A comprehensive rehabilitation management system for patients with peripheral artery disease

Abstract

The application provides a peripheral arterial disease patient rehabilitation management system based on a whole process, and relates to the technical field of peripheral arterial disease patient rehabilitation management.The application comprises a patient screening and risk grading module, a rehabilitation evaluation module, a motion prescription generation and dynamic adjustment module, an intelligent propaganda content pushing module and a quality control closed loop management module.The application realizes accurate screening of PAD patients through a medical data interface and NLP technology, dynamically evaluates the risk level of the patients in combination with an intelligent risk grading system;generates personalized motion prescriptions according to evaluation reports, complications and biological monitoring data, and guarantees the safety of motion through a dynamic threshold model and collaborative early warning;realizes dynamic matching and timeliness pushing of propaganda content by using an improved KL divergence algorithm;and realizes whole process quality closed loop management through the quality control closed loop management module.The application can improve the accuracy and safety of PAD patient rehabilitation management, optimize medical resource allocation, and help whole process data-driven rehabilitation practice.

Description

Technical Field

[0001] This invention relates to the fields of medical information technology and rehabilitation management technology, and more specifically, to a rehabilitation management system for patients with peripheral artery disease based on the entire process. Background Technology

[0002] Peripheral arterial disease (PAD) is a chronic disease caused by atherosclerosis, leading to narrowing or blockage of peripheral arteries. Its incidence is rising globally, particularly among the elderly. PAD not only severely impacts patients' quality of life but can also lead to serious complications such as amputation or even death. Studies have shown that regular exercise therapy can effectively improve walking ability, physical function, and quality of life in PAD patients, and reduce morbidity and all-cause cardiovascular mortality. Therefore, it is recommended as a first-line treatment in domestic and international guidelines.

[0003] However, the rehabilitation management of PAD patients still faces many challenges. Traditional rehabilitation management models mainly rely on manual operations, including patient assessment, exercise prescription development, data recording and analysis. This approach is not only inefficient but also prone to errors due to human factors, making it difficult to meet the needs of long-term follow-up and precise management. Meanwhile, while existing information systems have achieved some degree of automation, they still have significant shortcomings. For example, patient screening mechanisms largely rely on structured diagnostic coding, which has limited ability to identify ambiguous diagnoses such as "suspected PAD," potentially leading to the omission of high-risk patients. Risk grading systems generally adopt a static judgment model based on a single clinical indicator, failing to achieve multi-dimensional dynamic fusion analysis and unable to capture changes in patient risk levels in a timely manner. The personalized education and delivery modules lack sufficient personalization, relying solely on disease diagnosis tags for content matching, lacking the ability to adaptively adjust to real-time scenarios such as postoperative changes and decreased behavioral compliance. Quality control management also largely remains at the data statistics level, lacking a closed-loop quality control system based on multi-dimensional indicators, and unable to automatically identify process loopholes and trigger improvement measures. Furthermore, there are barriers to integrating real-time biometric data monitored by IoT devices with medical systems. The data collected by the devices has not been effectively incorporated into the exercise prescription generation and adjustment algorithms, resulting in a lack of individualized accuracy in exercise intensity calculations. These issues directly affect the timeliness and safety of rehabilitation management, necessitating a comprehensive rehabilitation management system that covers intelligent data analysis, dynamic risk prediction, and multimodal collaborative analysis to improve the rehabilitation outcomes and management efficiency for PAD patients.

[0004] In view of the above, this application is hereby submitted. Summary of the Invention

[0005] This invention aims to provide a full-process rehabilitation management system for patients with peripheral artery disease, in order to address the shortcomings of existing methods, such as low data management efficiency, insufficient intelligence in exercise prescriptions, low personalization of health education, and imperfect quality control systems.

[0006] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0007] A full-process rehabilitation management system for patients with peripheral artery disease includes: a patient screening and risk grading module, a rehabilitation assessment module, an exercise prescription generation and dynamic adjustment module, an intelligent push module for health education content, and a quality control closed-loop management module.

[0008] The patient screening and risk grading module is used to connect with the hospital information system through a medical data interface, obtain and analyze the basic data of patients with peripheral artery disease in real time, and build an intelligent risk grading system based on Fontaine staging, ankle-brachial index and contraindications to automatically update the patient's risk level and trigger an early warning.

[0009] The rehabilitation assessment module is used to generate an assessment report based on the patient's basic data and real-time acquired biomonitoring data, with time as the axis.

[0010] The exercise prescription generation and dynamic adjustment module is used to generate an exercise prescription based on the assessment report, comorbidities, and real-time acquired biomonitoring data, according to applicable items and safety principles, and to dynamically optimize and adjust it according to the intelligent risk grading system.

[0011] The intelligent push module for health education content is used to collect multi-dimensional features of patients to construct patient feature profiles, establish a health education knowledge base, generate initial health education content, and dynamically adjust and personalize the initial health education content for adaptive push.

[0012] The quality control closed-loop management module is used to construct a standardized quality control indicator system, generate quality control data, and provide early warnings and improvements based on the processing and feedback from the rehabilitation assessment module, the exercise prescription generation and dynamic adjustment module, and the intelligent push module for educational content.

[0013] Preferably, the patient's basic data includes the patient's basic information, diagnostic code, and test report;

[0014] If the diagnostic code is a fuzzy diagnosis of suspected peripheral artery disease, then NLP technology is used to analyze and extract key disease features based on the patient's basic data, and the patient screening results are dynamically corrected in conjunction with the patient's test reports.

[0015] Preferably, the biomonitoring data includes exercise time, distance, steps, real-time speed, metabolic equivalent, dynamic heart rate, blood oxygen, plantar pressure distribution, and blood pressure; the assessment report includes: the assessed patient, assessment date, height, weight, BMI, medications used, blood glucose, blood lipids, blood pressure, smoking status, alcohol consumption status, exercise habits, wound condition, visual analog scale (VAS) for pain, lower limb muscle strength, simplified physical fitness test (SPPB), 6-minute walking distance (6MWD) and steps, limping distance, maximum walking distance, ankle-brachial index (ABI), leg circumference, quality of life scale, gait disorder questionnaire, hope scale, and subjective fatigue score;

[0016] Wherein, BMI = weight / height 2 ;

[0017] The 6MWD is divided into male and female types. The male 6MWD is calculated as 1140 (meters) - 5.61 × BMI - 6.94 × age, and the female 6MWD is calculated as 1017 (meters) - 6.24 × BMI - 5.83 × age, where age is in years.

