Diabetic patient data management method and system based on big data

Abstract

The invention discloses a diabetic patient data management method and system based on big data. The method comprises the steps that basic information and stable blood glucose data of a patient are collected, similar cases are matched through a cloud database to obtain influence factors, and a personalized data collection scheme is generated; customizing a user terminal and establishing a data management container, adaptively acquiring blood glucose and influence factor data, calculating a change degree through a sliding window algorithm and performing early warning; and dynamically updating influence factor weights by adopting a time-varying weight model, and generating an intervention strategy according to the weights. The system comprises a cloud server, a comprehensive management terminal, a user terminal and other modules, and realizes data acquisition-analysis-intervention closed-loop management. According to the scheme, through big data analysis and the dynamic weight model, personalized accurate management of the diabetic patient is realized, the acquisition strategy and the intervention scheme can be adjusted according to individual differences, and the risk early warning accuracy and the management efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the field of healthcare, and specifically to a method and system for managing diabetes patient data based on big data. Background Art

[0002] Traditional diabetes management methods mainly rely on patients' self-monitoring and regular medical visits, and there are many limitations. On the one hand, patients manually record information such as blood glucose data, diet, and exercise. This is not only cumbersome but also prone to incomplete or inaccurate recording, making it difficult to comprehensively reflect the true situation of patients. On the other hand, doctors make diagnoses and formulate treatment plans based on limited outpatient data, lacking real-time and comprehensive understanding of patients' daily conditions, and it is difficult to achieve personalized and precise management. In addition, traditional methods are difficult to effectively capture the complex relationships between blood glucose fluctuations and various influencing factors, and are unable to detect potential risk factors in a timely manner and intervene. In recent years, with the rapid development of technologies such as big data, the Internet of Things, and artificial intelligence, new ideas and means have been provided for diabetes management. Big data technology can integrate a large amount of patient data, including blood glucose monitoring data, physiological indicators, living habits, environmental factors, etc., and deeply analyze and mine the laws and patterns behind the data. The Internet of Things technology enables the interconnection of various sensors, making real-time and continuous data collection possible. Artificial intelligence algorithms can analyze and process complex data, build accurate prediction models, and provide support for personalized diabetes management. However, existing diabetes data management systems still have some deficiencies. For example, the data collection method is single and cannot comprehensively cover all factors affecting blood glucose; the data processing and analysis capabilities are limited, and it is difficult to meet the real-time processing requirements of a large amount of data; there is a lack of a dynamic adjustment and personalized management strategy generation mechanism, making it difficult to adapt to individual differences and dynamic changes in patients' conditions. Therefore, there is an urgent need for a method and system for managing diabetes patient data based on big data to achieve more comprehensive, precise, and dynamic diabetes management. Summary of the Invention

[0003] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a method for managing diabetes patient data based on big data, including the following steps: Step 1, a comprehensive management terminal collects the basic information of a patient and blood glucose data in a stable state. According to the basic information of the patient, the patient's blood glucose influencing factors are obtained from a cloud data server. According to the patient's blood glucose influencing factors, the information of the corresponding influencing factor data collection module is obtained, and an influencing factor data collection plan corresponding to the patient is generated; Step 2: According to the influencing factor data collection scheme of the corresponding patient, generate the user terminal of the corresponding patient. The comprehensive management terminal generates a patient data management container, communicates with the user terminal of the corresponding patient, and sends the blood glucose data of the patient in a stable state to the patient data management container. Step 3: The user terminal of the corresponding patient collects the patient's blood glucose data and influencing factor data according to the collection period, and obtains the degree of change in blood glucose data at the start and end times of the influencing factors based on the blood glucose data and influencing factor data; conducts risk early warning according to the degree of change in blood glucose data, and sends the influencing factor data and the degree of change in blood glucose data at the start and end times of the influencing factors to the patient data management container. Step 4: The patient data management container obtains the weight of the influencing factor on blood glucose data based on the influencing factor data obtained in different periods and the degree of change in blood glucose data at the start and end times of the influencing factors. Step 5: Generate a patient blood glucose management strategy according to the weights of various influencing factors on blood glucose data obtained, and complete the data management of diabetic patients.

[0004] Furthermore, the comprehensive management terminal collects the basic information of the patient and the blood glucose data in a stable state, and obtains the patient's blood glucose influencing factors from the cloud data server according to the patient's basic information, including: Collect the patient's identity information, medical history records, medication records, and the set fasting / postprandial blood glucose benchmark values for consecutive n hours through a standardized form. Call the patient database of the cloud data server for similar case matching, and obtain a set of influencing factors with a correlation greater than the threshold for the current patient's blood glucose fluctuations. The influencing factors include: dietary components, exercise intensity, sleep quality, medication compliance, environmental temperature and humidity, and emotional fluctuation indicators. Establish a list of priorities for the patient's individual influencing factors. The priority is calculated by a dynamic weighting algorithm, and the initial weight is a preset value.

[0005] Furthermore, obtaining the corresponding influencing factor data collection module information according to the patient's blood glucose influencing factors and generating the influencing factor data collection scheme for the corresponding patient includes: Construct a module configuration mapping table to associate each type of influencing factor with the corresponding collection module; generate a collection module activation sequence according to the priority list, and set the high-priority collection module to the high-frequency sampling mode; and generate a collection module configuration file, including the sampling frequency, data format standard, and outlier determination threshold.

