Blood pressure dynamic monitoring and risk assessment system based on artificial intelligence

Abstract

The invention relates to the technical field of blood pressure dynamic monitoring systems, and provides an artificial intelligence-based blood pressure dynamic monitoring and risk assessment system, which comprises a patient terminal, a doctor terminal, a data center terminal, a local anomaly detection terminal and a strategy generation terminal, the patient terminal is used for collecting blood pressure, body position change and endocrine hormone fluctuation data of a patient in real time; the data center terminal is used for fusing, storing and processing the data uploaded by the patient terminal and performing overall risk assessment; the local anomaly detection terminal is used for collecting local blood flow rate and pulse wave conduction velocity data and carrying out local anomaly detection index analysis; the strategy generation terminal is used for generating strategy information according to the overall risk assessment result and the local anomaly detection index; and the doctor terminal is used for receiving and displaying the overall risk assessment result and the local anomaly detection index, and receiving and executing the strategy information. According to the invention, the comprehensiveness and timeliness of hypertension risk prediction and intervention are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic blood pressure monitoring, and in particular to an artificial intelligence-based dynamic blood pressure monitoring and risk assessment system. Background Art

[0002] In modern medicine, the use of artificial intelligence technology for dynamic blood pressure monitoring and risk assessment has become a research hotspot. Traditional blood pressure monitoring methods, such as Korotkoff sound auscultation and oscillometric methods, have problems such as complex operation, patient discomfort, and difficulty in achieving continuous monitoring. To overcome these limitations, researchers have begun to explore non-invasive blood pressure detection methods based on photoplethysmography signals to achieve continuous, non-invasive, and accurate monitoring of blood pressure, thereby improving early warning and management capabilities for cardiovascular diseases such as hypertension. With the development of digital medicine, blood pressure management systems have gradually transitioned from traditional measurement equipment to integrated health management platforms based on artificial intelligence and big data technologies. In recent years, researchers have conducted in-depth explorations into individualized prediction, data integration, and real-time warning of chronic hypertension, and have proposed a software system design that integrates blood pressure monitoring, follow-up, and re-examination.

[0003] The intelligent model integrates home blood pressure, clinic blood pressure, and ambulatory blood pressure data to generate a dynamic blood pressure trajectory for each patient. It also combines the ClinRank heterogeneous star network, knowledge graph algorithms, and machine learning methods to establish a data-driven risk warning model. Furthermore, based on the accumulated experience of hundreds of thousands of hypertension home monitoring data, multiple intelligent auxiliary diagnosis submodules and data interface specifications have been developed, providing a solid engineering foundation for the construction of this system.

[0004] Many blood pressure monitoring systems have been developed. After extensive research and reference, we found that existing blood pressure monitoring systems include those disclosed in publication numbers CN115770027A, CN103976721B, and CN104173036A. These blood pressure monitoring systems generally include: a blood pressure data acquisition terminal, a blood pressure data analysis terminal, and a monitoring result output terminal; the blood pressure data acquisition terminal is used to acquire the patient's blood pressure data in real time; the blood pressure data analysis terminal is used to process and analyze the blood pressure data; and the monitoring result output terminal is used to output corresponding monitoring results based on the analysis results. Due to the relatively simple data processing and analysis process of the above-mentioned blood pressure monitoring process, the accuracy of the blood pressure monitoring results is reduced, resulting in a defect that reduces the comprehensiveness and timeliness of hypertension risk prediction and intervention. Summary of the Invention

[0005] The purpose of the present invention is to address the shortcomings of the above-mentioned blood pressure monitoring system and propose a blood pressure dynamic monitoring and risk assessment system based on artificial intelligence.

[0006] The present invention adopts the following technical solutions:

[0007] A blood pressure dynamic monitoring and risk assessment system based on artificial intelligence includes a patient terminal, a doctor terminal, a data center terminal, a local anomaly detection terminal and a strategy generation terminal; the patient terminal is used to collect the patient's blood pressure, body position change and endocrine hormone fluctuation data in real time; the data center terminal is used to integrate, store and process the data uploaded by the patient terminal, and perform overall risk assessment; the local anomaly detection terminal is used to collect local blood flow rate and pulse wave conduction velocity data, and perform local anomaly detection index analysis; the strategy generation terminal is used to generate strategy information based on the overall risk assessment results and local anomaly detection indicators; the doctor terminal is used to receive and display the overall risk assessment results and local anomaly detection indicators, and receive and execute strategy information.

[0008] Optionally, the patient terminal includes a data acquisition module, a data preprocessing module and a wireless transmission module; the data acquisition module is used to collect blood pressure, body position change and endocrine hormone fluctuation data in real time; the data preprocessing module is used to perform preliminary encryption and formatting on the collected blood pressure, body position change and endocrine hormone fluctuation data; the wireless transmission module is used to upload the processed data to the data center terminal via a wireless network.

[0009] Optionally, the data center terminal includes a data fusion module, a storage management module and a risk assessment module; the data fusion module is used to integrate multi-source data from the patient terminal; the storage management module is used to store and manage the fused data; and the risk assessment module is used to perform an overall risk assessment based on the multi-source data from the patient terminal.