[0018] Preferably, the exercise prescription includes the patient's basic exercise parameters and displays recommendations for comorbidities; the patient's basic exercise parameters include exercise frequency, exercise intensity, exercise duration, exercise type, and exercise volume;

[0019] The intensity of the exercise is calculated using the following method:

[0020] ;

[0021] Where I represents exercise intensity; 6MWD represents a 6-minute walking distance.

[0022] The quantification model for the amount of exercise is as follows:

[0023] ;

[0024] in, I represents the amount of exercise; I represents the intensity of exercise. Indicates the total duration of the exercise; This indicates the actual number of steps taken in a 6-minute walking test. Safety Factor This represents the safety adjustment factor.

[0025] Preferably, when dynamically optimizing and adjusting according to the intelligent risk grading system, the matching degree of the exercise prescription is intelligently judged, and the data parameters of the exercise prescription are automatically checked and adjusted for threshold conflicts.

[0026] Preferably, the patient profile includes disease diagnosis results, clinical data, medication list, surgical records, exercise prescriptions, and patient lifestyle habits;

[0027] When generating initial missionary content, each piece of content in the missionary knowledge base is labeled with multi-dimensional tags, and then content matching is performed based on KL divergence to filter out the missionary content with the highest matching degree to the current time point; the matching degree is calculated using the following formula:

[0028] ;

[0029] in, Indicates the degree of matching; This indicates the timeliness of knowledge at the current time point t. , For decay rate, Indicates the interval between the knowledge publication date and the current time; This represents the total number of feature dimensions in a patient profile. Represents the feature dimension variable; Represents the weight coefficient of the current feature dimension; Represents the patient feature vector; This represents the i-th knowledge entry; A patient feature vector representing the feature dimension; Knowledge entries representing feature dimensions; Indicates intersection; Represents the union; Indicates the value to be retrieved;

[0030] When the matched content of the selected missionary materials is equal, the content with the higher reading completion rate is selected as the initial missionary material.

[0031] Preferably, the initial educational content is dynamically adjusted and personalized for adaptive delivery, specifically as follows:

[0032] The system collects multi-dimensional characteristics of patients in real time and updates the educational knowledge base accordingly.

[0033] Based on the updated mission knowledge base, the knowledge weights of the initial mission content are dynamically adjusted using a time decay mechanism, expressed as:

[0034] ;

[0035] in, The knowledge weights of the updated initial missionary content; The knowledge weight of the initial educational content before the update; n is the number of push cycles; This represents the number of days since the last valid read. Indicates periodic decay; Indicates continuous decay; Forgetting factor; The event correlation degree for the i-th push period Indicator functions.

[0036] Analyze patients' real-time reading behavior data of the initial educational content, calculate the similarity of patients' reading behavior, and generate a patient recommendation list. The expression is:

[0037] ;

[0038] ;

[0039] in, This indicates the similarity in reading behavior between patient u and patient v; This represents the set of numbers of shared push cycles for patients u and v. This indicates the patient u's reading status of the educational content in the i-th push cycle. A value of 1 indicates a reading time of ≥30s, representing in-depth reading; a value of 0.5 indicates a reading time of <30s, representing fast reading; and a value of 0 indicates that the content was not opened. This indicates the patient v's reading status of the educational content in the i-th push cycle; This represents the average reading time of patient u across all push notification cycles for educational content. This represents the average reading time of patient v across all push notification cycles for educational content;

[0040] Indicates the patient Recommended list; This represents a set of patients with similar reading behaviors to patient u. This indicates the patient v's reading status of the educational content in the j-th push cycle;

[0041] Based on the multi-dimensional characteristics of patients collected in real time, when abnormal indicators of patients are detected, content related to the abnormal indicators is searched from the health education knowledge base, inserted at the top of the patient recommendation list, and an alert is sent to the medical staff.

[0042] Preferably, the method further includes: when generating an exercise prescription, establishing a dynamic threshold model based on real-time acquired biomonitoring data and performing real-time feedback training; the dynamic threshold model includes a maximum heart rate threshold, the expression of which is:

[0043] Maximum heart rate threshold = 220 - patient age ± daily blood pressure fluctuation value.

[0044] Preferably, the method further includes: during the execution of the exercise prescription, calculating the daily exercise target achievement rate using a dynamic target achievement algorithm based on the exercise prescription and real-time acquired biomonitoring data, and providing real-time feedback on the patient's exercise results; wherein, the formula for calculating the daily exercise target achievement rate is:

[0045] Daily exercise target achievement rate = Σ (Daily completion amount × Time decay coefficient) / Exercise prescription requirement; Σ represents the sum of all exercise completed on that day;

[0046] Time decay coefficient = 1 / (1 + 0.1 × number of days of delay);

[0047] The exercise results are fed back through the Vascular Health Index (VHI) score, calculated using the following formula:

[0048] VHI = 0.3 × 6MWD progress rate + 0.2 × gait symmetry improvement + 0.5 × prescription compliance;

[0049] Among them, the 6MWD progress rate represents the results of two 6-minute walking distance tests before and after treatment, i.e., (6MWD after treatment - 6MWD before treatment) / 6MWD before treatment × 100%; the gait symmetry improvement rate represents the degree of improvement in the patient's gait symmetry, which is obtained by comparing the difference in gait between the left and right sides of the patient before and after treatment; the prescription compliance rate represents the degree to which the patient follows the exercise prescription issued by the doctor, i.e., (actual number of exercise sessions / number of exercise sessions prescribed by the prescription) × 100%.

[0050] Preferably, the standardized quality control indicator system includes the timeliness of rehabilitation assessment, the timeliness of patient enrollment, the matching degree of exercise prescription, the effectiveness rate of patient exercise, the implementation rate of health education, the timeliness rate of health education notification, the health education reading rate, and the effective reading rate.