[0006] Furthermore, the user terminal of the corresponding patient collects the patient's blood glucose data and influencing factor data according to the collection period, and obtains the degree of change in blood glucose data at the start and end times of the influencing factors, including: Adopt an adaptive sampling period: The basic period is the first set duration. When the detected blood glucose change rate exceeds the change rate threshold, encrypted sampling is performed. The encrypted sampling period is the second set duration, and the second set duration is less than the first set duration; The sliding window algorithm is used to calculate the degree of change. The window length is set to twice the action period of the influencing factor. The degree of change is: ΔG = Σ|G(t+τ)-G(t)| / (τ_max - τ_min) where G(t) is the blood glucose value and τ is the action duration of the influencing factor.

[0007] Furthermore, the risk warning based on the degree of change in blood glucose data includes: According to different degrees of change in blood glucose data, the corresponding warning strategy is called in the cloud data server, and the patient is warned according to the warning strategy.

[0008] Furthermore, the patient data management container obtains the weight of the influencing factor on blood glucose data based on the influencing factor data obtained in different periods and the degree of change in blood glucose data at the start and end times of the influencing factor, including: A time-varying weight calculation model is used to calculate the weight of the influencing factor on blood glucose data, and the weight is updated as: W_i(t) = α·ΔG_i(t) / ΣΔG_j(t) + (1-α)·W_i(t-Δt) where α is the forgetting factor (0 < α < 1), Δt is the update period; W_i(t) is the dynamic influence weight of the influencing factor i on blood glucose data at time t, representing the contribution ratio of the i-th influencing factor to blood glucose fluctuations at time t. The sum of the weights of all influencing factors ΣW_i(t)=1; ΔG_i(t) is the blood glucose change amount caused by the influencing factor i at time t; ΣΔG_j(t) is the total blood glucose change amount caused by all influencing factors at time t; W_i(t-Δt) is the historical weight of the influencing factor i in the previous calculation period (t-Δt).

[0009] Furthermore, the generation of the patient blood glucose management strategy based on the obtained weights of each influencing factor on blood glucose data includes: An intervention priority list of influencing factors is generated according to the weights of the influencing factors on blood glucose data. For the influencing factors with weights greater than the weight threshold, the corresponding intervention strategies are obtained in the cloud data server. The intervention strategies corresponding to each influencing factor with a weight greater than the weight threshold constitute the patient blood glucose management strategy.

[0010] A diabetes patient data management system based on big data, which applies the described diabetes patient data management method based on big data, includes a cloud data server, a comprehensive management terminal, a communication module, a user terminal, a risk assessment module, and a health status tracking module; The cloud data server, the comprehensive management terminal, the user terminal, the risk assessment module, and the health status tracking module are respectively communicatively connected to the communication module; The cloud data server is used to provide a patient database and a policy database; The comprehensive management terminal is used to customize the influencing factor data collection module in the user terminal according to the influencing factors of the patient's health data, and manage the user management data and the user terminal; The user terminal is used to collect the management data of the patient; The risk assessment module is used to perform risk assessment according to the management data of the patient; the risk assessment includes a health risk assessment module and a data integrity assessment; The health status tracking module is used to track the patient's health status according to the patient management data and the risk assessment.

[0011] Preferably, the user terminal includes an influencing factor data collection module, an identity information module, a location information collection module, a management data generation module, a data processing module, and a communication device; The influencing factor data collection module, the identity information module, the location information collection module, the management data generation module, and the communication device are respectively connected to the data processing module; The influencing factor data collection module is used to collect data on the influencing factors that affect the patient's blood glucose index; The identity information collection module is used to collect the patient's identity information; The location information collection module is used to collect the location where the patient is located; The management data generation module is used to generate user health management data according to the data collected by the user terminal within the management cycle.

[0012] The beneficial effects of the present invention are as follows: By collecting the patient's basic information and blood glucose baseline value through a standardized form, the standardization and integrity of the initial data are ensured. The personalized customization of the influencing factor data collection scheme determines the collection module and sampling frequency according to the specific situation of the patient, realizing the comprehensive collection of various factors affecting blood glucose. The adaptive sampling period mechanism can increase the sampling frequency when the blood glucose changes violently, timely capture the rapid changes in blood glucose, and improve the accuracy and real-time of blood glucose data collection. At the same time, the determination and processing of outliers during the data collection process further ensure the quality of the data. Through similar case matching and dynamic weighting algorithms, a personalized list of influencing factors priorities for patients is established, enabling data collection and management to be tailored to the specific circumstances of the patients. The time-varying weight calculation model continuously updates the weights of influencing factors based on real-time data, reflecting the dynamic changes in the patient's condition. The personalized management strategy generated based on the weights can provide precise intervention measures for factors that have a greater impact on the patient's blood glucose, improving the pertinence and effectiveness of diabetes management. By calculating the degree of change in blood glucose data through a sliding window algorithm, the full process impact of influencing factors on blood glucose changes can be comprehensively captured. Based on the degree of blood glucose change, the corresponding early warning strategy is invoked to achieve timely early warning of blood glucose risks. The risk assessment module not only focuses on the changes in blood glucose itself but also comprehensively considers various influencing factors, improving the accuracy and comprehensiveness of risk assessment. Timely risk early warning can help patients and medical staff take measures in a timely manner to reduce the risk of complications. The system integrates multi-dimensional data such as the patient's blood glucose data, influencing factor data, medical history records, and medication records, and uses big data analysis technology to mine the laws and patterns behind the data. The time-varying weight calculation model can dynamically analyze the complex relationship between influencing factors and blood glucose, providing data support for in-depth understanding of the pathogenesis and development laws of diabetes. The health status tracking module analyzes long-term data to draw a curve of health status changes, providing a strong basis for evaluating the treatment effect and adjusting the management strategy. Description of the Drawings

[0013] Figure 1 It is a schematic flowchart of a method for managing data of diabetic patients based on big data. Detailed Implementation Manner

[0014] The technical solution of the present invention will be further described in detail below in conjunction with the drawings, but the protection scope of the present invention is not limited to the following description.

