Diabetes decision-making method, device and equipment fusing blood glucose monitoring and clinical data

Abstract

The invention discloses a diabetes decision-making method, device and equipment fusing blood glucose monitoring and clinical data, and relates to the field of intelligent data management.The method comprises the steps that blood glucose related data of a target object is collected in real time, and the data is uploaded to a central database and preprocessed; based on a signal processing technology and a machine learning algorithm, extracting features related to diabetes management from the preprocessed data so as to identify and obtain key clinical events; according to the key clinical event obtained through recognition, in combination with medical knowledge and clinical guidelines in a pre-created evidence-based medical knowledge base, clinical decision suggestions are generated and visually displayed. According to the application, deep combination of patient data and evidence-based medical knowledge can be realized, and a real-time, scientific and individualized clinical intervention scheme is provided.

Description

Technical Field

[0001] This application relates to the field of data intelligence management, specifically to a diabetes decision-making method, device, and equipment that integrates blood glucose monitoring and clinical data. Background Technology

[0002] Currently, the clinical management of diabetes mainly relies on the doctor's experience and judgment, as well as the patient's regular follow-up visits. In the traditional model, doctors obtain the patient's blood glucose data through a blood glucose meter or continuous glucose monitoring (CGM) device, and then manually analyze it in conjunction with the patient's diet, exercise, medication, and other factors to propose adjustment plans.

[0003] However, the above-mentioned model has problems such as low data processing efficiency, lack of intelligent clinical decision support, insufficient causal explanation, untimely response, and difficulty in doctor-patient collaboration. Specifically, it is manifested as follows: (1) Low data processing efficiency: Doctors and nurses need to process a large amount of raw blood glucose curves and recorded information within a limited time, making it difficult to discover potential risks in a timely manner; (2) Lack of intelligent support: Existing systems mostly focus on data display and trend analysis, failing to deeply integrate evidence-based medicine guidelines and dynamic patient data, and lacking automated and personalized suggestions; (3) Insufficient causal explanation: Abnormal blood glucose events are difficult to automatically associate with specific lifestyle factors, limiting doctors and nurses' comprehensive understanding of the patient's condition; (4) Untimely response: Clinical intervention relies on patient follow-up visits or emergency reports, making it difficult to deal with acute risks such as hypoglycemia or persistent hyperglycemia in a timely manner; (5) Difficulty in doctor-patient collaboration: Existing solutions are difficult to achieve effective connection between patient-side data and medical staff workflow, resulting in medical staff being unable to dynamically grasp the patient's condition.

[0004] It is evident that the aforementioned issues directly impact the scientific rigor and timeliness of clinical decision-making, increasing the risk of diabetic complications. Summary of the Invention

[0005] This application provides a diabetes decision-making method, device, and equipment that integrates blood glucose monitoring and clinical data, enabling a deep integration of patient data and evidence-based medicine knowledge, and providing real-time, scientific, and individualized clinical intervention plans.

[0006] In a first aspect, embodiments of this application provide a diabetes decision-making method that integrates blood glucose monitoring and clinical data, the diabetes decision-making method integrating blood glucose monitoring and clinical data including: Real-time collection of blood glucose-related data from target subjects, uploading of data to a central database and preprocessing of data; Based on signal processing technology and machine learning algorithms, features related to diabetes management are extracted from preprocessed data to identify key clinical events; Based on the identified key clinical events, combined with medical knowledge and clinical guidelines from a pre-created evidence-based medicine knowledge base, clinical decision recommendations are generated and visualized.

[0007] In conjunction with the first aspect, in one implementation, the real-time collection of blood glucose-related data of the target object, uploading the data to a central database, and performing data preprocessing specifically includes: CGM data of the target object is collected in real time by a CGM device worn on the target object, and multidimensional health information of the target object is collected through a smart terminal or a dedicated recording device. Data collected is uploaded to a central database via wireless or wired transmission for storage and preprocessing.

[0008] In conjunction with the first aspect, in one implementation, the uploading of data to a central database and the preprocessing of the data specifically include: The CGM data and clinical data of the target subjects collected by the CGM device are integrated and cleaned to achieve data preprocessing; For CGM data, the validity of the data is tested to check whether the blood glucose value is within a reasonable range and to remove outliers caused by equipment failure or signal interference. The timestamps of the data are also standardized. For clinical data, data format standardization is performed to convert data in different formats into a standard format; For CGM data and clinical data, this also includes imputing missing data using mean imputation, median imputation, or predictive imputation methods based on machine learning algorithms.