[0051] The calculation method for the timeliness rate of rehabilitation assessment is: (number of patients who completed the assessment on time / number of patients who should be assessed) × 100%, and on-time completion is defined as: assessment date ≤ preset date + 1 working day;

[0052] The timely patient enrollment rate is calculated as follows: (Number of patients whose enrollment time is ≤ 24 hours after admission / Total number of newly enrolled patients) × 100%;

[0053] The exercise prescription matching rate is calculated as follows: (achieving rate × 80% - 120% × number of patients / number of patients enrolled in the system) × 100%; Achieving rate = (actual amount of exercise completed by the patient / amount of exercise required by the exercise prescription) × 100%;

[0054] The method for calculating the patient's exercise effectiveness rate is as follows: Improvement rate of primary indicators = (Current 6MWD - Baseline 6MWD) / Baseline 6MWD × 100%;

[0055] The implementation rate of health education campaigns is calculated as follows: (Number of patients actually reached with health education content / Total number of patients who should have received health education) × 100%;

[0056] The timely notification rate is calculated as follows: (Number of patients notified of educational content within 3 days / Total number of patients who did not read the notification) × 100%;

[0057] The health education reading rate is calculated as follows: (Number of patients who received health education reading ≥ 1 time / Total number of patients received) × 100%;

[0058] The effective reading rate is calculated as follows: (Number of patients who engage in deep reading / Total number of patients who engage in reading) × 100%. Deep reading is defined as a single dwell time of ≥120 seconds and completion of the knowledge test.

[0059] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0060] First, the system of this invention intelligently integrates multi-source patient data and dynamic risk management, significantly improving the identification response speed for PAD patients. Specifically, based on the FHIR standard and NLP unstructured text parsing technology, the system extracts PAD patient information from the hospital information system in real time, breaking through the limitations of traditional structured coding and accurately identifying patients with ambiguous diagnoses. By dynamically fusing Fontaine staging, ABI index, and contraindications using a four-color risk matrix, a minute-level risk scanning mechanism is established, greatly reducing the missed screening rate and false alarm rate for high-risk patients. Second, the intelligent generation and dynamic adjustment module of exercise prescriptions effectively reduces the occurrence of exercise-related adverse events. This module adopts a three-layer architecture of "multimodal data input - tagged action library matching - conflict detection," combined with a dynamic threshold model and a collaborative early warning network, to monitor the patient's physiological indicators during exercise in real time, ensuring exercise safety.

[0061] The intelligent content delivery module for health education is based on an improved KL divergence algorithm. The system constructs a deep matching model between patient feature vectors and a knowledge base, enabling timely delivery of health education content with decaying effectiveness and event-triggered updates. During exercise prescription execution, IoT devices collect real-time biometric data such as foot pressure and gait symmetry to dynamically optimize the achievement rate and achieve precise matching of exercise intensity and physiological state. The quality control closed-loop management module drives management upgrades, enabling automatic problem identification and standardized improvement process embedding, reducing manpower costs for quality control management.

[0062] This invention utilizes an automated risk grading and early warning freezing mechanism to prioritize the allocation of medical resources to high-risk patients, thereby improving the efficiency of managing critical rehabilitation values. The end-to-end management system maintains a traceable data chain for all operations, from prescription generation to quality control improvements, providing high-value datasets for clinical research and supporting the dynamic optimization of PAD patient rehabilitation and precision medicine practices.

[0063] In summary, this invention addresses the pain points of existing PAD patient rehabilitation management systems in areas such as data management, intelligent exercise prescriptions, personalized education and delivery, and improved quality control systems through multimodal data fusion, intelligent algorithm collaboration, and closed-loop quality control mechanisms, significantly improving the accuracy and safety of rehabilitation management. Attached Figure Description

[0064] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0065] Figure 1 This is a schematic diagram of a full-process peripheral artery disease patient rehabilitation management system provided in Example 1.

[0066] Figure 2 This is a schematic diagram of the compliance exception handling process provided in Example 1.

[0067] Figure 3 This is a schematic diagram of the three-level response mechanism provided in Example 1.

[0068] Figure 4 This is a schematic diagram of the educational knowledge base structure provided in Example 1.

[0069] Figure 5 This is a schematic diagram of the three-level response process for handling abnormal heart rate provided in Example 1.

[0070] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Detailed Implementation

[0071] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0072] Example 1

[0073] Embodiment 1 of the present invention provides a full-process peripheral artery disease patient rehabilitation management system, which can be implemented by a full-process peripheral artery disease patient rehabilitation management device (hereinafter referred to as management device), specifically, executed by one or more processors within the management device.

[0074] In this embodiment, the management device may be an electronic device equipped with a processor, which carries a computer program for the full-process peripheral artery disease patient rehabilitation management system and the computer program can be executed, such as a computer, smartphone, smart tablet, workstation, etc., which are not limited here.

[0075] In this embodiment, the implementation of the full-process peripheral artery disease patient rehabilitation management system is divided into two parts: a PC terminal and an APP terminal. The PC terminal is mainly used by rehabilitation administrators, while the APP terminal is used by patients and their families. The PC terminal includes modules for patient management, rehabilitation assessment, exercise prescription generation, exercise data recording, education, and quality control; the APP terminal includes intelligent rehabilitation, adaptive education, and intelligent safety monitoring modules. The system adopts a dual-channel encrypted transmission architecture to achieve efficient synchronization of medical data and IoT device data, and supports real-time push of vital sign data, daily batch synchronization of assessment reports, and immediate forced synchronization of exercise prescription adjustments, ensuring flexible configuration of data update frequency and a balance between real-time performance and accuracy.

[0076] like Figure 1 As shown, a full-process rehabilitation management system for patients with peripheral artery disease includes a patient screening and risk grading module, a rehabilitation assessment module, an exercise prescription generation and dynamic adjustment module, an intelligent push module for health education content, and a quality control closed-loop management module.