[0015] The features and performance of the present invention will be further described in detail below in conjunction with embodiments.

[0016] As Figure 1 shown, a method for managing data of diabetic patients based on big data Step 1: The comprehensive management terminal collects patient information and generates a data collection plan for influencing factors As the starting point of data collection and management, the comprehensive management terminal undertakes the important task of obtaining the patient's basic information and initial blood glucose data. When collecting the patient's basic information, a carefully designed standardized form is used to ensure the standardization and integrity of the data. The standardized form covers multiple key dimensions: Patient identity information: including name, age, gender, contact information, ID number, etc. These information are the basis for identifying patients and ensure the accurate correspondence between the data and the patients. Medical history record: Record in detail the onset time of the patient's diabetes, past medical history, complication status, etc., providing important historical references for subsequent disease analysis and management. Medication record: Record information such as the types, dosages, and administration times of the hypoglycemic drugs currently used by the patient. The impact of drug factors on blood glucose is crucial, and accurate medication records are the basis for subsequent analysis. Blood glucose baseline value: Set the fasting / postprandial blood glucose baseline value for consecutive n hours. Here, n hours need to be determined according to medical standards and clinical practice. Usually, a duration that can comprehensively reflect the patient's blood glucose stability over a period of time is selected, such as the fasting and postprandial blood glucose data for 72 consecutive hours. By continuously collecting data, the impact of accidental factors on blood glucose values can be reduced, and more representative baseline data can be obtained. After obtaining the patient's basic information, the comprehensive management terminal needs to obtain the blood glucose influencing factors of the patient from the cloud data server according to this information. The cloud data server stores a large amount of patient data and medical knowledge, and similar case matching is performed by calling the patient database therein. Here, advanced machine learning algorithms are used to identify similar cases. For example, clustering algorithms can be used to group patients with similar characteristics (such as age, gender, medical history, blood glucose level, etc.) into one category. By calculating the feature similarity between the current patient and other patients in the database, a set of influencing factors with a greater correlation with the blood glucose fluctuation of the current patient than the preset threshold is obtained. These influencing factors mainly include the following categories: Dietary components: Different food components have significant differences in their impact on blood glucose, such as the content and types of carbohydrates. Exercise intensity: Exercise is an important factor affecting blood glucose. The intensity, duration, and mode of exercise will all have different impacts on blood glucose. Sleep quality: Lack of sleep or poor sleep quality may lead to blood glucose fluctuations. Drug compliance: Whether the patient takes the medicine correctly according to the doctor's advice directly affects the blood glucose control effect. Environmental temperature and humidity: Environmental factors may indirectly affect blood glucose by influencing the patient's physiological state. Emotional fluctuation index: Severe emotional fluctuations may cause changes in the body's hormone levels, thereby affecting blood glucose. After obtaining the set of influencing factors, it is necessary to establish an individualized list of influencing factor priorities for the patient. This process is achieved through a dynamic weighting algorithm, and the initial weights are preset based on medical research and clinical experience. The dynamic weighting algorithm will consider various factors to adjust the weights, such as the influence degree of different influencing factors in similar cases and the individual characteristics of the patient. In this way, the relative importance of each influencing factor can be determined according to the specific situation of the patient, providing a basis for generating subsequent data collection plans. Next, according to the blood glucose influencing factors of the patient, obtain the corresponding information of the influencing factor data acquisition module and generate the corresponding acquisition plan. First, construct a module configuration mapping table to associate each type of influencing factor with the corresponding acquisition module. For example, the acquisition of dietary components may need to be associated with the diet record module, and the acquisition of exercise intensity needs to be associated with the exercise monitoring module, etc. Then, generate the activation sequence of the acquisition module according to the previously established priority list. For the acquisition modules with high priority, set them to the high-frequency sampling mode to ensure that the data of key influencing factors can be obtained more frequently. At the same time, generate a detailed acquisition module configuration file, which includes important parameters such as sampling frequency, data format standard, and outlier determination threshold. The setting of the sampling frequency needs to be determined according to the characteristics of the influencing factor and medical requirements. For example, for factors that are more sensitive to blood glucose changes, a higher sampling frequency may be required; the unification of data format standards helps with subsequent data processing and analysis; the outlier determination threshold is used to identify and filter abnormal data that may occur during the acquisition process. Step 2: Generate the user terminal and establish the data management container According to the influencing factor data acquisition plan generated in Step 1, generate a dedicated user terminal for the corresponding patient. The generation of the user terminal needs to consider the hardware and software configurations to meet the functional requirements and data processing requirements of different acquisition modules. In terms of hardware, it is necessary to integrate corresponding sensors and communication modules, such as a blood glucose meter sensor for collecting blood glucose data, an acceleration sensor for collecting exercise intensity, etc.; in terms of software, it is necessary to install an adapted operating system and application programs to achieve the functions of data collection, processing, and transmission. At the same time, the comprehensive management terminal generates a patient data management container. The patient data management container is a logical unit specifically used for storing and managing patient data. It can be constructed based on cloud computing technology and has powerful data storage and processing capabilities. The data management container adopts an advanced database architecture, such as a distributed database, to cope with the storage and management requirements of massive data and ensure the security and reliability of the data. Then, establish a communication connection between the comprehensive management terminal and the corresponding patient user terminal. The communication module adopts a secure and reliable communication protocol, such as the MQTT (Message Queuing Telemetry Transport) protocol, which has the characteristics of being lightweight, low-power, and suitable for Internet of Things device communication. During the communication process, encrypt the data and use encryption algorithms such as AES (Advanced Encryption Standard) to ensure the security and privacy of the data during transmission. After