[0009] In conjunction with the first aspect, in one implementation, the extraction of diabetes management-related features from preprocessed data based on signal processing techniques and machine learning algorithms to identify key clinical events specifically includes: Based on CGM data, blood glucose statistics are calculated and the dynamic change characteristics of blood glucose are obtained through time series analysis. By combining the clinical data of the target subjects, life events associated with blood glucose fluctuations can be identified; Based on the dynamic characteristics of blood glucose changes and life events associated with blood glucose fluctuations, hypoglycemic events, hyperglycemic events, and events with rapid blood glucose fluctuations are detected and labeled.

[0010] In conjunction with the first aspect, in one implementation, the step of generating clinical decision recommendations based on identified key clinical events, combined with medical knowledge and clinical guidelines from a pre-created evidence-based medicine knowledge base, specifically includes: By combining the identified key clinical events, risk assessments are performed on the target subjects to determine the severity of diabetes and the risk of complications. Based on the risk assessment results and the individual characteristics of the target group, appropriate treatment strategies are matched from a pre-created evidence-based medicine knowledge base to generate clinical decision recommendations.

[0011] In conjunction with the first aspect, in one implementation method, the creation of an evidence-based medicine knowledge base specifically includes: A structured knowledge graph for diabetes diagnosis and treatment was constructed and stored and managed using the Neo4j graph database. The node types of the knowledge graph include diseases, drugs, interventions, guideline recommendations, and adverse reactions. The edge relationship definitions include drug treatment relationships, contraindication relationships, recommendation strength, evidence level, and side effect associations. The system utilizes natural language processing technology to automate knowledge extraction and updating, employs a BERT-based model for entity recognition and relation extraction from medical literature, and uses rule-based methods to process structured guide documents. After the extraction results are filtered based on confidence level, they are pushed to the expert review platform for active confirmation, modification or rejection. The knowledge metadata that passes the review is automatically integrated into the knowledge base.

[0012] In conjunction with the first aspect, in one implementation method, For visualization, Vue.js is used in combination with ECharts to achieve a cross-platform visualization interface, with data synchronization between the web and mobile ends; The visualization components include a comprehensive dashboard, multi-dimensional time-series curves, event timelines, and decision interpretation diagrams; In the integrated dashboard, the pie chart displays the blood glucose target achievement rate, the trend chart shows the blood glucose fluctuation over the most recent preset time period, and the indicator card presents the current blood glucose value and coefficient of variation. The multidimensional time-series curve supports the overlay display and linked scaling of blood glucose, exercise, and diet data, and adopts a dual Y-axis design to support the display and concealment control of data sequences. The event timeline is used to mark hypoglycemia, hyperglycemia events, and life events along the timeline. The decision interpretation diagram uses a tree structure to visualize the decision path, showing the key nodes in the reasoning process of generating clinical decision recommendations, the matching guideline clauses, and the confidence scores.

[0013] In conjunction with the first aspect, in one implementation method, The diabetes decision-making method that integrates blood glucose monitoring and clinical data also includes building a real-time stream processing pipeline based on Apache Flink to process blood glucose data points in blood glucose-related data and achieve safety warnings. The security early warning system includes a multi-level architecture for building monitoring strategies, early warning classification, and response.

[0014] Secondly, embodiments of this application provide a diabetes decision-making device that integrates blood glucose monitoring and clinical data, the diabetes decision-making device comprising: The data acquisition module is used to collect blood glucose-related data of the target subjects and upload it to the central database; The central database is used to store the data uploaded by the data acquisition module. The data preprocessing module is used to preprocess blood glucose-related data; The feature extraction and event recognition module is used to extract features related to diabetes management from preprocessed data based on signal processing technology and machine learning algorithms in order to identify key clinical events. The clinical decision processing module is used to generate clinical decision recommendations based on the identified key clinical events, combined with medical knowledge and clinical guidelines from a pre-created evidence-based medicine knowledge base. Evidence-based medicine knowledge base, which is used to store clinical guidelines and standards; An interactive display module is used to display the target subject's blood glucose information and clinical decision suggestions; The safety warning module is used to monitor the blood glucose information of the target object in real time to provide safety warnings.

[0015] Thirdly, embodiments of this application provide a diabetes decision-making device that integrates blood glucose monitoring and clinical data. The diabetes decision-making device that integrates blood glucose monitoring and clinical data includes a processor, a memory, and a diabetes decision-making program that integrates blood glucose monitoring and clinical data and is stored in the memory and can be executed by the processor. When the diabetes decision-making program that integrates blood glucose monitoring and clinical data is executed by the processor, it implements the steps of the diabetes decision-making method that integrates blood glucose monitoring and clinical data described above.