[0010] Optionally, the local abnormality detection terminal includes a local data acquisition module, a signal processing module and an abnormality detection module; the local data acquisition module is used to collect local blood flow rate and pulse wave conduction velocity data; the signal processing module is used to perform noise reduction and normalization processing on the local blood flow rate and pulse wave conduction velocity data; the abnormality detection module is used to perform local abnormality detection index analysis based on the local blood flow rate and pulse wave conduction velocity data.

[0011] Optionally, the strategy generation terminal includes a data analysis module and a strategy output module; the data analysis module is used to read and analyze the overall risk assessment results and local abnormality detection indicators; the strategy output module is used to generate corresponding strategy information based on the analysis results and transmit it to the doctor terminal.

[0012] Optionally, the doctor terminal includes an information display module, a remote interaction module and an instruction receiving module; the information display module is used to display the overall risk assessment results and local abnormality detection indicators in real time; the remote interaction module is used to realize remote communication between the doctor and the data center and other terminals; the instruction receiving module is used to receive, display and execute the policy information issued by the policy generation terminal.

[0013] Optionally, the risk assessment module includes an overall risk index calculation submodule and an overall risk assessment result generation submodule; the overall risk index calculation submodule is used to calculate the overall risk index based on the patient's blood pressure, body position changes and endocrine hormone fluctuation data; the overall risk assessment result generation submodule is used to generate a corresponding overall risk assessment result based on the overall risk index;

[0014] When the overall risk index calculation submodule is working, the following formula is satisfied:

[0015]

[0016] Among them, R represents the overall risk index; P i represents the blood pressure data of the i-th measurement, specifically the systolic pressure in this embodiment; P0 represents the baseline blood pressure data, specifically the standard systolic pressure in this embodiment; N represents the total number of measurements corresponding to the data not used in the previous calculation; K1 represents the adjustment constant, which is obtained by fitting historical measurement data or set by the administrator based on experience; Δt i It represents the time difference between the i-th measurement and the previous measurement. The time difference value range is between 2s and 30s. It is helpful to highlight the situation where the smaller the time difference, the greater the deviation of the patient's blood pressure data and the fluctuation of endocrine hormones, and thus improve the sensitivity of the calculation of the overall risk index. i represents the fluctuation rate of endocrine hormones at the time of the i-th measurement, specifically the standard deviation of cortisol in this embodiment; P avg represents the average blood pressure data, specifically the average systolic blood pressure in this embodiment; PCF i Indicates the frequency of body position changes; N change represents the total number of changes in the patient's body position during the observation period before the i-th measurement; T all represents the total length of the observation period before the i-th measurement; R reflects the patient's long-term blood pressure fluctuation risk. When R>r ref When the overall risk assessment result generation submodule generates an overall risk assessment result indicating that the patient needs timely care, for example, the system automatically triggers an early warning to remind the patient and the doctor to take intervention measures, such as adjusting medication, increasing monitoring frequency or conducting clinical review.

[0017] An artificial intelligence-based dynamic blood pressure monitoring and risk assessment method is applied to the artificial intelligence-based dynamic blood pressure monitoring and risk assessment system as described above. The artificial intelligence-based dynamic blood pressure monitoring and risk assessment method includes:

[0018] S1, real-time collection of patient blood pressure, body position changes and endocrine hormone fluctuation data;

[0019] S2, integrating, storing and processing the data uploaded by the patient terminal and performing an overall risk assessment;

[0020] S3, collect local blood flow rate and pulse wave velocity data, and perform local abnormality detection index analysis;

[0021] S4, generates strategy information based on the overall risk assessment results and local anomaly detection indicators;

[0022] S5, receives and displays the overall risk assessment results and local anomaly detection indicators, and receives and executes policy information.

[0023] The beneficial effects achieved by the present invention are:

[0024] 1. Through the systematic setting of patient terminals, doctor terminals, data center terminals, local abnormality detection terminals and strategy generation terminals, the full process closed loop of data collection, fusion, analysis, early warning and feedback in the process of dynamic blood pressure monitoring can be realized, which is conducive to improving the intelligence level of the monitoring system and then realizing the risk perception mechanism of multi-terminal collaboration, which is conducive to improving the comprehensiveness and timeliness of hypertension risk prediction and intervention.

[0025] 2. By setting up a data acquisition module, a data preprocessing module and a wireless transmission module in the patient terminal, it is beneficial to accurately acquire and format multi-source data such as blood pressure, body position changes and hormone fluctuations, thereby improving data quality and privacy security, which is conducive to improving the stability of system data upload and the reliability of subsequent risk calculations.

[0026] 3. By setting up a data fusion module, a storage management module, and a risk assessment module in the data center terminal, it is beneficial to uniformly integrate and process the multi-dimensional physiological signals uploaded by the patient terminal, thereby supporting overall risk modeling based on global trends, thereby facilitating trend analysis and personalized risk grading of medium- and long-term blood pressure fluctuations.

[0027] 4. By setting up a local data acquisition module, a signal processing module and anomaly detection module in the local anomaly detection terminal, it is beneficial to perform high-frequency monitoring and noise reduction processing of the local blood flow rate and pulse wave velocity, and then extract the microscopic dynamic characteristics of the local circulatory system, which is conducive to the real-time detection of potential acute blood pressure abnormalities and improve the system's response capability to sudden high-pressure events.