completing the communication connection, the integrated management terminal sends the blood glucose data of the patient in a stable state to the patient data management container. These initial blood glucose data serve as the baseline data of the patient, providing a reference for subsequent data analysis and management. Before being sent, the data needs to be formatted and verified to ensure its accuracy and integrity. Step 3: The user terminal collects data, analyzes the degree of change, and issues a risk warning The user terminal of the patient collects the patient's blood glucose data and influencing factor data according to the set collection period. To improve the efficiency and accuracy of data collection, an adaptive sampling period mechanism is adopted. The basic period is set to the first set duration, which is determined according to the requirements of medical routine monitoring and the actual needs of data collection. For example, it can be set to 30 minutes. When it is detected that the blood glucose change rate exceeds the preset change rate threshold, the encrypted sampling mode is automatically triggered. At this time, the encrypted sampling period is the second set duration, and the second set duration is less than the first set duration. For example, it can be set to 5 minutes. This adaptive sampling method can reduce the sampling frequency, power consumption, and data volume when the blood glucose changes relatively smoothly; while increasing the sampling frequency when the blood glucose changes violently, timely capturing the rapid changes in blood glucose, and improving the real-time and accuracy of the data. After obtaining the blood glucose data and influencing factor data, it is necessary to obtain the degree of change in blood glucose at the start and end times of the influencing factor based on these data. Here, a sliding window algorithm is used to calculate the degree of change, and the window length is set to twice the action period of the influencing factor. This is because considering that the influence of the influencing factor on blood glucose may have a certain delay and duration, setting the window length to twice the action period can more comprehensively capture the entire process of the influence of the influencing factor on blood glucose changes. The formula for calculating the degree of change is: ΔG = Σ|G (t+τ)-G (t)| / (τ_max - τ_min) Among them, G (t) represents the blood glucose value at time t, τ is the action duration of the influencing factor, and τ_max and τ_min are the maximum and minimum values of the action duration of the influencing factor respectively. By calculating the sum of the absolute changes in blood glucose values within the sliding window and dividing by the range of the action duration, a quantitative index that can reflect the degree of blood glucose change is obtained. Based on the calculated degree of change in blood glucose data, risk warnings are issued. First, warning strategies corresponding to different degrees of blood glucose change are pre-stored in the cloud data server. These warning strategies are formulated based on medical knowledge and clinical experience, including different warning levels (such as low risk, medium risk, high risk) and corresponding warning measures. When the user terminal calculates the degree of blood glucose change, it sends it to the cloud data server. The server calls the corresponding warning strategy according to the degree of change and warns the patient through the user terminal. The warning methods can include voice prompts, APP push notifications, SMS reminders, etc., to ensure that the patient can timely understand their own blood glucose risk status. Step 4: Calculate the weight of influencing factors on blood glucose data The patient data management container receives the influencing factor data of different cycles and the corresponding degree of change in blood glucose data from the user terminal, and then uses a time-varying weight calculation model to calculate the weight of the influencing factors on the blood glucose data. The formula of the time-varying weight calculation model is: W_i (t) = α・ΔG_i (t) / ΣΔG_j (t) + (1-α)・W_i(t-Δt) Where α is the forgetting factor, and its value range is between 0 and 1, which is used to adjust the weight ratio of new data and historical data. When α is larger, it means that more attention is paid to the influence of current data; when α is smaller, the influence of historical data is relatively greater. Δt is the update period, which is determined according to the data collection frequency and actual needs. For example, it can be set to 1 day. W_i(t) represents the dynamic influence weight of the influencing factor i on the blood glucose data at time t, which reflects the contribution ratio of the i-th influencing factor to blood glucose fluctuations at time t. ΣΔG_j (t) is the total amount of blood glucose change caused by all influencing factors at time t. By dividing the blood glucose change amount ΔG_i (t) corresponding to each influencing factor by the total change amount, the relative contribution ratio of this influencing factor in the current cycle is obtained, and then multiplied by the forgetting factor α, plus the historical weight W_i (t-Δt) multiplied by (1-α), to obtain the updated weight. This time-varying weight calculation model has the following advantages: it can continuously update the weight according to new data over time, reflecting the dynamic changes in the influence of influencing factors on blood glucose; the introduction of the forgetting factor enables the model to adapt to changes in the patient's condition and the emergence of new influencing factors, improving the flexibility and adaptability of the model; at the same time, by considering the historical weight, the continuity and stability of weight changes are maintained, avoiding drastic changes in weight caused by individual data fluctuations. In the process of calculating weights, it is necessary to preprocess the data, including operations such as outlier removal and data smoothing, to reduce the impact of data noise on weight calculation. Statistical methods, such as the 3σ principle, can be used to identify and remove outliers; methods such as moving average can be used to smooth the data and improve the data quality. Step Five: Generate a blood glucose management strategy for the patient Generate a blood glucose management strategy for the patient based on the weights of various influencing factors on blood glucose data obtained in Step Four. First, sort the influencing factors according to their weights to generate a list of intervention priorities for the influencing factors. The greater the weight of an influencing factor, the greater its impact on blood glucose, so it should be given priority in intervention. Set a weight threshold, and use the influencing factors with weights greater than this threshold as the main intervention targets. The setting of the weight threshold needs to combine medical knowledge and clinical experience. For example, it can be set to 0.1, that is, the influencing factors with weights exceeding 0.1 are regarded as factors having a significant impact on blood glucose. Then, for each influencing factor with a weight greater than the weight threshold, obtain the corresponding intervention strategy in the strategy database of the cloud data server. The strategy database stores a wealth of intervention strategies, which are formulated by medical experts based on a large number of clinical cases and research results and are scientific and effective. Different influencing factors correspond to different intervention strategies. For example, for the influencing factor of dietary composition, possible intervention strategies may include formulating a personalized diet plan, controlling the intake of carbohydrates, and reasonably matching the diet, etc.; for the influencing factor of exercise intensity, the intervention strategy may include formulating an exercise plan suitable for the patient's physical condition, determining