[0016] The beneficial effects of the technical solutions provided in this application include: (1) Improve data processing efficiency and accuracy: Automated data collection and preprocessing significantly reduce manual processing time and error rate, ensuring data accuracy and reliability; (2) Achieve intelligent clinical decision support: Deeply integrate evidence-based medicine knowledge, provide scientific and traceable individualized intervention suggestions, and improve the level of intelligence in clinical decision-making; (3) Enhance causal explanation and risk warning capabilities: Through multi-factor association analysis, enhance the causal explanation capability for abnormal blood glucose events and automatically trigger high-priority alarms to ensure patient safety; (4) Promote doctor-patient collaboration and remote management: realize data linkage between the patient end and the medical care end, support remote dynamic management, and improve the continuity and effectiveness of diabetes management; (5) Improve the efficiency and scientific nature of clinical decision-making: Provide each patient with a scientific, standardized and causally explanatory individualized intervention plan, which significantly improves the efficiency and scientific nature of clinical decision-making; (6) Support knowledge base iteration and continuous optimization: continuously optimize and update the evidence-based medicine knowledge base and decision-making rules through data feedback mechanism to ensure the advancement and accuracy of diabetes decision-making. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the diabetes decision-making method that integrates blood glucose monitoring and clinical data, as described in this application. Figure 2 This is a schematic diagram of the functional modules of the diabetes decision-making device that integrates blood glucose monitoring and clinical data according to this application. Figure 3 This is a schematic diagram of the hardware structure of the diabetes decision-making device that integrates blood glucose monitoring and clinical data, as described in this application. Detailed Implementation

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

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0020] In the first aspect, the embodiments of this application provide a diabetes decision-making tool that integrates blood glucose monitoring and clinical data, overcoming the problems of heavy manual analysis burden, decision-making reliance on personal experience, lack of intelligent support, and untimely response in existing diabetes clinical management. It provides a diabetes clinical decision support method that integrates dynamic blood glucose monitoring and clinical data, realizes the deep integration of patient data and evidence-based medicine knowledge, and provides real-time, scientific and individualized clinical intervention plans.

[0021] In one embodiment, reference is made to Figure 1 , Figure 1 This is a flowchart illustrating the diabetes decision-making method that integrates blood glucose monitoring and clinical data, as described in this application. Figure 1 As shown, diabetes decision-making methods that integrate blood glucose monitoring and clinical data include: S1: Real-time collection of blood glucose-related data of the target object, uploading of data to the central database and preprocessing of data; S2: Based on signal processing technology and machine learning algorithms, extract features related to diabetes management from preprocessed data to identify key clinical events; S3: Based on the identified key clinical events, combined with medical knowledge and clinical guidelines from a pre-created evidence-based medicine knowledge base, generate clinical decision recommendations and display them visually.

[0022] Furthermore, in one embodiment, blood glucose-related data of the target object is collected in real time, uploaded to a central database, and preprocessed, specifically including: S101: The target subject's CGM data is collected in real time through a CGM (continuous glucose monitoring) device worn on the target subject, and multidimensional health information of the target subject is collected through a smart terminal or a dedicated recording device; it should be noted that in this application, the target subject refers to a diabetic patient; S102: Uploads the collected data to the central database via wireless or wired transmission for data storage and preprocessing.

[0023] Specifically, the process begins with data collection. Patients wear continuous glucose monitoring devices (CGM devices) to obtain real-time blood glucose curves and record multidimensional health information such as diet, exercise, sleep, psychological state, and medication adherence via smart terminals (such as smartphones and tablets) or dedicated recording devices. The data is then uploaded to a central database via wired or wireless transmission methods (such as Bluetooth or Wi-Fi) for storage and preprocessing. The central database, acting as a data hub, stores the raw data uploaded by patients, providing a solid data foundation for subsequent processing. The central database is designed with a highly available and scalable architecture to ensure data security and reliability.

[0024] Furthermore, in one embodiment, data is uploaded to a central database and preprocessed, wherein the data preprocessing specifically includes: The process integrates and cleans CGM data and clinical data collected from the target subjects by CGM devices to achieve data preprocessing. Specifically, for CGM data, validity testing is performed to ensure blood glucose levels are within a reasonable range and to remove outliers caused by device malfunction or signal interference. Data timestamps are also standardized. For clinical data, data format standardization is implemented to convert data from different formats to a standard format. Furthermore, for both CGM and clinical data, missing data is imputed using mean imputation, median imputation, or predictive imputation methods based on machine learning algorithms.

