockchain node deployment architecture diagram

[0078] The blockchain node deployment architecture diagram is divided into five layers from bottom to top: data layer, network layer, consensus layer, smart contract layer, application layer, and entity node layer. It illustrates the functional modules and interrelationships at different layers of the blockchain system, as explained below:

[0079] Data Layer: The foundation for storing data, including "distributed ledger" (storing permission change records and operation credential hashes) and "IPFS storage" (storing encrypted medical reports and task verification evidence), is the basis for on-chain data storage and traceability.

[0080] Network layer: Based on the "P2P network protocol", it enables peer-to-peer communication and data synchronization among nodes, ensuring the distributed connection and data transmission of the blockchain network.

[0081] Consensus layer: Relying on "supervisory nodes" and "PBFT consensus mechanism" (Practical Byzantine Fault Tolerance), data consistency is achieved among distributed nodes, solving the trust problem and ensuring the validity of transactions / data on-chain.

[0082] The smart contract layer contains clusters of smart contracts for "permission change, data audit, and points redemption," which automatically execute rules to automate business logic. For example, permission changes trigger the execution of the corresponding contract.

[0083] Application layer: Provides "permission management DApp" and "data query interface", connects to user / business systems, supports application functions such as permission control and data query, and serves as the entry point for user interaction.

[0084] Entity Node Layer: At the top layer, smartphones connect to the cloud service gateway via "HTTPS / API" as terminal nodes. The cloud service gateway is responsible for connecting with the lower application layer and forwarding requests.

[0085] The overall architecture, through a layered design, enables the collaboration of data storage, network communication, consensus mechanisms, business logic, and user interaction, building a blockchain system that connects the underlying data to the upper-layer applications, and is adapted to business scenarios such as permission management and data auditing.

[0086] (two) Figure 2 Dynamic permission adjustment and early warning logic diagram

[0087] Explanation of the dynamic permission adjustment and early warning logic diagram:

[0088] 1. User Registration

[0089] When users use the mini-program for the first time, they need to register and select a role (family member user (family member role or patient)).

[0090] 2. Character Selection

[0091] 2) Select family member user (family member role: no permission configuration, can only view information (status, collaborative tasks) shared by the main user (patient role), and cannot perform operations such as permission adjustment.

[0092] 3) Select Patient Role + Create Family Association: The main user (patient role) can enter the permission management interface to assign permissions to family members, view permission assignment details, and modify permissions.

[0093] 3. Permission Configuration

[0094] The primary user (patient role) assigns different sets of permissions to family members:

[0095] Permission 1: Cannot view all blood glucose data + Cannot view plan completion data.

[0096] Permission 2: Cannot view all blood glucose data + can view plan completion data.

[0097] Permission 3: Cannot view abnormal blood glucose data + Cannot view plan completion data.

[0098] Permission 4: Cannot view abnormal blood glucose data + Can view plan completion data.

[0099] Permission 5: Can view all blood glucose data + cannot view plan completion data.

[0100] Permission 6: View all blood glucose data + View plan completion data.

[0101] 4. Risk Triggering Conditions

[0102] For permissions 5 and 6, which grant the ability to view all blood glucose data, further risk trigger conditions can be set:

[0103] Risk trigger condition 1: Blood glucose level exceeds the preset emergency threshold.

[0104] Risk trigger condition 2: Blood glucose levels exceed the target range for 3 consecutive times.

[0105] Risk trigger condition 3: Blood glucose level exceeds the preset emergency threshold, or blood glucose level exceeds the target range 3 times consecutively.

[0106] 5. Risk Warning

[0107] For risk trigger condition 1: the primary user (patient role) receives a pop-up notification, and family member users (family member role) receive an SMS reminder.

[0108] For risk trigger conditions 2 and 3: the primary user (patient role) receives a pop-up notification, family member users (family member role) receive an SMS reminder, and the system pushes an LSTM analysis report.

[0109] (three) Figure 3 Multimodal verification technology flowchart

[0110] This multimodal validation flowchart, using a top-down tree structure, clearly illustrates the multimodal validation process involving the primary user (patient role) and family members around different types of tasks within the blood glucose management mini-program. The following is a detailed explanation of the flowchart:

[0111] 1. User Relationships

[0112] 1) Primary User (Patient Role): Located at the core top level of the flowchart, this user is the initiator and key decision-maker of the entire process, responsible for creating tasks and coordinating task management with family members. In the blood glucose management scenario, the primary user (patient role) is the patient themselves, who has the highest control and management authority over their own blood glucose data and related health management tasks.