the intensity, time, and frequency of exercise, etc. Integrate the intervention strategies corresponding to each eligible influencing factor to form the patient's blood glucose management strategy. The generated management strategy has the characteristics of being personalized, can target the specific situation of the patient, focus on the factors having a greater impact on blood glucose, and provide precise intervention measures. At the same time, the management strategy also needs to consider the individual differences of the patient, such as age, physical condition, living habits, etc., to ensure the feasibility and effectiveness of the strategy. Diabetes patient data management system based on big data The diabetes patient data management system based on big data applies the above management method, mainly including components such as a cloud data server, an integrated management terminal, a communication module, a user terminal, a risk assessment module, and a health status tracking module. Each module realizes communication connection through the communication module. The cloud data server is the data core of the system, which is used to provide a patient database and a policy database. A large amount of patient information is stored in the patient database, including basic information, medical history records, medication records, blood glucose monitoring data, influencing factor data, etc. These data come from various comprehensive management terminals and user terminals, and through data cleaning, integration, and storage, a rich patient data resource is formed. The policy database stores various intervention strategies and early warning strategies, which are important bases for the system to generate management strategies and conduct risk early warnings. The cloud data server adopts a distributed storage architecture to meet the storage requirements of massive data and ensure the high availability and scalability of the data. At the same time, it is equipped with a powerful data processing and analysis engine, which can support complex data analysis and model calculations. The comprehensive management terminal is mainly used to customize the influencing factor data collection module in the user terminal according to the influencing factors of the patient's health data, and manage the user management data and the user terminal. It provides a friendly user interface for the convenience of medical staff to operate. When customizing the collection module, the comprehensive management terminal configures the corresponding collection module parameters, such as sampling frequency, data format, etc., according to the patient influencing factors and priority list obtained in step one. At the same time, the comprehensive management terminal is also responsible for managing the user terminal, including terminal registration, configuration update, status monitoring, etc., to ensure the normal operation of the user terminal. In addition, the comprehensive management terminal can also conduct preliminary analysis and display of the patient's management data to provide decision-making support for medical staff. The user terminal is a device directly facing patients and is used to collect the management data of patients. It includes the following main modules: Influencing factor data collection module: It is used to collect data on various influencing factors that affect the patient's blood glucose index, and is equipped with corresponding sensors according to different influencing factors, such as blood glucose meter sensors, motion sensors, sleep monitoring sensors, etc. Identity information module: It is used to collect and store the identity information of patients to ensure the accurate correspondence of the data with the patients. Location information collection module: It uses positioning technologies such as GPS or Beidou to collect the location information of the patients, which is of great significance for analyzing the impact of environmental factors on blood glucose and for positioning and rescue in case of emergencies. Management data generation module: According to the data collected by the user terminal within the management cycle, it conducts data integration and processing to generate user health management data. Data processing module: It is the core processing unit of the user terminal, responsible for processing, analyzing, and storing the data collected by each module, and at the same time controlling the working status of each module. Communication device: It is used to realize the data transmission between the user terminal and the communication module, supports multiple communication methods, such as 4G, 5G, Wi-Fi, etc., to ensure the stable transmission of data. The risk assessment module is used to conduct risk assessments based on the management data of patients, including health risk assessment and data integrity assessment. The health risk assessment module evaluates the health status of patients based on the degree of blood glucose change and other relevant data in Step 3, in combination with the risk assessment model in the cloud data server, determines whether the patient is at risk of hypoglycemia, hyperglycemia or other complications, and gives the corresponding risk level. The data integrity assessment module checks the collected data and evaluates the integrity and quality of the data, such as checking whether the data collection frequency meets the requirements and whether there are missing values, etc., providing a reliable data basis for subsequent data analysis and management. The health status tracking module is used to track the health status of patients based on the patient management data and the risk assessment results. It analyzes the long-term blood glucose data, influencing factor data and risk assessment results of patients, draws the health status change curve of patients, and intuitively displays the blood glucose control situation and health trend of patients. At the same time, the health status tracking module can also identify the change patterns of patients' health status, timely detect potential health problems, and provide a basis for adjusting management strategies. In addition, the health status tracking module can also provide health reports for patients to help patients understand their own health status and improve patients' self-management awareness. The communication module is the bridge connecting the various components of the system and is responsible for realizing data transmission and communication between the cloud data server, the comprehensive management terminal, the user terminal, the risk assessment module and the health status tracking module. It supports a variety of communication protocols and technologies and selects appropriate communication methods according to the requirements of different modules and communication scenarios. For example, for the communication between the user terminal and the communication module, considering that the user terminal may be in a mobile state and has requirements for power consumption, wireless communication technologies such as 4G, 5G or Wi-Fi are adopted; for the communication between the cloud data server and other modules, since the data volume is large and high stability is required, wired network communication technologies are adopted. The communication module also needs to have security mechanisms such as data encryption, identity authentication and access control to ensure the security and privacy of data during transmission, preventing data leakage and illegal access. Example 1: A case of personalized management of a 45-year-old male patient with type 2 diabetes I. Basic situation of the patient Patient Mr. Zhang, 45 years old, has been diagnosed with type 2 diabetes for 3 years, with a BMI of 27.5 kg / m². His previous medical history shows hyperlipidemia. Currently, he is taking metformin (500 mg / time, 3 times a day) combined with acarbose (50 mg / time, 3 times a day). The comprehensive management terminal collects his fasting and postprandial blood glucose data for 72 