[0025] Specifically, data preprocessing is performed, integrating and cleaning data from multiple sources, including continuous glucose monitoring devices and hospital clinical information systems. For continuous glucose monitoring data, validity is first verified to ensure blood glucose levels are within a reasonable physiological range, eliminating outliers caused by device malfunctions or signal interference. Simultaneously, the timestamps are standardized to ensure the accuracy of the time series data. For clinical data, such as patient basic information, medical history, and laboratory test results, data formats are standardized, converting data from different formats into a recognizable standard format. Furthermore, missing data is addressed. Depending on the characteristics and importance of the data, methods such as mean imputation, median imputation, or predictive imputation based on machine learning algorithms are employed to ensure data integrity and usability, providing a high-quality data foundation for subsequent feature extraction and event recognition.

[0026] Furthermore, in one embodiment, based on signal processing techniques and machine learning algorithms, features related to diabetes management are extracted from the preprocessed data to identify key clinical events, specifically including: S201: Based on CGM data, calculate blood glucose statistics and obtain the dynamic change characteristics of blood glucose through time series analysis; S202: By combining the clinical data of the target subjects, life events associated with blood glucose fluctuations are identified; S203: Based on the dynamic characteristics of blood glucose changes and life events associated with blood glucose fluctuations, hypoglycemic events, hyperglycemic events, and events with rapid blood glucose fluctuations are detected and labeled.

[0027] Specifically, feature extraction and event recognition are performed. These are the core of data analysis, utilizing advanced signal processing techniques and machine learning algorithms to extract features of significant importance for diabetes management from preprocessed data and identify key clinical events. For CGM data, not only are basic blood glucose statistics (such as mean, standard deviation, and coefficient of variation) calculated, but time-series analysis methods are also used to capture dynamic blood glucose change characteristics, such as the rate of rise / fall in blood glucose and the time of postprandial blood glucose peak. Simultaneously, combined with patient clinical data, such as dietary records, exercise status, and medication use, life events associated with blood glucose fluctuations are identified, such as meals, exercise, and insulin injections. In terms of event recognition, based on preset medical rules and algorithms, abnormal situations such as hypoglycemic events, hyperglycemic events, and events with rapid blood glucose fluctuations are automatically detected and marked, providing timely and accurate alerts for clinical decision-making.

[0028] Furthermore, in one embodiment, based on the identified key clinical events, and in conjunction with medical knowledge and clinical guidelines in a pre-created evidence-based medicine knowledge base, clinical decision recommendations are generated, specifically including: S301: Combine the identified key clinical events to conduct a risk assessment of the target subjects, and obtain the severity of diabetes and risk of complications in the target subjects; S302: Based on the risk assessment results and the individual characteristics of the target subjects, match the corresponding treatment strategies from a pre-created evidence-based medicine knowledge base to generate clinical decision recommendations.

[0029] Specifically, clinical decision processing involves using information obtained from feature extraction and event identification, combined with medical knowledge and clinical guidelines from an evidence-based medicine knowledge base, to provide personalized clinical decision recommendations to healthcare professionals. First, a risk assessment is conducted on the patient, such as evaluating the severity of diabetes and the risk of complications based on blood glucose fluctuations and complication status. Then, based on the risk assessment results and the patient's individual characteristics, appropriate treatment strategies are matched from the evidence-based medicine knowledge base. For example, for patients with poor blood glucose control and a risk of cardiovascular complications, adjustments to the hypoglycemic medication regimen and enhanced cardiovascular protection therapy are recommended. The patient's treatment history and current treatment status are also considered to avoid duplicate or inappropriate decision recommendations. When providing decision recommendations, they are presented in a clear and easy-to-understand manner, including detailed explanations of the treatment plan, expected effects, and possible side effects, assisting healthcare professionals in making scientific and rational clinical decisions and improving the treatment outcomes and management of diabetic patients.

[0030] Furthermore, in one embodiment, the creation of the evidence-based medicine knowledge base specifically includes: a1: Construct a structured diabetes diagnosis and treatment knowledge graph, which is stored and managed using the Neo4j graph database. The node types of the knowledge graph include diseases (type 2 diabetes), drugs (such as metformin and insulin), interventions (such as exercise therapy and nutritional intervention), guideline recommendations (such as ADA standards and CDS guidelines), and adverse reactions. The edge relationship definitions include drug treatment relationships (such as "used for"), contraindication relationships ("contraindicated for"), recommendation strength (such as "strong recommendation"), evidence level (such as "level A evidence"), and side effect associations. a2: Automated knowledge extraction and updating are achieved through natural language processing technology. A BERT-based model (a pre-trained language model based on the Transformer architecture) is used for entity recognition and relation extraction from medical literature, and a rule-based approach is used to process structured guideline documents. Specifically, a BERT-based model is used to perform entity recognition and relation extraction on the latest literature from sources such as PubMed and Cochrane Library, and a rule-based approach is used to process structured guideline documents. a3: After the extraction results are filtered based on confidence level, they are pushed to the expert review platform for active confirmation, modification, or rejection. Approved knowledge metadata is automatically integrated into the knowledge base. Medical experts can confirm, modify, or reject the results via a web interface; approved knowledge metadata is automatically integrated into the knowledge base.