[0113] 2) Family members: As collaborators of the main user (patient role), they are located at the bottom branch of the flowchart. They undertake various collaborative tasks assigned by the main user (patient role) and perform corresponding operations. They are important participants in the family collaboration model for blood glucose management and assist the main user (patient role) in better completing blood glucose management-related tasks.

[0114] 2. Task Classification and Verification Process

[0115] 1) Execute the task

[0116] Diet / Exercise / Knowledge / Glucose Monitoring / Medication: These are the core blood glucose management tasks that the primary user (patient role) needs to complete. Taking diet as an example, the primary user (patient role) needs to follow the system's prescribed diet plan for three meals a day; exercise involves executing the corresponding exercise plan; knowledge tasks involve learning about blood glucose management; blood glucose monitoring requires the primary user (patient role) to regularly monitor their blood glucose levels using a blood glucose meter or similar device; and medication tasks involve recording the user's medication intake as prescribed by their doctor.

[0117] 2) Publish collaborative tasks

[0118] Diet / Exercise / Knowledge Collaboration: While completing their own tasks, the primary user (patient role) can collaboratively assign some tasks to family members, corresponding to diet collaboration tasks, exercise collaboration tasks, and knowledge collaboration tasks. For example, in the diet collaboration task, the primary user (patient role) can invite family members to participate in the implementation of the diet plan, such as preparing healthy meals together; the exercise collaboration task involves inviting family members to participate in exercise together, providing mutual supervision and encouragement; and the knowledge collaboration task involves learning about blood glucose management together with family members, improving the whole family's awareness and ability in blood glucose management.

[0119] 3) Task Verification

[0120] Dietary task verification: You need to take a photo of your food and upload it to the mini-program. The system uses OCR technology to identify and extract key information such as food type from the photo, and then compares it with the diet plan to verify whether the dietary task meets the requirements.

[0121] Exercise data verification: For exercise tasks, users (primary user (patient role) and family member users assisting with the task (family member role users) record exercise data (such as steps, heart rate, etc.) through wearable devices or tools such as WeChat Sports and synchronize them to the system. The system analyzes and processes this exercise data to verify whether the exercise participation meets the task setting standards, such as whether the specified exercise duration has been completed and whether the exercise intensity matches the exercise data.

[0122] Knowledge Test Verification: In the knowledge tasks, after learning the course, students need to complete corresponding test questions (multiple-choice questions). The system automatically evaluates the submitted answers, verifying the student's mastery of the knowledge based on indicators such as the accuracy of the answers and the video playback progress bar, and determining whether the knowledge learning task has been completed.

[0123] 4) Results Feedback and Incentives

[0124] If the primary user (patient role) completes the task, points are awarded to each member of the entire family; if a family member user (family member role) completes the task, only the family member user (family member role) receives points, which can be redeemed for rewards.

[0125] (Four) Figure 4 Blockchain verification flowchart

[0126] The flowchart for implementing permission operation verification based on blockchain technology covers core aspects such as permission change, smart contract deployment, and consensus verification, and aligns with the permission storage and traceability logic in consortium blockchain scenarios:

[0127] 1) Triggering scenario

[0128] The blockchain verification process is initiated only when the primary user (patient role) initiates permission configuration or modification, in accordance with the "primary user (patient role) dominates permissions" rule in the RBAC model.

[0129] 2) Core Verification Link

[0130] Operation credential generation: solidifies key information about permission changes (timestamp, change content, etc.) to provide a basis for subsequent traceability.

[0131] Smart contract intervention: Automatically verify the legitimacy of the main user (patient role)'s identity through predefined contracts, and deny permission operations to non-main users (patient roles).

[0132] PBFT consensus mechanism: The consortium blockchain nodes reach a consensus through the Byzantine fault tolerance algorithm to ensure the consistency and immutability of permission change records (requiring confirmation from more than 2 / 3 of the nodes).

[0133] Distributed storage: The ledger stores only the hashes of key records (saving space), while the original vouchers are stored in IPFS, achieving efficient traceability through hash association.

[0134] 3) Query verification closed loop

[0135] When a family member (family member role) accesses data, the system automatically queries the blockchain ledger to verify whether the current permission matches the record. If they match, the data is decrypted using the bound key; otherwise, access is denied, thus achieving "traceable permission changes and verifiable data access".

[0136] 4) Integration with other modules

[0137] With the security and access management module: Blockchain records provide an immutable basis for access verification, and together with AES-256 encryption, they achieve dual security of "access + encryption".

[0138] Dynamic permission adjustment: This process is triggered every time permission is adjusted, ensuring that all dynamic changes are recorded on the blockchain, which meets the design requirement of "critical operations are fixed on the blockchain".