[0028] 5. By setting up a data analysis module and a strategy output module in the strategy generation terminal, it is helpful to automatically analyze the multi-level information of overall risk and local anomalies, and then formulate hierarchical intervention recommendations based on risk levels, which is conducive to realizing a dynamic and personalized intervention strategy push mechanism and enhancing the closed-loop control capability of the system.

[0029] 6. By setting up an information display module, a remote interaction module and a command receiving module in the doctor's terminal, it is convenient for doctors to obtain system evaluation results in a timely manner and make decisions in conjunction with the data center, which makes it easier for doctors to remotely intervene in high-risk patients and adjust treatment plans, thereby improving medical response efficiency and individual health management level.

[0030] 7. By constructing a nonlinear weighted function model based on blood pressure deviation, frequency of body position changes and hormone fluctuations, it is helpful to quantitatively characterize the degree of chronic blood pressure abnormalities in patients from the global trend dimension, and then realize the dynamic assessment of medium- and long-term risks, which will help doctors formulate more scientific and reasonable individualized hypertension intervention and management plans.

[0031] 8. By constructing a nonlinear fractional model that includes blood flow rate deviation, pulse wave velocity changes, and abnormal duration, it is helpful to reflect in real time the possibility of patients experiencing acute blood pressure abnormalities in a short period of time, thereby realizing a rapid warning and emergency response mechanism at the second level, thereby helping to avoid serious threats to patients' lives caused by sudden hypertension events.

[0032] To further understand the features and technical contents of the present invention, please refer to the following detailed description and drawings of the present invention. However, the drawings provided are only for reference and illustration and are not intended to limit the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic diagram of the overall structure of the present invention;

[0034] Figure 2 Schematic diagram of the structure of the risk assessment module in the present invention;

[0035] Figure 3 This is the radar chart of the overall risk index in the present invention;

[0036] Figure 4 Schematic diagram of a method flow of a dynamic blood pressure monitoring and risk assessment method based on artificial intelligence in the present invention;

[0037] Figure 5 This is a local anomaly detection heat map according to another embodiment of the present invention. DETAILED DESCRIPTION

[0038] The following is an explanation of the embodiments of the present invention through specific embodiments. Those skilled in the art can understand the advantages and effects of the present invention from the contents disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and the details in this specification can also be modified and changed based on different viewpoints and applications without departing from the spirit of the present invention. In addition, the drawings of the present invention are only for simple schematic illustrations and are not depicted according to actual dimensions. It is stated in advance. The following embodiments will further explain the relevant technical contents of the present invention in detail, but the disclosed contents are not intended to limit the scope of protection of the present invention.

[0039] Hypertension is a common chronic disease and a major contributor to the global burden of disease. As of 2019, the cumulative number of people suffering from the disease worldwide exceeded 1.28 billion, with 245 million in my country, and the prevalence is increasing annually. Cardiovascular disease currently leads the world in both morbidity and mortality, with the number of deaths from cardiovascular disease projected to reach 23.6 million by 2030. Poorly controlled blood pressure is a significant contributing factor, with each 10 mmHg increase in average blood pressure associated with a 30% increase in risk. Hypertension is also a leading cause of cardiovascular disease (CVD) mortality. In 2020, 4.5766 million deaths from cardiovascular and cerebrovascular diseases were reported in my country, with 70% of strokes and 50% of ischemic heart disease attributable to hypertension. Strengthening hypertension prevention and control is urgent.

[0040] However, hypertension prevention and control are hampered by numerous factors. According to statistics, more than half of patients are unaware of their hypertension, and only 30%-50% are aware of their condition and regularly take their medication. Key factors include patient compliance and inadequate management. In 2018, the prevalence of hypertension was 27.5%, while awareness, treatment, and control rates among adults aged 18 and older were only 41.0%, 34.9%, and 11.0%, respectively. Chronic disease prevention and control requires lifelong, ongoing self-management. Blood pressure monitoring and risk management are fundamental to hypertension prevention and control, and are crucial for increasing awareness, treatment, and control rates, and reducing the health risks of hypertension.

[0041] Data-driven chronic disease management is an effective means of improving the efficiency and quality of chronic disease management, and some initiatives have adopted this approach. For example, the use of information technology to establish a remote chronic disease management model has been widely adopted and has achieved significant results. Other initiatives utilize the National DNA Management System to combine DNA and patient medical data to provide personalized diagnosis and treatment.

[0042] At the same time, new technologies are providing a convenient way to manage chronic diseases. One solution's health management platform stores user health and medical data, including medication prescriptions and medical history. It connects to wearable devices and collects users' health signs and exercise status data online in real time, serving both doctors and patients. This system provides patients with tools to manage their digital medical records and establish their own historical health profiles. Based on a patient's daily physiological data, doctors can pre-diagnose potential health issues and determine whether contact with the patient is necessary for further diagnosis and treatment, thus achieving comprehensive health promotion and management tracking for users. However, subsequent research has also indicated that the continued provision and protection of this information requires follow-up and resolution.

[0043] The development of the Internet Plus, Internet of Things Plus, Visual Internet Plus, wearable devices, and 5G industries is driving digital, intelligent self-management of health and hypertension, becoming a key tool for strengthening hypertension prevention and control at the grassroots level. In the absence of antihypertensive medications, new models based on digital diagnosis and treatment systems have significantly improved blood pressure reduction through 24-hour ambulatory blood pressure monitoring and office blood pressure monitoring compared to conventional management models.