[0031] Furthermore, in one embodiment, for visualization display, Vue.js (a progressive JavaScript framework) combined with ECharts (a JavaScript data visualization library) is used to implement a cross-platform visualization interface, with data synchronization between the web (web page) and mobile devices; the visualization display components include a comprehensive dashboard, multi-dimensional time series curves, event timelines, and decision interpretation charts; In the integrated dashboard, the pie chart displays the blood glucose target achievement rate (TIR), the trend chart shows the blood glucose fluctuation over the most recent preset time period (e.g., the last 14 days), and the indicator card presents the current blood glucose value and coefficient of variation; The multi-dimensional time-series curve supports the overlay display and linked scaling of blood glucose, exercise, and diet data. It adopts a dual Y-axis design and supports the display and concealment control of data sequences. The event timeline is used to mark hypoglycemia, hyperglycemia, and life events (such as eating and exercising) along the timeline. Clicking on an event marker will display detailed contextual data. The decision interpretation diagram uses a tree structure to visualize the decision path, showing the key nodes in the reasoning process of generating clinical decision recommendations, the matching guideline clauses, and the confidence scores.

[0032] Furthermore, in terms of visualization and interaction, medical staff can customize personalized threshold parameters (such as setting individualized blood glucose target ranges), provide feedback on suggestions (acceptance, recording reasons for modification, and selection of reasons for rejection), and add clinical annotations. All interactive operations are recorded in the audit log, including operation time, user ID, operation type, and modification content. This data is subsequently used to optimize the feedback learning of the decision-making model.

[0033] Furthermore, in one embodiment, the diabetes decision-making method that integrates blood glucose monitoring and clinical data also includes building a real-time stream processing pipeline based on Apache Flink (an open-source distributed big data processing engine) to process blood glucose data points in blood glucose-related data and achieve safety warnings; the safety warnings include building monitoring strategies, warning classification and response using a multi-level architecture.

[0034] Specifically, for data security alerts, a real-time stream processing pipeline built on Apache Flink processes thousands of blood glucose data points per second to achieve security alerts. The monitoring strategy adopts a multi-level architecture: Level 1 monitoring: static threshold rules, such as immediately marking blood glucose values ​​<3.9 mmol / L; Level 2 monitoring: dynamic adaptive thresholds, adjusting threshold sensitivity based on the patient's recent blood glucose levels (moving percentiles) and patterns (such as rate of decline >1 mg / dL / min); Context awareness: risk correction is performed by combining current activity status (such as sleep monitoring data from the device's gyroscope) and historical patterns (such as a history of nocturnal hypoglycemia).

[0035] The warning classification and response include three levels: Yellow warning (intermediate): Blood glucose levels exceed the range for 15 minutes, triggering notifications in relevant terminals and desktop reminders on medical staff's devices; Red warning (high risk): If blood glucose is <3.0mmol / L or persistently high blood glucose is >13.9mmol / L for 1 hour, triggering multi-channel alarms (SMS with emergency guidance, hospital intranet pop-up reminders, and push notifications on family member apps); Emergency protocol linkage: Pushing personalized treatment plans, such as "Immediately ingest 15g of fast-acting carbohydrates and retest blood glucose after 15 minutes" for hypoglycemia, and simultaneously notifying the responsible doctor and nurse. The effectiveness of the warning is continuously evaluated, and the warning response time, false positive rate, and clinical intervention effect are statistically analyzed. Threshold parameters and inference rules are regularly optimized.

[0036] The following section provides a detailed explanation of the diabetes decision-making method that integrates blood glucose monitoring and clinical data as described in this application.