[0077] In the patient screening and risk stratification module, administrators can use a new function to filter patients diagnosed with PAD in the hospital information HIS system and import them into this system. This function uses the FHIR medical data interface and NLP text parsing technology to extract key disease features from unstructured text, such as intermittent claudication or ABI values.

[0078] Patient basic data includes the patient's basic information, diagnostic code (e.g., ICD-1: PAD), and laboratory reports. If the diagnostic code is a vague diagnosis of suspected PAD, NLP technology is used to analyze the patient's basic data and extract key disease features. Combined with the patient's laboratory reports, the patient screening results are dynamically corrected.

[0079] After screening, the system automatically labels the patient's basic data, including name, gender, age, diagnosis, time joined for rehabilitation, contact information, and whether the case is closed or not. Clicking on a patient's name will take you to their profile page to view their general information, which is automatically imported from the HIS system and includes height, weight, disease diagnosis, surgery, medical records, rehabilitation assessment, exercise prescription, and exercise records. Within the patient profile page, administrators can also set filter criteria to view lists of closed and open cases, and click to view detailed information for each patient.

[0080] Then, based on this, an intelligent risk grading system was constructed using Fontaine staging (a commonly used clinical staging method for lower extremity arteriosclerosis obliterans), ankle-brachial index, and contraindications to automatically update patient risk levels and trigger alerts. As shown in Table 1, the four-color risk matrix automatically labels different risk groups with colors. The risk status is automatically scanned and updated every morning. When a patient's risk level upgrades, an alert is sent to the administrator. When an extremely high risk is detected, the generation of exercise prescriptions is immediately frozen and an alert is sent to the administrator, achieving automatic identification and dynamic monitoring of high-risk patients.

[0081] Table 1. Four-color risk classification of PAD patients

[0082]

[0083] The rehabilitation assessment module is structured chronologically, with the system defaulting to assessment dates for each patient: upon joining rehabilitation, weekly during hospitalization, upon discharge, and monthly thereafter. Administrators can use the filtering function to view lists of patients requiring assessment today, patients whose assessment dates have not yet arrived, and patients whose assessment dates have passed. Clicking on a patient's name leads to the specific assessment interface, which includes assessment content such as height, weight, BMI, medication status, blood glucose, blood lipids, blood pressure, smoking status, alcohol consumption status, exercise habits, wound condition, VAS score, lower limb muscle strength, SPPB, 6MWD and steps, limp distance, maximum walking distance, ABI, leg circumference, quality of life scale, gait disorder questionnaire, hope scale, and Borg scale. BMI is automatically calculated by the system using the formula: BMI = weight / height. 2 The ideal 6MWD value is automatically generated according to a formula, categorized as male and female. For males, 6MWD = 1140 - 5.61 × BMI - 6.94 × age; for females, 6MWD = 1017 - 6.24 × BMI - 5.83 × age, with age in years. This facilitates administrator assessment of patient performance. After the assessment, the system automatically generates an assessment report and prompts whether to generate an exercise prescription. If the system selects to generate a prescription, it will redirect to the exercise prescription generation interface; otherwise, it will remain on the current interface.

[0084] The assessment data statistics interface also allows you to view trend charts and data for each assessment item by day, week, month / year (1. Select the assessment date range (e.g., "2024 / 07 / 01-2024 / 07 / 31"); 2. The system automatically retrieves: electronic medical record data (ABI value, medication records, etc.), wristband data (daily average step count change curve), and video assessment results (wound healing progress analysis); 3. Generate a PDF assessment report (including risk trend radar chart) to determine the patient's recovery effectiveness).

[0085] The exercise prescription generation and dynamic adjustment module is based on a three-layer architecture: multimodal data input, tagged action library matching, and conflict detection. The bottom layer integrates assessment report results, comorbidities, and real-time vital signs from multimodal patient data. The middle layer generates a basic prescription based on applicable principles for comorbidities. The top layer combines safety rules and physician confirmation to output the final exercise prescription. The exercise prescription includes the patient's basic exercise parameters and displays recommendations for comorbidity adaptation. These basic parameters include exercise frequency, intensity, duration, type, and volume. Specifically, the exercise intensity calculation algorithm first quantifies the exercise volume based on the actual number of steps in a 6-minute walk test and a safety adjustment coefficient. Then, it adapts specific exercise methods to the patient's comorbidities, such as adding resistance training if diabetes is present.

[0086] The intensity of the exercise is calculated using the following method:

[0087] ;

[0088] Where I represents exercise intensity; 6MWD represents the distance walked in 6 minutes;

[0089] The quantification model for the amount of exercise is as follows:

[0090] ;

[0091] in, Indicates the amount of exercise; Indicates exercise intensity; Indicates the total duration of the exercise; This indicates the actual number of steps taken in a 6-minute walking test. Safety Factor This indicates a safety adjustment factor, such as 0.8 to 1.2, which can be set according to the patient's age or the severity of complications.

[0092] In addition, the system incorporates a static priority rule-based exercise safety conflict detection mechanism. When dynamically optimizing and adjusting the exercise prescription based on an intelligent risk grading system, it intelligently assesses the matching degree of the exercise prescription and automatically checks and adjusts the data parameters of the prescription for threshold conflicts. For example, when the exercise intensity requirement conflicts with the heart rate limit, the upper limit of the heart rate is prioritized. Real-time monitoring via a wristband detects when the heart rate exceeds the safety threshold during exercise, triggering a three-level warning: Level 1 is an app pop-up reminder, Level 2 is automatic intensity reduction, and Level 3 is emergency pause and notification to the doctor. The phased adjustment of exercise prescription parameters is set according to Table 2, allowing administrators to dynamically optimize the prescription content based on the patient's actual exercise situation.