[0044] The team collected household monitoring data for 500,000 patients with major chronic diseases, including hypertension. Leveraging IoT technologies, the data was uploaded to a smart chronic disease management platform via Bluetooth, Wi-Fi, or other mobile app or web-based connections. This ensured consistency and comparability in the collection, storage, publication, and exchange of household monitoring information for patients with major chronic diseases, including hypertension. They also developed interface specifications for data transmission and sharing for home health monitoring devices, effectively connecting household monitoring data with medical data. By developing regional standards for sharing household monitoring information for chronic disease patients, the team collected household monitoring data and uploaded it to the Hunan Provincial Family Wearable Device Data Platform, achieving regional data interoperability. To address the lack of decision-making support tools for chronic disease data like hypertension and to support grassroots chronic disease prevention and control, the team innovatively developed a regional decision-making system for major chronic diseases. They pioneered a series of big data-assisted decision-making support applications and tools based on knowledge graphs, the ClinRank weighted heterogeneous star network multi-layer ranking algorithm, and a similar case intelligent search engine. The medication recommendation software uses a multi-layer ranking algorithm based on a weighted heterogeneous star network knowledge graph of clinical data and literature data, and provides data-driven reminder, warning, and push decision support for patients with chronic diseases such as hypertension.

[0045] Research and application of digital and intelligent approaches to the prevention and treatment of chronic hypertension primarily focus on hypertension risk prediction and information-based management. A machine learning meta-algorithm was used to examine the heterogeneous treatment effects of intensive hypertension treatment among individuals and to predict the 3-year risk of cerebrovascular disease. The predicted area under the receiver operating characteristic (ROC) curve was 0.60, superior to the AUC of traditional logistic regression. Furthermore, deep neural networks were used to predict blood pressure variability in hypertensive patients 1-4 weeks later, and ML techniques were used to risk stratify individuals with different levels of blood pressure variability, with stability scores of 0.98 and 0.91, respectively. Studies have shown that approximately one-seventh of people with high blood pressure variability are at increased risk of stroke and heart failure. These risk predictions may provide additional insights into physicians' clinical judgments but have yet to be truly applied in hypertension prevention and control practices.

[0046] A study using an information management system on hypertensive patients who did not receive drug treatment found that the blood pressure reduction effect after intervention was better than that of the group with simple lifestyle adjustment. In addition, we divided 480 hypertensive patients into an intervention group and a control group, and used the system to conduct remote information management of the intervention group. The research results showed that the medication compliance of the intervention group was significantly improved and better than that of the control group.

[0047] However, intelligent hypertension management systems involve multiple fields, including medicine, statistics, artificial intelligence, and computer science. They are still cutting-edge technologies, requiring further exploration and improvement in specific rule-building, operational procedures, and functional refinement. For patients and hypertension professionals, intelligent hypertension management systems should offer a variety of functions, including blood pressure monitoring and analysis, blood pressure warnings, medication reminders, and information sharing while protecting personal data privacy. However, the application market shows a wide range of functionalities, requiring further standardization and improvement. Most applications have low download counts, lack clear user instructions, and are likely developed by a lack of professional medical professionals. Furthermore, most applications have a single function, while a few offer comprehensive functionality. Furthermore, most lack a clear theoretical foundation. An analysis of 73 hypertension management applications from the perspectives of data collection and functional application revealed that no single application addresses all management needs. Most offer limited functionality, such as demographic information collection and vital sign collection, while other intelligent features are insufficient. Furthermore, data security and privacy protection are also lacking.

[0048] New data-driven chronic disease management models that integrate digitalization, intelligentization, and other technologies are the future trend. Therefore, based on current developments and major trends both domestically and internationally, and the needs of hypertension prevention and management in the context of future digitalization, this study aims to develop an intelligent, integrated software-based medical device that integrates hypertension monitoring and follow-up data to enable real-time blood pressure monitoring, follow-up, return visits, risk assessment and early warning, and education. This research aims to improve patient compliance and reduce cardiovascular and cerebrovascular disease risks while respecting the practical needs of both physicians and patients and adhering to scientific objectivity. While ensuring data security and privacy, it aims to meet the needs of digital hypertension prevention and control that are efficient, convenient, easy to use, share information, reduce costs, increase efficiency, and manage patients flexibly. The development of this software holds significant value and broad application prospects for its widespread future application and promotion of intelligent hypertension management, improving prevention and control efficiency, increasing hypertension awareness, treatment, and control rates, and reducing cardiovascular and cerebrovascular disease risks.

[0049] Example 1: This example provides a blood pressure dynamic monitoring and risk assessment system based on artificial intelligence. Figure 1 As shown, the system includes a patient terminal, a doctor terminal, a data center terminal, a local abnormality detection terminal and a strategy generation terminal; the patient terminal is used to collect the patient's blood pressure, body position changes and endocrine hormone fluctuation data in real time; the data center terminal is used to integrate, store and process the data uploaded by the patient terminal, and perform overall risk assessment; the local abnormality detection terminal is used to collect local blood flow rate and pulse wave conduction velocity data, and perform local abnormality detection index analysis; the strategy generation terminal is used to generate strategy information based on the overall risk assessment results and local abnormality detection indicators; the doctor terminal is used to receive and display the overall risk assessment results and local abnormality detection indicators, and receive and execute strategy information.