[0139] (five) Figure 5 LSTM report generation timing diagram

[0140] The above is a flowchart of LSTM report generation in a blood glucose management scenario. The core logic of the flowchart is explained below:

[0141] 1) Triggering mechanism

[0142] The LSTM report generation process is only initiated when blood glucose data meets risk criteria, thus avoiding unnecessary resource consumption.

[0143] 2) Data Link

[0144] From raw data input to preprocessing, noise is rigorously filtered and multi-source information (diet / exercise, etc.) is correlated to provide high-quality input for the LSTM model and ensure prediction accuracy.

[0145] 3) Model Core

[0146] By leveraging the collaborative computation of LSTM gating units, long-term dependencies in blood glucose data (such as the lagged correlation between diet and blood glucose) are captured, generating reliable predictions of future trends.

[0147] 4) Access Control Linkage

[0148] Before the report is pushed out, the recipient's permissions must be verified through blockchain to ensure that the main user (patient role) and family member users (family member role) can receive the report (only family member users with permissions 5 and 6 can receive the report, in accordance with dynamic permission rules).

[0149] 5) Closed-loop design

[0150] After terminal interaction, the system automatically adjusts the daily planned tasks based on the report's suggestions, forming a complete closed loop of "generation-application-iteration". Detailed Implementation

[0151] (I) System Setup

[0152] 1. Hardware parameters and network requirements

[0153] CPU: 8-core Intel Xeon E5-2670v3 (or equivalent AMD processor)

[0154] Memory: 32GB DDR4 2133MHz

[0155] Storage: 1TB SSD (read / write speed ≥ 500MB / s)

[0156] Operating System: Ubuntu 20.04 LTS 64-bit

[0157] Bandwidth: ≥100Mbps leased line, symmetrical uplink / downlink speeds

[0158] Latency: Network latency between nodes ≤ 50ms

[0159] Reliability: Supports dual-machine hot standby (automatic switching between primary and backup nodes)

[0160] 2. Server-side deployment

[0161] The system backend is built on a cloud server, and a database is configured to store user information, blood glucose data, permission rules, and other data. Blockchain nodes are deployed to build a consortium blockchain network for storing and verifying smart contracts for critical permission operations.

[0162] 3. Mini Program Development

[0163] Based on the WeChat Mini Program platform, a user-end Mini Program was developed to implement the functional interface and interaction logic of the data collection and interaction module, the family collaborative management module, the intelligent analysis and reminder module, and the security permission management module.

[0164] (II) Method Implementation Steps

[0165] 1. Establish family relationships

[0166] After family members and the main user (patient role) register for the mini-program for the first time, they select the corresponding role and fill in personal information, including contact information, on the "Settings" page. Then, the main user (patient role) enters the contact information of family member users (family member roles) on the "Manage Family" interface and submits, and the family relationship is successfully established.

[0167] 2. Deployment of Blockchain Smart Contracts

[0168]

[0169] 3. Configure dynamic permission rules

[0170] The primary user (patient role) accesses the permission management interface and sets separate access permissions for blood glucose data for each family member. For example, the primary user can set permissions for their spouse to view all blood glucose data, view behavioral data, and trigger conditions; and permissions for parents can set permissions to view blood glucose data (but not abnormal blood glucose data) and to view planned completion behavioral data. The primary user (patient role) can adjust the permissions for each family member at any time through the permission management interface as needed.

[0171] 4. Blood glucose data synchronization and permission filtering

[0172] Users measure their blood sugar using a Bluetooth blood glucose meter linked to the mini-program. After the measurement, the data acquisition and interaction module automatically acquires the blood glucose data via Bluetooth communication protocol and uploads it to the cloud server. The server dynamically filters and encrypts the blood glucose data based on the permission sets of each family member. After a family member logs into the mini-program, the system verifies their permissions. If the permission verification is successful, the system uses the corresponding decryption key to decrypt the encrypted data, displaying the data content that matches their permissions, and presenting it in the form of multi-dimensional visual charts.

[0173] 5. Targeted push and collaboration

[0174] The primary user (patient role) creates collaborative tasks in the mini-program, such as a diet monitoring task. They select the diet plan to be monitored and send an invitation to a designated family member. Upon receiving the task notification, family members complete the task within a specified time, such as taking and uploading photos of their food with a timestamp. The system uses OCR technology to identify the food type and compares it with the preset diet plan. Upon successful verification, the family member receives corresponding points, and the primary user's (patient role's) task loop closes accordingly. If the primary user (patient role) achieves target blood glucose levels three consecutive times, the system automatically generates reward points and distributes them to all family members. The system periodically generates a family task completion leaderboard and pushes it to each family member's mini-program.