[0037] (1) Data collection and transmission: After the patient wears the CGM device, the device collects blood glucose data at regular intervals and transmits it to the smart terminal wirelessly. At the same time, the patient records information such as diet, exercise, sleep, psychological state and medication through the application and uploads it to the central database simultaneously. (2) Data preprocessing: The central database cleans and formats the received raw data, including time alignment, missing value imputation, outlier removal and noise filtering. By optimizing algorithms and technical means, the quality and availability of the data are improved, providing a reliable data foundation for subsequent analysis. (3) Feature extraction and event recognition: Advanced machine learning algorithms are used to automatically extract blood glucose features and combine them with patients’ lifestyle data to identify potential abnormal blood glucose events and their causes. This allows for in-depth analysis of the complex relationships behind the data and provides precise support for clinical decision-making. (4) Clinical decision generation: The patient's individualized data is matched with the evidence-based medicine knowledge base to generate specific intervention suggestions and display them to medical staff through an interactive interface. Medical staff can make more reasonable and effective decisions based on the suggestions provided and their own experience. (5) Safety warning and emergency response: Monitor patient data in real time, trigger an alarm and notify medical staff immediately if serious abnormalities are detected. At the same time, provide emergency response guidelines and process support to ensure that medical staff can respond to acute risk events quickly and effectively.

[0038] This application achieves the synergistic use of multidimensional patient data and evidence-based medicine knowledge through a closed-loop mechanism of "data collection → data processing and event identification → knowledge base matching → personalized intervention generation → clinical implementation → data feedback and update". This improves the real-time nature of clinical decision-making, the scientific nature of intervention, and the accuracy of diabetes management, fully demonstrating the technical characteristics of personalization, intelligence, and safety traceability.

[0039] The diabetes decision-making method integrating blood glucose monitoring and clinical data in this application embodiment (1) improves data processing efficiency and accuracy: automated data collection and preprocessing significantly reduce manual processing time and error rate, ensuring the accuracy and reliability of data; (2) realizes intelligent clinical decision support: deeply integrates evidence-based medicine knowledge, provides scientific and traceable individualized intervention suggestions, and improves the level of intelligence in clinical decision-making; (3) enhances causal explanation and risk warning capabilities: through multi-factor association analysis, enhances the causal explanation capability for abnormal blood glucose events, and automatically triggers high-priority alarms to ensure patient safety; (4) promotes doctor-patient collaboration and remote management: realizes data linkage between the patient end and the medical care end, supports remote dynamic management, and improves the continuity and effectiveness of diabetes management; (5) improves the efficiency and scientific nature of clinical decision-making: provides each patient with a scientific, standardized and causally explanatory individualized intervention plan, significantly improving the efficiency and scientific nature of clinical decision-making; (6) supports knowledge base iteration and continuous optimization: continuously optimizes and updates the evidence-based medicine knowledge base and decision rules through a data feedback mechanism to ensure the advancement and accuracy of diabetes decision-making.

[0040] Secondly, embodiments of this application also provide a diabetes decision-making device that integrates blood glucose monitoring and clinical data.

[0041] In one embodiment, reference is made to Figure 2 , Figure 2 This is a schematic diagram of the functional modules of the diabetes decision-making device integrating blood glucose monitoring and clinical data, as described in this application. Figure 2 As shown, the diabetes decision-making device integrating blood glucose monitoring and clinical data includes: a data acquisition module, a central database, a data preprocessing module, a feature extraction and event recognition module, a clinical decision processing module, an evidence-based medicine knowledge base, an interactive display module, and a safety warning module.

[0042] The data acquisition module collects blood glucose-related data from the target subjects and uploads it to the central database. This module utilizes CMG equipment and a patient-side app for data collection. The central database stores the data uploaded by the acquisition module. The data preprocessing module preprocesses the blood glucose-related data. The feature extraction and event recognition module extracts diabetes management-related features from the preprocessed data using signal processing techniques and machine learning algorithms to identify key clinical events. The clinical decision processing module generates clinical decision recommendations based on the identified key clinical events, combined with medical knowledge and clinical guidelines from a pre-created evidence-based medicine knowledge base. This knowledge base stores treatment guidelines and standards. The interactive display module displays the target subjects' blood glucose information and clinical decision recommendations. The safety alert module monitors the target subjects' blood glucose information in real time to provide safety alerts.

[0043] It should be noted that the above modules are interconnected through data interfaces, forming a closed-loop clinical decision support system. The data preprocessing module is the foundation of the entire diabetes decision-making device, responsible for integrating and cleaning data from multiple sources, including continuous glucose monitoring devices and hospital clinical information systems. The feature extraction and event recognition module is the core of data analysis; it utilizes advanced signal processing techniques and machine learning algorithms to extract features of significant importance for diabetes management from the preprocessed data and identify key clinical events. The clinical decision processing module is the core decision-making unit of the diabetes decision-making device. Based on the information output by the feature extraction and event recognition module, combined with medical knowledge and clinical guidelines from the evidence-based medicine knowledge base, it provides personalized clinical decision recommendations to healthcare professionals.