[0093] Table 2. Adjustment of Exercise Prescription Parameters

[0094]

[0095] During the implementation of the exercise prescription, a dynamic target achievement algorithm is used to calculate the daily exercise target achievement rate based on the exercise prescription and real-time acquired biomonitoring data, and the patient's exercise results are fed back in real time; wherein, the formula for calculating the daily exercise target achievement rate is:

[0096] Daily exercise target achievement rate = Σ (Daily completion amount × Time decay coefficient) / Exercise prescription requirement; Σ represents the sum of all exercise completed on that day;

[0097] Time decay coefficient = 1 / (1 + 0.1 × number of days of delay);

[0098] The exercise results are fed back through the Vascular Health Index (VHI) score, calculated using the following formula:

[0099] VHI = 0.3 × 6MWD progress rate + 0.2 × gait symmetry improvement + 0.5 × prescription compliance;

[0100] Among them, the 6MWD progress rate represents the results of two 6-minute walking distance tests before and after treatment, i.e., (6MWD after treatment - 6MWD before treatment) / 6MWD before treatment × 100%; the gait symmetry improvement rate represents the degree of improvement in the patient's gait symmetry, which is obtained by comparing the difference in gait between the left and right sides of the patient before and after treatment; the prescription compliance rate represents the degree to which the patient follows the exercise prescription issued by the doctor, i.e., (actual number of exercise sessions / number of exercise sessions prescribed by the prescription) × 100%.

[0101] The principle of merging applicable items means that when a patient has multiple diseases and special circumstances, identical items from multiple exercise rehabilitation guidance programs that are applicable to the patient's different diseases and special circumstances are merged, while different items appear in parallel; that is, all items in all exercise rehabilitation guidance programs are retained, identical items appear only once, and different items appear in parallel.

[0102] The safety principle means that when different health conditions have different regulations for the same exercise, and a certain disease has clear safety regulations for that exercise, safety should be given priority.

[0103] In addition, the intelligent exercise record dashboard displays the patient's exercise performance in real time, assesses prescription matching, and provides treatment suggestions. The length of the progress bar represents the number of weekly exercises required by the prescription. The color scheme is as follows: all green indicates 100% completion and the intensity meets the standard; yellow and green stripes indicate completion ≥80% but with fluctuating intensity; and a red box warning indicates failure to meet the standard for 3 consecutive days.

[0104] This invention also includes a mechanism for handling abnormal compliance, the specific process of which can be found in [link to documentation]. Figure 2 This is a flowchart illustrating the handling process for abnormal exercise target achievement. When a patient's weekly exercise target achievement rate is abnormal, such as below 80% or above 120%, the system will analyze the cause (e.g., determine the type of exceedance, analyze the distribution of exercise time periods, retrieve heart rate and blood oxygen data, etc.) and pop up a floating prompt box for the administrator to choose to contact the patient, adjust the prescription, or continue observation. For example, when patient Zhang San's weekly target achievement rate is displayed as 72%, clicking the "Prescription?" icon will bring up an analysis window displaying indicators such as exercise volume, exercise frequency, and average intensity. The system suggests increasing the exercise volume, after which the administrator can choose to adjust the prescription. The system will also automatically generate optimization plans for selection, such as video communication with the patient, one-click prescription adjustment, and marking for observation.

[0105] In addition, a data management system can be built using exercise prescriptions and real-time collected patient exercise data to form patient exercise records, which can then be integrated with the hospital's information system. Patient exercise records include:

[0106] (1) Generate a dynamic data dashboard, including: 1) Cumulative exercise days: a circular progress chart (current exercise progress / target exercise progress); 2) Exercise heat map: displaying exercise distribution by hour; 3) Three-dimensional trajectory map: integrating GPS data to generate exercise paths (e.g., displaying the exercise area within the hospital for inpatients); 4) Historical comparison module: sliding to select the time axis, displaying the exercise change curve (red line is the prescription value), heart rate-blood oxygen correlation scatter plot, and the system automatically marking abnormal points (e.g., heart rate > 180 is marked in red).

[0107] (2) Generating intelligent exercise logs: Combining wristband and manual input to form a dual-channel log entry mechanism, it automatically generates standardized medical documents and can synchronize disease progress and nursing records. The specific steps are as follows:

[0108] 1) Collect the following data in real time:

[0109] Basic data (via the fitness tracker): exercise time, distance, steps;

[0110] Intensity metrics (via wristband): real-time speed, metabolic equivalent;

[0111] Physiological parameters: dynamic heart rate (wristband), blood oxygen (wristband), plantar pressure distribution (smart insole), blood pressure (IoT blood pressure monitor) (recorded every 5 minutes).

[0112] 2) Manually supplement data:

[0113] Pain location labeling: Visual analogue scale (VAS) pain rating;

[0114] Subjective fatigue rating: Subjective fatigue Borg score.

[0115] Standardized medical documentation: For hospitalized patients, standard progress notes are automatically generated, such as "The patient completed x sets of walking training today, with an average heart rate of xx bpm, blood pressure fluctuating between xx-xx / xx-xx mmHg, and blood oxygen fluctuation between xx-xx%. When critical values ​​occur during exercise, prompts for appropriate measures and records are provided, such as automatically adding "Oxygen therapy observation" when blood oxygen is <90%. For discharged patients, home rehabilitation reports and suggestions are generated: "This week's target achievement rate is 82%, resistance training is recommended," with outlier values ​​highlighted in yellow.

[0116] This module establishes a three-tiered security system through real-time biomonitoring, dynamic prescription optimization, and multi-terminal collaboration to monitor patients' exercise health. Details are as follows:

[0117] (1) Real-time biological monitoring

[0118] 1) Dynamic threshold monitoring: The baseline threshold is automatically generated based on the patient's records (e.g., heart rate max = 220 - age); the threshold is adjusted in real time: it automatically fluctuates up or down by 5% based on the blood pressure on the day.

[0119] 2) Anomalies in inpatient data trigger a three-tiered response mechanism: screen flashing alert, video and voice warning paused, and automatic call to the responsible nurse. See details for the process. Figure 3 .

[0120] (2) Dynamic optimization of prescriptions:

[0121] 1) Generate a monthly "Prescription Compatibility Report": including retention rate analysis and statistics on prescription clause implementation;

[0122] Progress curve: such as showing the improving trend of key indicators;

[0123] Intelligent suggestions: If there is significant improvement or stagnation, it is recommended to advance to the next stage or change the training mode.

[0124] 2) Prescription confirmation process: Major adjustments require dual review (doctor + rehabilitation therapist / nurse); historical versions can be traced and compared.