[0175] When abnormal blood sugar levels occur, the intelligent analysis and alert module selects the appropriate push mode based on preset rules. If the blood sugar value exceeds the preset emergency threshold, the mini-program immediately sends an alert pop-up to the main user (patient role) and simultaneously sends a text message containing blood sugar data to family members with "all data permissions." If the blood sugar value exceeds the target range three times consecutively, an image and text analysis report is generated based on the LSTM neural network model and pushed to the mini-program terminals of the main user (patient role) and associated family member users (family member roles), providing users with accurate health management advice.

[0176] The above specific embodiments are merely examples of the present invention. In practical applications, the system and method can be appropriately adjusted and optimized according to specific needs, and all such adjustments and optimizations should fall within the protection scope of the present invention.

[0177] (III) Experimental Validation and Clinical Efficacy Evaluation

[0178] To verify the practical application effect of the system described in this invention, the research team conducted a 4-week intervention experiment in Beijing, which included 94 patients with type 2 diabetes aged 60 and above. They were divided into an experimental group (n=30, who used the mini-program of this invention and invited family members to participate (family member role collaboration), a control group (n=33, who only used the mini-program), and a blank group (n=31, who did not use the mini-program).

[0179] The results of this study on blood glucose management behavior, blood glucose levels, and family support are as follows:

[0180] 1. Improved blood glucose management behavior

[0181] After intervention, the scores of the experimental group patients on the Diabetes Self-Management Behavior Scale (DSCS) increased from 65.33±11.72 to 71.03±10.01, which were significantly higher than those of the control group and the blank group (significant interaction effect, F=11.597, p<0.001).

[0182] 2. Improved blood glucose levels

[0183] The fasting blood glucose level in the experimental group decreased significantly from 7.15±0.90 mmol / L before intervention (significant interaction effect, F=4.558, p=0.013); the 2-hour postprandial blood glucose level decreased from 11.07±2.34 mmol / L to 10.43±1.95 mmol / L, which was better than that in the control group and the blank group.

[0184] 3. Enhanced perception of family support

[0185] The family support scale score of the experimental group increased from 27.57±2.58 to 28.70±1.84, which was significantly higher than that of the control group (significant interaction effect, F=4.111, p=0.020).

[0186] The final conclusion is that the system described in this invention significantly improves the self-management behavior, blood glucose control level, and perceived family support of elderly diabetic patients through functions such as family collaboration mechanism, dynamic permission control, task incentives, and data visualization, and has good clinical application prospects.

Claims

1. A family-based collaborative blood glucose management method based on dynamic access control, characterized in that, Includes the following steps: Establish Family Association: In response to the main user's (patient role) actions, a family group is created on the server side. At least one family member user is associated with the family group via mobile phone verification. The identity information of the family member user is bound to the unique identifier of the family group. Dynamic Permission Rules Configuration: Based on the RBAC model, when a user registers for the first time, they select a role. If they choose to be a family member user (relative), they do not have the right to configure dynamic permissions. If they choose to be a patient (main user (patient role)), after establishing the family association, they need to assign an independent set of blood glucose data access permissions to each family member user in the family group. This permission set includes permissions to view blood glucose data and permissions to view behavioral data. Viewing blood glucose data permissions include all blood glucose data, no viewing of abnormal blood glucose data, and no viewing of all blood glucose records. For family members with permission to view all blood glucose data, the main user (patient role) can also set risk trigger condition permissions. These risk trigger condition permissions include those based on a preset blood glucose threshold range, time rules, or both blood glucose thresholds and time rules. Viewing behavioral data permissions include those that allow viewing of planned completion data and those that do not allow viewing of planned completion data. Viewing completed data, the permission set is dynamically generated based on the settings of the main user (patient role) and supports real-time adjustment; Blood glucose data synchronization and permission filtering: Blood glucose data is received from blood glucose monitoring devices associated with the main user (patient role) or manually entered, encrypted using the AES-256 encryption algorithm, and stored on the cloud server. The blood glucose data is dynamically filtered according to the configured real-time permission set; Targeted push and collaborative management: When blood glucose data meets the configured risk trigger conditions, trend analysis of historical blood glucose data is performed based on the LSTM neural network model to generate a graphic report containing blood glucose fluctuation curves, correlation analysis of triggering factors, and suggestions. The graphic report and warning information are targeted and pushed to the main user (patient role) and family member users with permission to view all blood glucose data. Push methods include mini-program pop-ups, SMS, and mini-program customer service information. The main user (patient role) pushes dietary supervision, exercise accompaniment, or knowledge learning collaborative tasks to family member users. Family member users complete the tasks by verifying dietary photos with OCR technology, synchronizing exercise data with wearable devices, or taking knowledge tests. The task completion status is fed back through points.