[0044] Thirdly, embodiments of this application provide a diabetes decision-making device that integrates blood glucose monitoring and clinical data. The diabetes decision-making device that integrates blood glucose monitoring and clinical data can be a personal computer (PC), laptop computer, server, or other device with data processing capabilities.

[0045] Reference Figure 3 , Figure 3 This is a schematic diagram of the hardware structure of a diabetes decision-making device that integrates blood glucose monitoring and clinical data, as described in an embodiment of this application. In this embodiment, the diabetes decision-making device integrating blood glucose monitoring and clinical data may include a processor, a memory, a communication interface, and a communication bus.

[0046] The communication bus can be of any type and is used to interconnect the processor, memory, and communication interface.

[0047] The communication interface includes input / output (I / O) interfaces, physical interfaces, and logical interfaces used for interconnecting internal components of the diabetes decision-making device that integrates blood glucose monitoring and clinical data, as well as interfaces used for interconnecting the device with other devices (such as other computing devices or user devices). Physical interfaces can be Ethernet interfaces, fiber optic interfaces, ATM interfaces, etc.; user devices can be displays, keyboards, etc.

[0048] Memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0049] The processor can be a general-purpose processor, which can call a diabetes decision-making program that integrates blood glucose monitoring and clinical data stored in memory and execute the diabetes decision-making method that integrates blood glucose monitoring and clinical data provided in the embodiments of this application. For example, the general-purpose processor can be a central processing unit (CPU). The method executed when the diabetes decision-making program that integrates blood glucose monitoring and clinical data is called can refer to the various embodiments of the diabetes decision-making method that integrates blood glucose monitoring and clinical data in this application, and will not be repeated here.

[0050] Those skilled in the art will understand that Figure 3 The hardware structure shown does not constitute a limitation of this application and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0051] The terms "comprising" and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus. The terms "first," "second," and "third," etc., are used to distinguish different objects, etc., and do not indicate a sequence, nor do they limit "first," "second," and "third" to different types.

[0052] In the description of the embodiments of this application, terms such as "exemplary," "for example," or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "exemplary," "for example," or "for instance" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary," "for example," or "for instance" is intended to present the relevant concepts in a concrete manner.

[0053] In the description of the embodiments of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. The "and / or" in the text is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, in the description of the embodiments of this application, "multiple" means two or more.

[0054] In some processes described in the embodiments of this application, multiple operations or steps are included in a specific order. However, it should be understood that these operations or steps may not be executed in the order they appear in the embodiments of this application, or they may be executed in parallel. The sequence number of the operation is only used to distinguish different operations, and the sequence number itself does not represent any execution order. In addition, these processes may include more or fewer operations, and these operations or steps may be executed sequentially or in parallel, and these operations or steps may be combined.

[0055] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device to execute the methods described in the various embodiments of this application.

[0056] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method of diabetes decision making that fuses blood glucose monitoring with clinical data, characterized in that, The fusion blood glucose monitoring and clinical data diabetes decision-making method comprises the following steps: Real-time collection of blood glucose related data of a target object, uploading of the data to a central database and preprocessing of the data; Based on signal processing technology and machine learning algorithm, features related to diabetes management are extracted from the preprocessed data to identify key clinical events; According to the identified key clinical events, combined with the medical knowledge and clinical guidelines in the pre-created evidence-based medical knowledge base, clinical decision-making suggestions are generated and visualized.

2. The diabetes decision-making method integrating blood glucose monitoring and clinical data as described in claim 1, characterized in that, The real-time collection of blood glucose related data of a target object, uploading of the data to a central database and preprocessing of the data, specifically includes: Real-time collection of CGM data of a target object through a CGM device worn on the target object, collection of multi-dimensional health information of the target object through a smart terminal or a special recording device; Through wireless transmission or wired transmission, the collected data is uploaded to the central database for storage and preprocessing.

3. The diabetes decision-making method integrating blood glucose monitoring and clinical data as described in claim 1, characterized in that, The uploading of the data to the central database and the preprocessing of the data, wherein the preprocessing of the data specifically includes: Integrating and cleaning the CGM data of the target object collected by the CGM device and the clinical data of the target object to realize the preprocessing of the data; Among them, for CGM data, the validity of the data is detected to detect whether the blood glucose value is within a reasonable range and to eliminate abnormal values caused by device failure or signal interference, and the timestamp of the data is standardized; Among them, for clinical data, the data format is unified to convert different formats of data into a standard format; Among them, for CGM data and clinical data, it also includes using mean filling, median filling or machine learning algorithm based prediction filling method to fill the missing data.