consecutive hours through a standardized form. The fasting blood glucose baseline value is 7.8 - 8.5 mmol / L, and the 2-hour postprandial blood glucose baseline value is 11.2 - 13.0 mmol / L. II. Influence Factor Collection and Priority Determination Similar Case Matching: Through the K-means clustering algorithm, the cloud data server matches 100 patients with similar age, BMI, and medication regimens to Mr. Zhang and finds that dietary carbohydrate intake (accounting for 38%), sedentary duration (accounting for 27%), and nighttime sleep duration (accounting for 15%) are the main influencing factors with a correlation exceeding the threshold (0.7). Dynamic Weight Calculation: The initial weights are set as diet (0.4), exercise (0.3), and sleep (0.2). Considering Mr. Zhang's dietary records of excessive daily carbohydrate intake (average 280 g / day, far exceeding the recommended value of 200 g / day) and average daily sedentary duration (8.5 hours), after adjustment by the dynamic weighting algorithm, the priority list is: dietary components (0.45), exercise intensity (0.32), and sleep quality (0.23). III. Data Collection Scheme and Terminal Configuration Module Activation Sequence: The high-priority dietary component collection module (associated with the intelligent diet record APP) is set to the real-time sampling mode, and the carbohydrate content needs to be scanned and entered by scanning the food barcode after each meal; the exercise intensity collection module (integrated with an acceleration sensor) is set to high-frequency sampling at 10 minutes / time to monitor sedentary duration and the number of steps walked. Sampling Parameter Configuration: The sleep quality collection module (wrist intelligent bracelet) adopts a basic cycle of 30 minutes. When the heart rate variability (HRV) is detected to be lower than 50 ms after falling asleep, it triggers an encrypted sampling cycle of 5 minutes to record the deep sleep / light sleep duration. IV. Blood Glucose Change Analysis and Risk Warning Adaptive Sampling Trigger: At 14:00 on the 3rd day, the user terminal detects that the postprandial blood glucose change rate of Mr. Zhang reaches 0.8 mmol / L / 10 min (exceeding the threshold of 0.5 mmol / L / 10 min), and the encrypted sampling cycle is started, and three blood glucose values (12.5 → 13.2 → 14.1 mmol / L) are continuously collected within 5 minutes. Sliding Window Calculation: Taking the carbohydrate intake (75 g) at lunch (12:00 - 12:30) as the influencing factor, with an action period set to 2 hours and a window length of 4 hours (10:00 - 14:00), it is calculated that ΔG = Σ|G(t + τ) - G(t)| / (2h - 0h) = (|8.2 - 7.8| + |9.5 - 8.2| + … + |14.1 - 13.2|) / 2h = 1.2 mmol / L / h, triggering a medium-risk warning (sharp increase in blood glucose). The user terminal pushes an intervention prompt of "It is recommended to walk immediately for 15 minutes" through the APP. V. Weight Update and Generation of Management Strategies Time-varying weight calculation: On the 7th day, α = 0.6, Δt = 1 day, the ΔG_i(t) caused by daily dietary factors = 1.8 mmol / L, the total ΔG_j(t) of all factors = 2.5 mmol / L, and the historical weight W_i(t - Δt) = 0.45. Then the updated dietary weight W_i(t) = 0.6×(1.8 / 2.5)+(1 - 0.6)×0.45 = 0.432 + 0.18 = 0.612. Management strategy generation: The weight threshold is set to 0.1, and diet (0.612) and exercise (0.285) are the main intervention factors. Retrieve the corresponding intervention plan from the cloud strategy database: Dietary intervention: Develop a low-carbohydrate diet plan (carbohydrates ≤ 180 g per day), and recommend a food list with a GI value < 55; Exercise intervention: Set a sedentary reminder (stand and move for 3 minutes every hour), and accumulate 30 minutes of moderate-intensity exercise per day (such as brisk walking). Example 2: Dynamic management case of an elderly female patient with type 1 diabetes I. Patient's basic situation Grandma Li, 68 years old, has a 10-year history of type 1 diabetes and has hypertension. She uses an insulin pump (basal rate 0.5 U / h, bolus dose at mealtime calculated according to carbohydrate intake). The integrated management terminal collects her continuous 72-hour blood glucose data. The fasting blood glucose baseline value is 6.5 - 7.2 mmol / L, and the lowest night blood glucose value reached 3.8 mmol / L (at 2:00 am). II. Collection of influencing factors and determination of priorities Similar case matching: The cloud data server uses the hierarchical clustering algorithm to match 80 elderly patients with type 1 diabetes and finds that the insulin pump basal rate setting (correlation degree 0.85), nocturnal sleep apnea index (correlation degree 0.78), and drug compliance (insulin injection standardization degree, correlation degree 0.72) are the main influencing factors. Dynamic weight calculation: The initial weights are set as insulin (0.5), sleep (0.3), and compliance (0.2). Considering Grandma Li's nocturnal hypoglycemia record and mild sleep apnea (AHI = 12 times / hour) shown by sleep monitoring, the adjusted priority list is: insulin basal rate (0.55), sleep quality (0.3), and drug compliance (0.15). III. Data collection plan and terminal configuration Module activation sequence: The insulin basal rate acquisition module (directly reads the insulin pump setting parameters) is set to the real-time synchronization mode, and the basal rate and bolus data are updated every 15 minutes; The sleep quality acquisition module (integrates a blood oxygen saturation sensor) uses a basic cycle of 15 minutes. When the blood oxygen saturation < 90% lasts for 5 minutes, an encrypted sampling cycle of 1 minute is triggered. Outlier threshold setting: The outlier determination threshold for blood glucose data is set to < 3.9 mmol / L (hypoglycemia) and > 16.7 mmol / L (hyperglycemia), and the abnormal fluctuation threshold for the insulin pump basal rate is set to ±0.1 U / h (compared with the same period of the previous day). IV. Blood Glucose Change Analysis and Risk Warning Adaptive sampling trigger: At 1:30 am on the 5th day, the user terminal detected that Grandma Li's blood glucose dropped from 5.2 mmol / L to 4.1 mmol / L (change rate 0.7 mmol / L / 15 min), and at the same time, the blood oxygen saturation was 90% (lasting for 10 minutes). The encrypted sampling cycle was started, and blood glucose was collected once every 1 minute (4.1 → 3.9 → 3.8 mmol / L). Sliding window calculation: Taking the nocturnal insulin basal rate (0.5 U / h) as the influencing factor, the action cycle is set to 4 hours (23:00 - 3:00), and the window length is 8 hours (21:00 - 5:00). It is calculated that ΔG = Σ|G (t + τ) - G (t)| / (4h - 0h) = (|5.8 - 6.2| + |5.2 - 5.8| + … + |3.8 - 4.1|) / 4h = 0.6 mmol / L / h, triggering a high-risk warning (hypoglycemia combined with sleep apnea). The user terminal synchronously starts an audible and visual alarm and sends an emergency notice to the family member's mobile phone. V. Weight Update and Management Strategy Generation Time-varying weight calculation: On the 10th day, α = 0.4 (the elderly pay more attention to the stability of historical data), Δt = 12 hours, the ΔG_i (t) caused by the insulin factor on the same day = 1.0 mmol / L (hypoglycemia event), the total ΔG_j (t) = 1.5 mmol / L, and the historical weight W_i (t - Δt) = 0.55. Then the updated insulin weight W_i(t) = 0.4×(1.0 / 1.5) + (1 - 0.4)×0.55 ≈ 0.267 + 0.33 = 0.597. Management strategy generation: The weight threshold is set to 0.1, and insulin (0.597) and sleep (0.28) are the main intervention factors. The corresponding intervention plan is retrieved from the cloud strategy database: Insulin intervention: The basal rate from 0:00 to 3:00 in the early morning is reduced to 0.4 U / h, and the hypoglycemia warning threshold is set to 4.5 mmol / L (triggering intervention in advance); Sleep intervention: It is recommended to wear a home non-invasive ventilator, avoid large-dose insulin injection 2 hours before going to bed, and adjust the night blood glucose monitoring cycle to once every 30 minutes.