[0050] Optionally, the patient terminal includes a data acquisition module, a data preprocessing module and a wireless transmission module; the data acquisition module is used to collect blood pressure, body position change and endocrine hormone fluctuation data in real time; the data preprocessing module is used to perform preliminary encryption and formatting on the collected blood pressure, body position change and endocrine hormone fluctuation data; the wireless transmission module is used to upload the processed data to the data center terminal via a wireless network.

[0051] Optionally, the data center terminal includes a data fusion module, a storage management module and a risk assessment module; the data fusion module is used to integrate multi-source data from the patient terminal; the storage management module is used to store and manage the fused data; and the risk assessment module is used to perform an overall risk assessment based on the multi-source data from the patient terminal.

[0052] Optionally, the local abnormality detection terminal includes a local data acquisition module, a signal processing module and an abnormality detection module; the local data acquisition module is used to collect local blood flow rate and pulse wave conduction velocity data; the signal processing module is used to perform noise reduction and normalization processing on the local blood flow rate and pulse wave conduction velocity data; the abnormality detection module is used to perform local abnormality detection index analysis based on the local blood flow rate and pulse wave conduction velocity data.

[0053] Optionally, the strategy generation terminal includes a data analysis module and a strategy output module; the data analysis module is used to read and analyze the overall risk assessment results and local abnormality detection indicators; the strategy output module is used to generate corresponding strategy information based on the analysis results and transmit it to the doctor terminal.

[0054] Optionally, the strategy information may be, but is not limited to: 1. Pushing a medium- to long-term trend analysis report; 2. Reminding doctors to evaluate whether to adjust antihypertensive drugs; 3. Recommending regular review of hormones, electrocardiograms, etc.; 4. Increasing the blood pressure sampling frequency, such as from once an hour to once every 20 minutes; 5. Immediate alarm: The system issues a sound and visual alarm to remind patients of high risk; 6. Recommending patients to use emergency antihypertensive drugs; 7. Maintaining the current status.

[0055] Optionally, the doctor terminal includes an information display module, a remote interaction module and an instruction receiving module; the information display module is used to display the overall risk assessment results and local abnormality detection indicators in real time; the remote interaction module is used to realize remote communication between the doctor and the data center and other terminals; the instruction receiving module is used to receive, display and execute the policy information issued by the policy generation terminal.

[0056] Optional, combined Figure 2 As shown, the risk assessment module includes an overall risk index calculation submodule and an overall risk assessment result generation submodule; the overall risk index calculation submodule is used to calculate the overall risk index based on the patient's blood pressure, body position changes and endocrine hormone fluctuation data; the overall risk assessment result generation submodule is used to generate a corresponding overall risk assessment result based on the overall risk index, and the overall risk assessment result includes that the patient is currently in good condition, the patient needs to be cared for in time, and the patient needs to seek medical treatment in time;

[0057] When the overall risk index calculation submodule is working, the following formula is satisfied:

[0058]

[0059] Among them, R represents the overall risk index; P irepresents the blood pressure data of the i-th measurement, specifically the systolic pressure in this embodiment; P0 represents the baseline blood pressure data, specifically the standard systolic pressure in this embodiment; N represents the total number of measurements corresponding to the data not used in the previous calculation; K1 represents the adjustment constant, which is obtained by fitting historical measurement data or set by the administrator based on experience; Δt i It represents the time difference between the i-th measurement and the previous measurement. The time difference value range is between 2s and 30s. It is helpful to highlight the situation where the smaller the time difference, the greater the deviation of the patient's blood pressure data and the fluctuation of endocrine hormones, and thus improve the sensitivity of the calculation of the overall risk index. i represents the fluctuation rate of endocrine hormones at the time of measurement i, specifically the standard deviation of cortisol; P avg represents the average blood pressure data, specifically the average systolic blood pressure in this embodiment; PCF i Indicates the frequency of body position changes; N change represents the total number of changes in the patient's body position during the observation period before the i-th measurement; T all represents the total length of the observation period before the i-th measurement; R reflects the patient's long-term blood pressure fluctuation risk; when R>r ref1 When r ref1 ≥R≥r ref2 When r ref1 >R, the overall risk assessment result generating submodule generates an overall risk assessment result indicating that the patient is currently in a good condition; r ref1 、r ref2 The upper and lower thresholds representing the risk of blood pressure fluctuation are set by the administrator based on experience.