[0125] (3) When data anomalies exceed the threshold during the movement, a multi-terminal collaborative early warning will be activated, as follows:

[0126] 1) Patient side: The fitness tracker vibrates and gives a voice announcement, such as "Your current heart rate is too fast, please slow down immediately";

[0127] 2) Administrator side: A pop-up window displays the patient's real-time movement path and automatically links to the electronic medical record to display key information such as allergy history;

[0128] 3) Family member's side: The APP pushes a concise warning, such as "Zhang San's current blood pressure is 165 / 100 mmHg".

[0129] The intelligent push module for propaganda and education content consists of two parts: a propaganda and education knowledge base and personalized propaganda and education paths.

[0130] Collect multi-dimensional characteristics of patients, organize educational content on PAD-related diseases, diet, medication, exercise, lifestyle habits and surgery, etc., and upload it to the knowledge base. Administrators can manage this content through the upload, add, edit and delete functions.

[0131] A personalized education pathway generation method constructs a patient profile based on collected multi-dimensional features. This profile includes disease diagnosis results, clinical data, medication lists, surgical records, exercise prescriptions, and patient lifestyle habits. When generating initial education content, the system annotates each piece of content in the education knowledge base with multi-dimensional tags and establishes a knowledge vector matrix. An improved KL divergence algorithm is then used to select the education content with the highest relevance to the current time point.

[0132] The formula for calculating the matching degree is:

[0133] ;

[0134] in, Indicates the degree of matching; This indicates the timeliness of knowledge at the current time point t. , For decay rate, Indicates the interval between the knowledge publication date and the current time; This represents the total number of feature dimensions in a patient profile. Representing feature dimension variables, such as This corresponds to the disease diagnosis results in the patient feature profile. Represents the weight coefficient of the current feature dimension; Represents the patient feature vector; This represents the i-th knowledge entry; A patient feature vector representing the feature dimension; Knowledge entries representing feature dimensions; Indicates intersection; Represents the union; This indicates that a numerical value is being retrieved.

[0135] The initial path generation uses a priority ranking algorithm. If multiple pieces of content have similar Key-Level Distance (KL) values, the content with the higher completion rate is prioritized. For example, educational content uses mandatory relevance items: content directly related to diagnosis / surgery (such as "PAD nursing points"); high-risk items: content related to current abnormal indicators (such as pushing low-fat diet guidelines when blood lipid-related indicators are elevated); and behavioral intervention items: content matching unhealthy lifestyle habits (such as automatically adding the "Tobacco-related damage to blood vessels" module for smoking patients).

[0136] like Figure 4 As shown, a tree-like push path is generated.

[0137] In addition, the system is equipped with key monitoring node triggers, such as time nodes like the preoperative preparation period, postoperative acute period, in-hospital rehabilitation period, and home rehabilitation period, as well as event nodes like decreased steps, abnormal blood sugar, and decreased ABI, dynamically adjusting the knowledge weight of content and optimizing the push strategy. A time decay mechanism is used to dynamically adjust the knowledge weight of the initial educational content, expressed as:

[0138] ;

[0139] in, The knowledge weights of the updated initial missionary content; The knowledge weight of the initial educational content before the update; n is the number of push cycles; Indicates periodic decay; Indicates continuous decay; This represents the number of days since the last valid read. This is the forgetting factor, for example, set to 0.05; The event correlation degree for the i-th push period The indicator function, such as when an event is detected (ABI value decreases > 0.15), Otherwise, it is 0.

[0140] Analyze patients' real-time reading behavior data of the initial educational content, calculate the similarity of patients' reading behavior, and generate a patient recommendation list. The expression is:

[0141] ;

[0142] ;

[0143] in, This indicates the similarity in reading behavior between patient u and patient v; This represents the set of numbers of shared push cycles for patients u and v. This indicates the patient u's reading status of the educational content in the i-th push cycle. A value of 1 indicates a reading time of ≥30s, representing in-depth reading; a value of 0.5 indicates a reading time of <30s, representing fast reading; and a value of 0 indicates that the content was not opened. This indicates the patient v's reading status of the educational content in the i-th push cycle; This represents the average reading time of patient u across all push notification cycles for educational content. This represents the average reading time of patient v across all push notification cycles for educational content;

[0144] Indicates the patient Recommended list; This represents a set of patients with similar reading behaviors to patient u. This indicates the patient v's reading status of the educational content in the j-th push cycle;

[0145] Based on the multi-dimensional characteristics of patients collected in real time, when abnormal indicators of patients are detected, content related to the abnormal indicators is searched from the health education knowledge base, inserted at the top of the patient recommendation list, and an alert is sent to the medical staff.

[0146] For handling unread content, the system has a tiered reminder mechanism. When a patient has more than two unread items and the number of unread days is between 1 and 3, a level 1 reminder is triggered, which means that the app will automatically push a message at 09:00 every day; a text message reminder template will be sent on the 3rd day of unread content; and the intelligent outbound calling system will be activated on the 7th day of unread content. After a phone call, if the status is updated to "confirmed," it is marked as read; if the call is not connected, it is added to the next call queue; if the call is refused, it will be transferred to a human for follow-up.

[0147] The quality control closed-loop management module constructs a standardized quality control indicator system, generates quality control data, and provides early warnings and improvements based on the processing and feedback from the rehabilitation assessment module, the exercise prescription generation and dynamic adjustment module, and the intelligent push module for educational content.

[0148] This embodiment constructs a standardized quality control indicator system across eight dimensions to comprehensively ensure the timeliness, effectiveness, accuracy, and scientific rigor of PAD patient rehabilitation. As shown in Table 3, the standardized quality control indicator system across eight dimensions includes: timely rehabilitation assessment rate, timely patient enrollment rate, exercise prescription matching degree, patient exercise effectiveness rate, health education delivery implementation rate, timely health education notification rate, and health education reading rate and effective reading rate.

[0149] Table 3. Eight-Dimensional Quality Control Indicator Model

[0150]

[0151] The system also visualizes quality control data through heatmaps, trend comparison charts, and quality control analysis and improvement reports. The heatmap displays the completion status of eight indicators and generates red-yellow-green alerts, as shown in Table 4. The trend comparison chart supports historical data comparison, and the quality control analysis report is generated monthly and supports filtering by indicator category.