Claims

1. A method for managing diabetes patient data based on big data, characterized in that, It includes the following steps: Step 1: The comprehensive management terminal collects the basic information of the patient and the blood glucose data in the stable state. According to the basic information of the patient, the blood glucose influencing factors of the patient are obtained from the cloud data server. According to the blood glucose influencing factors of the patient, the information of the corresponding influencing factor data collection module is obtained, and an influencing factor data collection plan for the corresponding patient is generated; Step 2: According to the influencing factor data collection plan of the corresponding patient, a user terminal for the corresponding patient is generated. The comprehensive management terminal generates a patient data management container and communicates with the user terminal of the corresponding patient, and sends the blood glucose data of the patient in the stable state to the patient data management container; Step 3: The user terminal of the corresponding patient collects the patient's blood glucose data and influencing factor data according to the collection period, and obtains the degree of change in blood glucose data at the start and end times of the influencing factor based on the blood glucose data and the influencing factor data; performs risk warning according to the degree of change in blood glucose data, and sends the influencing factor data and the degree of change in blood glucose data at the start and end times of the influencing factor to the patient data management container; Step 4: The patient data management container obtains the weight of the influencing factor on the blood glucose data according to the influencing factor data obtained in different periods and the degree of change in blood glucose data at the start and end times of the influencing factor; Step 5: According to the obtained weights of each influencing factor on the blood glucose data, a patient blood glucose management strategy is generated to complete the data management of diabetic patients.