[0060] The constant "60" in the formula is used for time unit conversion and is commonly used in time normalization. The constant "15" in the formula is a scaling factor, indicating the relative weight of the deviation of blood flow rate or pulse wave velocity relative to the reference value. In clinical practice, the fluctuation of blood flow rate or pulse wave velocity typically does not exceed 10%-20% of the reference value. The value of 15 ensures that deviations within a common range (e.g., 10%-20%) are appropriately normalized. In physiology and medicine, deviations of blood flow rate and pulse wave velocity typically have a reasonable range of variation. Typically, deviations of blood flow rate, for example, do not exceed 10% to 20% of the baseline value, and in some cases may reach around 30%. Therefore, 15 is an empirical value that appropriately compresses the deviation and ensures that the calculated risk value remains within a reasonable range. For example, in many clinical data sets, blood flow rate fluctuations are commonly between ±10% and ±20% of the baseline value. Therefore, choosing 15 ensures that the influence of the deviation term is not excessive. Other values can also be used, but they must ensure that they reflect actual physiological variations. The selected value should fit within the range of the actual physiological model. For example, if the blood flow rate or pulse wave velocity varies significantly, 15 may be too low. Try larger values such as 20 or 30 to limit the impact of the deviation. If the variation is small, use 10 or 5 to make the model more sensitive to small changes. 10 to 30 is a common range. The constant "0.05" in the formula is a normalization factor, which is primarily used to adjust the impact of the pulse wave conduction factor (PCF) on the calculated results. It normalizes the range of the PCF and prevents excessively large values from having an extreme effect in the calculation. The value of 0.05 is chosen based on the distribution of the PCF_i data. This constant ensures that the data is not overly amplified by extreme PCF values. For example, if the PCF is within a small range (such as 0 to 0.1), using 0.05 will reasonably include it in the calculation. Other values can be used to adjust based on specific experimental data or different application scenarios. If the PCF value is generally small (such as 0 to 0.02), you can reduce the value appropriately and choose 0.02 or 0.01 to increase its contribution to the calculation results. If the PCF variation range is large (such as 0 to 1), 0.05 may appear too small, and you can consider using 0.1 or 0.2 to mitigate its impact. Generally, 0.01 to 0.1 is a more common range, and the specific value will be determined according to the actual distribution of the PCF data. If the PCF values are mostly less than 0.1, then 0.05 is a reasonable normalization factor. If the PCF value is large, you can choose a larger value (such as 0.1 or 0.2) to reduce its impact on the final result.

[0061] There are generally several ways to monitor changes in body position: Inertial sensor (IMU): used to capture real-time angle changes and displacement trends, and then judge changes; pressure sensor array: used to identify whether the center of gravity of the force-bearing area of the body and the contact surface has changed; visual recognition module: assists in judging body position changes through skeleton reconstruction or image contour extraction. In this embodiment, a visual recognition module is preferred, which can identify the user's posture contour, key bone points or center of gravity structure, and track the changes of these features over time. If the posture feature exceeds the set threshold range or a state label switch occurs, it is determined to be a body position change. The above design can help this embodiment to timely monitor the blood pressure of patients who are active at home.

[0062] The following is an example of calculating the above overall risk indicator:

[0063] Given: N = 3, Δt i =[30min,30min,30min], P i = [140mmHg, 150mmHg, 160mmHg], P0 = 130mmHg, P avg =150mmHg, H i =[0.3,0.4,0.5], PCF i =[0.02,0.015,0.025]、K1=10、r ref1 =2.

[0064] Substitute into the calculation:

[0065]

[0066] Since R=2.24>r ref1 Therefore, the overall risk assessment result generation submodule generates an overall risk assessment result indicating that the patient needs to seek medical treatment in a timely manner, so that the system provides the following policy information:

[0067] 1. Immediate alarm: The system issues an audible and visual alarm to indicate a high risk.

[0068] 2. Patients are advised to use emergency antihypertensive drugs.

[0069] Combine Figure 3 As shown, Figure 3 This is a radar chart of overall risk indicators, which is used to show the horizontal comparison of multiple patients in the overall blood pressure risk score dimension.

[0070] The following is the program code for the above calculation example:

[0071]

[0072]

[0073] In practice, the patient terminal app integrates blood pressure monitoring, follow-up, re-examination, and risk warning modules, continuously assessing the user's blood pressure trends through AI models. The system generates dynamic blood pressure trajectory charts, displaying fluctuation curves by day, week, month, and year. Doctors can use this information to determine the patient's disease progression and control effectiveness.

[0074] When an individual's trajectory approaches a warning threshold, the system automatically triggers a follow-up reminder or warning alert, allowing doctors to remotely adjust target blood pressure values or treatment plans. The software also supports customizable reminders, such as follow-up appointment dates and medication schedules, ensuring a seamless user experience and clinical applicability.

[0075] In summary, through the systematic setting of patient terminals, doctor terminals, data center terminals, local anomaly detection terminals and strategy generation terminals, the full process closed loop of data collection, fusion, analysis, early warning and feedback in the process of dynamic blood pressure monitoring can be realized, which is conducive to improving the intelligence level of the monitoring system and thus realizing the risk perception mechanism of multi-terminal collaboration; by setting up data collection modules, data preprocessing modules and wireless transmission modules in the patient terminal, it is conducive to the accurate acquisition and formatting management of multi-source data such as blood pressure, body position changes and hormone fluctuations, thereby improving data quality and privacy security; by setting up data fusion modules, storage management modules and risk assessment modules in the data center terminal, it is conducive to the unified fusion processing of multi-dimensional physiological signals uploaded by the patient terminal, thereby supporting overall risk modeling based on global trends; by setting up local data collection modules, signal processing modules and abnormality modules in the local anomaly detection terminal The regular detection module is conducive to high-frequency monitoring and noise reduction of local blood flow rate and pulse wave velocity, and then extracting the microscopic dynamic characteristics of the local circulatory system; by setting up a data analysis module and a strategy output module in the strategy generation terminal, it is conducive to automatically analyzing the multi-level information of overall risk and local abnormalities, and then formulating stratified intervention recommendations based on the risk level; by setting up an information display module, a remote interaction module and an instruction receiving module in the doctor terminal, it is conducive to doctors obtaining system evaluation results in a timely manner and making joint decisions with the data center, thereby facilitating doctors to remotely intervene in high-risk patients and adjust treatment plans; by constructing a nonlinear weighted function model based on blood pressure deviation, posture change frequency and hormone fluctuations, it is conducive to quantitatively characterizing the degree of chronic blood pressure abnormalities of patients from the global trend dimension, and then realizing dynamic assessment of medium- and long-term risks, which is conducive to doctors formulating more scientific and reasonable individualized hypertension intervention and management plans.