[0152] Table 4 Quality Control Three-Color Early Warning Processing Flowchart

[0153]

[0154] In terms of quality improvement, a closed-loop built-in PDCA (Plan-Do-Check-Act) cycle processing mechanism is adopted. The system automatically generates a "Quality Control Analysis Report," recommends standardized improvement processes, and tracks the implementation progress. When indicators deviate from their targets for multiple consecutive cycles, the system automatically calls upon the improvement measures knowledge base to recommend solutions and tracks the effectiveness verification. The PDCA cycle processing mechanism process is as follows:

[0155] 1) Problem discovery: The system automatically generates a "Quality Control Analysis Report" (including a root cause fishbone diagram).

[0156] 2) Improvement plan: Built-in knowledge base of improvement measures (including 50+ standardized improvement processes).

[0157] 3) Execution tracking: Establish and improve the task tracking dashboard (display progress in Gantt chart mode).

[0158] 4) Effect verification: The effect evaluation algorithm is automatically triggered after the improvement cycle ends.

[0159] The initial interface of the app adopts a layered and progressive architecture, mainly including an intelligent rehabilitation module, an adaptive education module, and an intelligent safety monitoring module. The intelligent rehabilitation module uses a "layered and progressive - dynamically optimized" intelligent prescription engine. Exercise programs are dynamically demonstrated using a 3D human body model, and key muscle groups are displayed with heatmaps showing activation levels. Each movement is labeled with multi-dimensional tags, such as indications, metabolic equivalents, and contraindications. Real-time biofeedback training collects data from wristbands and smart insoles via IoT devices, establishing a dynamic threshold model and triggering a three-level response in case of anomalies, such as... Figure 5 As shown. Intelligent progress management uses a dynamic achievement algorithm to calculate the achievement rate and establishes a vascular health index scoring system to incentivize patient compliance. A multimodal exercise log constructs a three-in-one recording system of "device acquisition - manual supplementation - intelligent analysis." The automated acquisition layer integrates GPS and indoor positioning data to mark exercise hotspot areas, and the dynamic curves of physiological parameters record changes in heart rate, blood oxygen, and foot pressure every 5 seconds. Manually supplemented data includes marking pain sites and manually entering daily blood pressure. The intelligent analysis report generates a daily "Exercise and Health Briefing" and correlates the exercise data from the previous 24 hours to generate a causal analysis tree.

[0160] The adaptive education module uses a "feature clustering-dynamic optimization" recommendation algorithm to build a six-dimensional feature vector to construct a patient profile. A real-time update mechanism triggers profile reconstruction when the ABI value changes by more than 0.15. The content matching algorithm applies improved KL divergence for knowledge matching. The interactive learning system improves patient learning outcomes through scenario simulation tests and knowledge mastery assessments. The multi-channel reminder system constructs a three-level intelligent reminder strategy: when there are more than 2 unread items and the unread duration is less than 3 days, an app push notification is triggered; 3 to 7 days of unread items trigger an SMS reminder; and more than 7 days of unread items triggers an intelligent outbound call and family member interaction. Emergency content is prioritized; for example, when the system detects a systolic blood pressure greater than 180 mmHg, it automatically pushes content on "Hypertensive Emergency Treatment." The AI ​​image recognition system analyzes the wound photos uploaded by the patient and automatically matches wound care guidelines.

[0161] The intelligent safety monitoring module constructs a three-tiered protection network of "device-algorithm-human intervention." A dynamic threshold system adjusts the upper limit of heart rate based on the morning blood pressure value, while a foot pressure feedback system generates a real-time plantar pressure cloud map and triggers gait correction suggestions when the unilateral pressure difference exceeds 15%. A multi-parameter correlation early warning system establishes a correlation matrix between blood pressure, blood oxygen, and gait; a red fall warning is triggered when "blood pressure decreases + blood oxygen decreases + gait symmetry decreases," and an LSTM model predicts fall risk trends, issuing warnings 30 minutes in advance. Remote rehabilitation monitoring supports voice interaction for querying exercise data through a virtual rehabilitation assistant, and an intelligent question-and-answer system analyzes common rehabilitation questions based on a medical knowledge graph. A family-hospital collaborative monitoring dashboard is established, automatically pushing concise reports when key indicators are abnormal, with human consultation serving as a supplementary means to answer complex questions.

[0162] Experiments showed that after using the system of this invention, the incidence of adverse motor events in patients with orange risk decreased from 12.7% to 4.3%; and the response time for identifying high-risk patients was shortened from an average of 4.2 hours to 11 minutes.

[0163] In summary, compared with existing technologies, the full-process PAD patient rehabilitation management system proposed in this invention achieves a revolution in PAD patient rehabilitation management through multimodal data fusion and an intelligent decision engine. Its beneficial effects are reflected in the following aspects:

[0164] (1) Intelligent integration of multi-source data and dynamic risk management

[0165] 1) Cross-system data deep analysis: Based on the FHIR standard and NLP unstructured text parsing technology, real-time integration of HIS / NIS system data is achieved, breaking through the limitations of traditional structured coding and accurately identifying patients with ambiguous diagnoses such as "suspected PAD". Through an intelligent risk grading system, multi-dimensional parameters such as Fontaine staging, ABI index, and contraindications are dynamically integrated to establish a minute-level risk scanning mechanism, improving the response speed for identifying extremely high-risk patients and reducing the false alarm rate.

[0166] 2) Intelligent generation and safety protection of exercise prescriptions: The three-layer architecture of "multimodal data input - tagged action library matching - conflict detection" is adopted to realize the accurate calculation of exercise prescription parameters. Combined with dynamic threshold model and three-level collaborative early warning network, the incidence of exercise-related adverse events is reduced.