2. The method for managing diabetes patient data based on big data according to claim 1, wherein The comprehensive management terminal collects the basic information of the patient and the blood glucose data in the stable state. According to the basic information of the patient, the blood glucose influencing factors of the patient are obtained from the cloud data server, including: Collecting the patient's identity information, medical history record, medication record and the set fasting / postprandial blood glucose baseline value for continuous n hours through a standardized form; Invoking the patient database of the cloud data server for similar case matching, and obtaining a set of influencing factors with a correlation greater than the threshold for the current patient's blood glucose fluctuation. The influencing factors include: dietary components, exercise intensity, sleep quality, medication compliance, environmental temperature and humidity, and emotional fluctuation indicators; Establishing a list of priorities for the patient's individual influencing factors, and the priority is calculated by a dynamic weighting algorithm, and the initial weight is a preset value.

3. The method for managing diabetes patient data based on big data according to claim 1, wherein, The obtaining of the information of the corresponding influencing factor data collection module according to the blood glucose influencing factors of the patient and the generation of an influencing factor data collection plan for the corresponding patient include: Constructing a module configuration mapping table to associate each type of influencing factor with the corresponding collection module; generating a collection module activation sequence according to the priority list, and setting the high-priority collection module to the high-frequency sampling mode; and generating a collection module configuration file, including the sampling frequency, data format standard and outlier determination threshold.

4. The method for managing diabetes patient data based on big data according to claim 1, wherein The user terminal of the corresponding patient collects the patient's blood glucose data and influencing factor data according to the collection period, and obtains the degree of change in blood glucose data at the start and end times of the influencing factor based on the blood glucose data and the influencing factor data, including: Adopting an adaptive sampling period: the basic period is the first set duration. When it is detected that the blood glucose change rate exceeds the change rate threshold, encrypted sampling is performed, and the encrypted sampling period is the second set duration, and the second set duration is less than the first set duration; The sliding window algorithm is used to calculate the degree of change. The window length is set to twice the action period of the influencing factor, and the degree of change is as follows: ΔG = Σ|G(t+τ)-G(t)| / (τ_max - τ_min) where G(t) is the blood glucose value and τ is the action duration of the influencing factor.

5. The method for managing diabetes patient data based on big data according to claim 1, characterized in that The risk warning based on the degree of change in blood glucose data includes: According to different degrees of change in blood glucose data, the corresponding warning strategy is called in the cloud data server, and the patient is warned according to the warning strategy.

6. The method for managing diabetes patient data based on big data according to claim 1, wherein The patient data management container obtains the weight of the influencing factor on blood glucose data based on the influencing factor data obtained in different periods and the degree of change in blood glucose data at the start and end times of the influencing factor, including: The time-varying weight calculation model is used to calculate the weight of the influencing factor on blood glucose data, and the weight is updated as: W_i(t) = α·ΔG_i(t) / ΣΔG_j(t) + (1-α)·W_i(t-Δt) where α is the forgetting factor (0<α<1), Δt is the update period; W_i(t) is the dynamic influence weight of the influencing factor i on blood glucose data at time t, indicating the contribution ratio of the i-th influencing factor to blood glucose fluctuations at time t. The sum of the weights of all influencing factors ΣW_i(t)=1; ΔG_i(t) is the blood glucose change amount caused by the influencing factor i at time t; ΣΔG_j(t) is the total blood glucose change amount caused by all influencing factors at time t; W_i(t-Δt) is the historical weight of the influencing factor i in the previous calculation period (t-Δt).

7. The method for managing diabetes patient data based on big data according to claim 1, characterized in that, The generation of the patient blood glucose management strategy based on the obtained weights of each influencing factor on blood glucose data includes: Sort the weights of the influencing factors on blood glucose data to generate a list of intervention priorities for the influencing factors. For the influencing factors with weights greater than the weight threshold, obtain the corresponding intervention strategies in the cloud data server. The intervention strategies corresponding to each influencing factor with a weight greater than the weight threshold constitute the patient blood glucose management strategy.

8. A data management system for diabetic patients based on big data, characterized in that, Applying the big data-based diabetes patient data management method according to any one of claims 1-7, including a cloud data server, a comprehensive management terminal, a communication module, a user terminal, a risk assessment module, and a health status tracking module; The cloud data server, the comprehensive management terminal, the user terminal, the risk assessment module, and the health status tracking module are respectively communicatively connected to the communication module; The cloud data server is used to provide a patient database and a strategy database; The comprehensive management terminal is used to customize the influencing factor data collection module in the user terminal according to the influencing factors of the patient's health data, and manage the user management data and the user terminal; The user terminal is used to collect the management data of the patient; The risk assessment module is used to perform risk assessment according to the management data of the patient; the risk assessment includes a health risk assessment module and a data integrity assessment; The health status tracking module is used to track the health status of the patient according to the patient management data and the risk assessment.

9. The data management system for diabetic patients based on big data according to claim 8, wherein, The user terminal described above includes an influencing factor data acquisition module, an identity information module, a location information acquisition module, a management data generation module, a data processing module, and a communication device; The influencing factor data acquisition module, the identity information module, the location information acquisition module, the management data generation module, and the communication device are respectively connected to the data processing module; The influencing factor data acquisition module is used to acquire data on influencing factors that affect the patient's blood glucose index; The identity information acquisition module is used to acquire the patient's identity information; The location information acquisition module is used to acquire the location where the patient is located; The management data generation module is used to generate user health management data based on the data collected by the user terminal within the management cycle.

Images