[0076] A blood pressure dynamic monitoring and risk assessment method based on artificial intelligence is applied to a blood pressure dynamic monitoring and risk assessment system based on artificial intelligence as described above, combined with Figure 4 As shown, the artificial intelligence-based dynamic blood pressure monitoring and risk assessment method includes:

[0077] S1, real-time collection of patient blood pressure, body position changes and endocrine hormone fluctuation data;

[0078] S2, integrating, storing and processing the data uploaded by the patient terminal and performing an overall risk assessment;

[0079] S3, collect local blood flow rate and pulse wave velocity data, and perform local abnormality detection index analysis;

[0080] S4, generates strategy information based on the overall risk assessment results and local anomaly detection indicators;

[0081] S5, receives and displays the overall risk assessment results and local anomaly detection indicators, and receives and executes policy information.

[0082] Example 2: This example includes all the contents of Example 1 and provides an artificial intelligence-based dynamic blood pressure monitoring and risk assessment system. The local abnormality detection terminal includes a local data acquisition module, a signal processing module, and an abnormality detection module. The local data acquisition module is used to collect local blood flow rate and pulse wave velocity data. The signal processing module is used to perform noise reduction and normalization processing on the local blood flow rate and pulse wave velocity data. The abnormality detection module is used to perform local abnormality detection index analysis based on the local blood flow rate and pulse wave velocity data. The abnormality detection module includes a local abnormality detection index calculation submodule and a local abnormality detection index output submodule. The local abnormality detection index calculation submodule is used to calculate the local abnormality detection index based on the processed local blood flow rate and pulse wave velocity data. The local abnormality detection index output submodule is used to output the local abnormality detection index to the strategy generation terminal.

[0083] When the local anomaly detection index calculation submodule is working, the following formula is satisfied:

[0084]

[0085] Among them, r part It represents the local abnormality detection index, which is used to determine the patient's short-term risk; ΔV represents the blood flow rate deviation, ΔV = v t -v0,v t represents the blood flow rate measured by the local data acquisition module at the current moment, v0 represents the reference value of the blood flow rate in the resting state; ΔW represents the pulse wave velocity deviation, ΔW=w t -w0,w t represents the pulse wave velocity measured by the local data acquisition module at the current moment, w0 represents the pulse wave velocity reference value in the resting state; T partIndicates the duration of abnormal blood pressure. The local abnormality detection index is calculated when abnormal blood pressure occurs. ref Indicates the safe duration of abnormality, which is set by the administrator based on experience; when r part >r ref3 When the local abnormality detection index output submodule outputs the corresponding local abnormality detection index and indicates that the patient needs to be treated in time. ref3 Indicates the short-term abnormal risk threshold, which is set by the administrator based on experience.

[0086] The following is an example of calculating the above local anomaly detection indicators:

[0087] Given: v0 = 0.2 m / s, w0 = 8.0 m / s, T ref =120s, v t =0.25m / s, w t =9.5m / s, T part =90s, r ref2 =1.0.

[0088] Substitute into the calculation:

[0089]

[0090] Due to r part >r ref2 , the patient needs to be treated promptly, and the system takes the following specific measures:

[0091] 1. Immediate alarm: The system issues audible and visual alarms to alert patients and administrators of high risk;

[0092] 2. Patient response suggestions: Patients should immediately stop current activities and lie down to rest; take self-regulation measures such as deep breathing to help stabilize blood pressure;

[0093] 3. Doctor remote monitoring: The system will push the data to the doctor's terminal. The doctor can intervene remotely based on the real-time data and recommend the patient to seek medical treatment or use emergency antihypertensive drugs when necessary;

[0094] 4. Data synchronization and early warning: The system automatically records the patient's real-time data and synchronizes it to the hospital data center for subsequent health monitoring.

[0095] Combine Figure 5 As shown, Figure 5 This is a local anomaly detection heat map, which shows the local anomaly detection indicators of multiple locations at multiple time sampling points.

[0096] The following is the program code for the above calculation example:

[0097]

[0098]

[0099] In summary, by constructing a nonlinear fractional model that includes blood flow rate deviation, pulse wave velocity change, and abnormal duration, it is beneficial to reflect in real time the possibility of patients developing acute blood pressure abnormalities in a short period of time, thereby realizing a rapid warning and emergency response mechanism at the second level, thereby helping to avoid serious threats to patients' life safety caused by sudden hypertension events.