4. The method of claim 3, wherein the method further comprises: The extraction of features related to diabetes management from the preprocessed data based on signal processing technology and machine learning algorithm to identify key clinical events, specifically includes: Based on CGM data, blood glucose statistics are calculated and dynamic change characteristics of blood glucose are obtained through time series analysis method; Combined with the clinical data of the target object, life events associated with blood glucose fluctuations are identified; According to the dynamic change characteristics of blood glucose and the life events associated with blood glucose fluctuations, hypoglycemia events, hyperglycemia events and rapid blood glucose fluctuation events are detected and marked.

5. The diabetes decision-making method integrating blood glucose monitoring and clinical data as described in claim 1, characterized in that, The generation of clinical decision-making suggestions according to the identified key clinical events, combined with the medical knowledge and clinical guidelines in the pre-created evidence-based medical knowledge base, specifically includes: Combined with the identified key clinical events, risk assessment is performed on the target object to obtain the severity of diabetes and complication risk of the target object; According to the risk assessment results and individual characteristics of the target object, the corresponding treatment strategy is matched from the pre-created evidence-based medical knowledge base to realize the generation of clinical decision-making suggestions.

6. The diabetes decision-making method integrating blood glucose monitoring and clinical data as described in claim 1, characterized in that, The creation of the evidence-based medical knowledge base specifically includes: A structured diabetes diagnosis and treatment knowledge graph is constructed and stored and managed by using a Neo4j graph database, wherein the node types of the knowledge graph include diseases, drugs, interventions, guideline recommendations, and adverse reactions, and the edge relationship definitions include drug treatment relationships, contraindications, recommendation strengths, evidence levels, and side effect associations; Automatic knowledge extraction and updating are realized by using natural language processing technology, BERT-based model is used for entity recognition and relationship extraction of medical literature, and a rule-based method is used to process structured guideline documents; After the extraction results are filtered based on confidence sorting, they are pushed to an expert review platform for active confirmation, modification or rejection operation, and the knowledge metadata that passes the review is automatically integrated into the knowledge base.

7. The diabetes decision-making method of claim 1, wherein: for visual display, a cross-platform visual interface is realized by using Vue.js combined with ECharts, and data of the web end and the mobile end are synchronized; the visual display components include a comprehensive dashboard, a multi-dimensional time series curve, an event timeline, and a decision explanation diagram; in the comprehensive dashboard, a ring chart displays a blood glucose compliance rate, a trend chart displays blood glucose fluctuations in a preset time period, and an index card presents a current blood glucose value and a coefficient of variation; the multi-dimensional time series curve supports superimposed display and linked scaling of blood glucose, exercise, and diet data, adopts a double-Y-axis design, and supports display and hiding control of data sequences; the event timeline is used to mark hypoglycemia, hyperglycemia events, and life events along a time axis; and the decision explanation diagram adopts a tree structure to visualize a decision path, displays key nodes in a reasoning process of generating a clinical decision suggestion, matches guideline clauses, and scores confidence.

8. The diabetes decision-making method of claim 1, wherein: the diabetes decision-making method further includes constructing a real-time stream processing pipeline based on Apache Flink to process blood glucose data points in blood glucose related data, and realizing a safety warning; the safety warning includes constructing a monitoring strategy by using a multi-level architecture, warning classification, and response. The diabetes decision-making device includes: a data acquisition module configured to acquire blood glucose related data of a target object and upload the data to a central database; the central database configured to store the data uploaded by the data acquisition module; a data preprocessing module configured to preprocess the blood glucose related data; a feature extraction and event recognition module configured to extract features related to diabetes management from the preprocessed data based on signal processing technology and machine learning algorithms to recognize key clinical events; a clinical decision processing module configured to generate a clinical decision suggestion based on the recognized key clinical events and medical knowledge and clinical guidelines in a pre-created evidence-based medical knowledge base; an evidence-based medical knowledge base configured to store diagnosis and treatment guidelines and specifications; and an interactive display module configured to display blood glucose information and clinical decision suggestions of the target object. ​ ​ ​ ​ ​ ​ ​ 9. A diabetes decision device that fuses blood glucose monitoring with clinical data, characterized by, ​ ​ ​ ​ ​ ​ ​ ​ A safety warning module is configured to monitor blood glucose information of the target object in real time to perform safety warning.

10. A diabetes decision device that fuses blood glucose monitoring with clinical data, characterized by, The diabetes decision device fusing blood glucose monitoring and clinical data comprises a processor, a memory, and a diabetes decision program fusing blood glucose monitoring and clinical data stored on the memory and executable by the processor, wherein the diabetes decision program fusing blood glucose monitoring and clinical data, when executed by the processor, implements the steps of the diabetes decision method fusing blood glucose monitoring and clinical data according to any one of claims 1 to 8.

Images