2. The family collaborative blood glucose management method based on dynamic access control according to claim 1, characterized in that, In the step of establishing family association, when associating a family member user with the family group, it is necessary to verify the identity information of the family member user. The identity information verification method is mobile phone number verification. If the verification fails, the association operation is rejected and an error message is sent to the main user (patient role).

3. The family collaborative blood glucose management method based on dynamic access control according to claim 1, characterized in that, In the blood glucose data synchronization and permission filtering process, when dynamically filtering blood glucose data, the filtered data is also encrypted. The encryption method adopts the AES-256 encryption algorithm, and each family member user has an independent decryption key. This decryption key is bound to the permission set of the family member user, and the data can only be decrypted and viewed when the permission verification is successful.

4. The family collaborative blood glucose management method based on dynamic access control according to claim 1, characterized in that, The collaborative tasks include dietary supervision tasks, exercise accompaniment tasks, and knowledge learning tasks. The dietary supervision task involves family member users receiving recipes shared by the main user (patient role), and completing the shared dietary requirements together by taking photos and uploading them or recording them in text, which counts as a successful check-in. The exercise accompaniment task involves family member users receiving the daily exercise course shared by the main user (patient role), and family members completing the same exercise tasks and checking in. The knowledge learning task involves family member users receiving courses shared by the main user (patient role), and family members answering questions after watching the courses.

5. The family collaborative blood glucose management method based on dynamic access control according to claim 1, characterized in that, Only the primary user (patient role) can share status within the defined roles. Family users can comment on and like the primary user's (patient role's) status. Comments and likes between family users are not visible to each other.

6. The family collaborative blood glucose management method based on dynamic access control according to claim 1, characterized in that, When the primary user (patient role) adjusts the permission set of any family member, the system automatically generates a smart contract containing a timestamp of the permission change and a comparison of the permission content before and after the change. The smart contract is deployed to the consortium blockchain network and bound to the unique identity of the primary user (patient role) and family members to ensure that all permission change records are immutable and traceable.

7. The family collaborative blood glucose management method based on dynamic access control according to claim 1, characterized in that, When executing targeted push notifications for abnormal blood glucose levels, the push mode is dynamically selected based on risk triggering conditions: based on a preset blood glucose threshold range, i.e., when the blood glucose value exceeds the preset emergency threshold, or based on time rules, i.e., when the blood glucose value exceeds the target range three times consecutively. After the risk warning is triggered: a warning pop-up is sent to the main user (patient role) within the mini-program, and a text message containing blood glucose data is sent to family members with all data permissions. Based on the LSTM neural network model, graphic analysis is pushed, including support for generating blood glucose fluctuation curves by time dimension, future trend prediction, correlation analysis of inducing factors, and further suggestions.

8. The family collaborative blood glucose management method based on dynamic access control according to claim 1, characterized in that, The multimodal collaborative task verification mechanism includes: Dietary supervision task: Upload food photos with time watermarks, identify food types using OCR technology, and compare them with the system's preset diet plan; Exercise coaching task: Synchronize exercise data to the system via wearable devices / WeChat Sports to verify the authenticity of the exercise; Knowledge learning task: After completing a short video course, a simple test is required to verify the learning effect.

9. The family collaborative blood glucose management method based on dynamic access control according to claim 1, characterized in that, Following the multimodal collaborative task verification mechanism step, a task incentive and feedback module is also included: After the main member completes the daily plan, the task loop closure is determined based on the completed / to-be-completed tasks; if the main member achieves the target blood glucose level three times consecutively, the system optimizes the blood glucose target based on the data, generates points rewards upon achievement, and distributes them to all family members; these points can be redeemed for health management-related rewards; when a family member completes a collaborative task and passes verification, the system accumulates points for that member; the system periodically generates a family task completion ranking and pushes it to all family members; and the priority and weight of tasks are reset based on user task completion status.

10. The family collaborative blood glucose management method based on dynamic access control according to claim 1, characterized in that, After the blood glucose data synchronization and permission filtering steps, the system provides multi-dimensional data visualization for data that has passed permission verification: it supports generating blood glucose fluctuation curves by time dimension and annotating event nodes such as diet, exercise, and medication; for abnormal blood glucose data, it highlights them with bright colors and animation effects and provides brief analysis and prompts on the reasons for the abnormality.

Images