[0100] The contents disclosed above are only preferred feasible embodiments of the present invention and do not limit the scope of protection of the present invention. Therefore, all equivalent technical changes made using the contents of the present invention description and drawings are included in the scope of protection of the present invention. In addition, the elements therein can be updated as technology develops.

Claims

1. A blood pressure dynamic monitoring and risk assessment system based on artificial intelligence, characterized in that: It includes a patient terminal, a doctor terminal, a data center terminal, a local anomaly detection terminal, and a strategy generation terminal; the patient terminal is used to collect real-time data on the patient's blood pressure, body position changes, and endocrine hormone fluctuations; the data center terminal is used to integrate, store, and process the data uploaded by the patient terminal and perform an overall risk assessment; The local abnormality detection terminal is used to collect local blood flow rate and pulse wave conduction velocity data, and perform local abnormality detection index analysis; the strategy generation terminal is used to generate strategy information based on the overall risk assessment results and local abnormality detection indicators; the doctor terminal is used to receive and display the overall risk assessment results and local abnormality detection indicators, and receive and execute strategy information.

2. The artificial intelligence-based dynamic blood pressure monitoring and risk assessment system according to claim 1, characterized in that: The patient terminal includes a data acquisition module, a data preprocessing module and a wireless transmission module; the data acquisition module is used to collect blood pressure, body position change and endocrine hormone fluctuation data in real time; the data preprocessing module is used to perform preliminary encryption and formatting processing on the collected blood pressure, body position change and endocrine hormone fluctuation data; the wireless transmission module is used to upload the processed data to the data center terminal via a wireless network.

3. The artificial intelligence-based dynamic blood pressure monitoring and risk assessment system according to claim 2, characterized in that: The data center terminal includes a data fusion module, a storage management module and a risk assessment module; the data fusion module is used to integrate multi-source data from the patient terminal; the storage management module is used to store and manage the fused data; and the risk assessment module is used to perform an overall risk assessment based on the multi-source data from the patient terminal.

4. The artificial intelligence-based dynamic blood pressure monitoring and risk assessment system according to claim 3, characterized in that: The local abnormality detection terminal includes a local data acquisition module, a signal processing module and an abnormality detection module; the local data acquisition module is used to collect local blood flow rate and pulse wave conduction velocity data; the signal processing module is used to perform noise reduction and normalization processing on the local blood flow rate and pulse wave conduction velocity data; the abnormality detection module is used to perform local abnormality detection index analysis based on the local blood flow rate and pulse wave conduction velocity data.

5. The artificial intelligence-based dynamic blood pressure monitoring and risk assessment system according to claim 4, characterized in that: The strategy generation terminal includes a data analysis module and a strategy output module; the data analysis module is used to read and analyze the overall risk assessment results and local abnormality detection indicators; the strategy output module is used to generate corresponding strategy information based on the analysis results and transmit it to the doctor terminal.

6. The artificial intelligence-based dynamic blood pressure monitoring and risk assessment system according to claim 5, characterized in that: The doctor terminal includes an information display module, a remote interaction module and an instruction receiving module; the information display module is used to display the overall risk assessment results and local abnormality detection indicators in real time; the remote interaction module is used to realize remote communication between the doctor and the data center and other terminals; the instruction receiving module is used to receive, display and execute the policy information issued by the policy generation terminal.

7. The artificial intelligence-based dynamic blood pressure monitoring and risk assessment system according to claim 6, characterized in that: The risk assessment module includes an overall risk index calculation submodule and an overall risk assessment result generation submodule; the overall risk index calculation submodule is used to calculate the overall risk index based on the patient's blood pressure, body position changes and endocrine hormone fluctuation data; the overall risk assessment result generation submodule is used to generate a corresponding overall risk assessment result based on the overall risk index; When the overall risk index calculation submodule is working, the following formula is satisfied: Among them, R represents the overall risk index; P i represents the blood pressure data of the ith measurement; P0 represents the baseline blood pressure data; N represents the total number of measurements corresponding to the data not used in the calculation since the last calculation; K1 represents the adjustment constant; Δt i represents the time difference between the i-th measurement and the previous measurement; H i represents the fluctuation rate of endocrine hormones at the time of the i-th measurement; P avg Indicates average blood pressure data; PCF i Indicates the frequency of body position changes; N change represents the total number of changes in the patient's body position during the observation period before the i-th measurement; T all represents the total length of the observation period before the i-th measurement; R reflects the patient's long-term blood pressure fluctuation risk. The larger the overall risk index, the greater the patient's current condition risk.

8. A method for dynamic blood pressure monitoring and risk assessment based on artificial intelligence, applied to the system for dynamic blood pressure monitoring and risk assessment based on artificial intelligence as claimed in claim 7, characterized in that: The artificial intelligence-based dynamic blood pressure monitoring and risk assessment method includes: S1, real-time collection of patient blood pressure, body position changes and endocrine hormone fluctuation data; S2, integrating, storing and processing the data uploaded by the patient terminal and performing an overall risk assessment; S3, collect local blood flow rate and pulse wave velocity data, and perform local abnormality detection index analysis; S4, generates strategy information based on the overall risk assessment results and local anomaly detection indicators; S5, receives and displays the overall risk assessment results and local anomaly detection indicators, and receives and executes policy information.

Images