[0167] (2) Intelligent decision-making and closed-loop quality control throughout the entire process

[0168] 1) Adaptive Rehabilitation Path Optimization: Based on an improved KL divergence algorithm, a deep matching model is constructed between patient feature vectors and the educational knowledge base. This enables the timely decay of educational content delivery and event-triggered updates (such as automatic triggering of nursing guide delivery for new wounds), improving the reading completion rate. During the execution of the exercise prescription, biometric data such as foot pressure and gait symmetry are collected in real time through IoT devices. The achievement rate algorithm is dynamically optimized to achieve precise matching between exercise intensity and physiological state.

[0169] 2) Eight-dimensional quality control engine drives management upgrade: Construct a quality control heatmap covering indicators such as assessment timeliness, prescription matching degree, and reading completion rate. Combine the PDCA cycle to achieve automatic problem identification (such as fishbone diagram root cause analysis triggered by continuous target abnormality) and embed standardized improvement processes to improve the scientific nature of rehabilitation program adjustment and reduce the manpower cost of quality control management.

[0170] (3) Multimodal interaction and positive incentives for patient behavior

[0171] 1) Contextualized Intelligent Reminder System: Breaking through the traditional time-series reminder model, this system establishes multi-dimensional event triggers such as abnormal physiological indicators and decreased behavioral compliance, enabling priority delivery of emergency educational content and linked early warnings for family members. Through a 3D motion visualization model and VHI score feedback, it enhances patients' understanding of rehabilitation goals.

[0172] 2) Enhanced Full-Cycle Rehabilitation Efficiency: Based on real-time data such as foot pressure cloud maps from smart insoles and physiological curves from wristbands, a motion-physiology correlation analysis model is constructed to improve the accuracy of individualized exercise prescriptions. Through closed-loop quality control and multi-terminal collaboration, the patient's disease status is improved, the incidence of postoperative complications is reduced, and the overall rehabilitation cycle is shortened.

[0173] (4) Optimization of medical resources and evidence-based decision support

[0174] 1) Data-driven resource reallocation: Through automated risk classification and early warning freezing mechanisms, medical resources are tilted toward high-risk patients, improving the efficiency of handling critical values ​​in rehabilitation.

[0175] 2) Construction of a full-process management system: From prescription generation (algorithm formula floating prompts) to quality control improvement (automatic generation of standardized medical documents), all operations retain a traceable data chain, providing high-value datasets for clinical research and helping to dynamically optimize the rehabilitation of PAD patients and practice precision medicine.

[0176] In summary, this invention constructs a management system covering the entire rehabilitation process for PAD patients through multimodal data fusion and intelligent algorithm collaboration, which significantly improves the accuracy and safety of rehabilitation management and solves the pain points of existing technologies such as low data management efficiency, insufficient intelligence of exercise prescriptions, low degree of personalization of education and promotion, and imperfect quality control system.

[0177] In the several embodiments provided in this invention, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus and method embodiments described above are merely illustrative. For example, the flowcharts in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of the invention. In this regard, each block in the flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0178] In addition, the functional modules in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0179] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, electronic device, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks. It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. In the absence of further restrictions, an element defined by the phrase "comprising a..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0180] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0181] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0182] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0183] The use of "first" and "second" in the embodiments is merely to distinguish similar objects and does not represent a specific ordering of objects. It is understood that "first" and "second" can be interchanged in a specific order or sequence where permitted. It should be understood that the objects distinguished by "first" and "second" can be interchanged where appropriate so that the embodiments described herein can be implemented in an order other than those illustrated or described herein.

[0184] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A full-process-based peripheral arterial disease patient rehabilitation management system, characterized in that, The application comprises a patient screening and risk grading module, a rehabilitation evaluation module, a sports prescription generation and dynamic adjustment module, an education content intelligent pushing module, and a quality control closed-loop management module. The patient screening and risk grading module is connected with a hospital information system through a medical data interface to obtain basic data of patients with peripheral arterial disease in real time for analysis, and an intelligent risk grading system is constructed based on Fontaine staging, ankle-brachial index, and contraindications to automatically update the risk level of the patients and trigger an early warning. The rehabilitation evaluation module generates an evaluation report with time as the axis based on the basic data of the patients and real-time biological monitoring data. The sports prescription generation and dynamic adjustment module generates a sports prescription based on the evaluation report, comorbidities, and real-time biological monitoring data, and dynamically optimizes and adjusts the sports prescription based on the intelligent risk grading system. The sports prescription includes patient sports basic parameters, and displays comorbidity adaptation suggestions. The sports intensity is calculated by the following method: ； wherein I represents the sports intensity, and 6MWD represents the 6-minute walking distance. The sports amount is quantified by the following model: ； wherein, represents the amount of exercise; I represents the intensity of exercise; represents the total duration of exercise; represents the actual number of steps in the 6-minute walk test; SafetyFactor represents the safety adjustment factor; The education content intelligent pushing module collects multi-dimensional characteristics of the patients to construct a patient characteristic portrait, establishes an education knowledge base, generates initial education content, and dynamically adjusts and individually adapts and pushes the initial education content. The quality control closed-loop management module constructs a standardized quality control index system based on the processing and feedback of the rehabilitation evaluation module, the sports prescription generation and dynamic adjustment module, and the education content intelligent pushing module, generates quality control data, and provides early warning and improvement.

2. The system for rehabilitation management of peripheral arterial disease patient based on whole process according to claim 1, characterized in that The basic data of the patients includes basic information, diagnosis codes, and test reports of the patients. If the diagnosis codes are fuzzy diagnoses of suspected peripheral arterial disease, the basic data of the patients are analyzed and key disease characteristics are extracted by using NLP technology, and the test reports of the patients are combined to dynamically correct the screening results of the patients. 3.The whole-process-based peripheral arterial disease patient rehabilitation management system according to claim 1, characterized in that The biological monitoring data includes exercise time, distance, steps, real-time speed, metabolic equivalent, dynamic heart rate, blood oxygen, plantar pressure distribution, and blood pressure. wherein BMI = weight / height 2 ; The content of the evaluation report includes evaluation patients, evaluation dates, height, weight, BMI, drugs used, blood glucose, blood lipids, blood pressure, whether smoking, whether drinking, exercise habits, wound conditions, pain visual analogue scale VAS, lower limb muscle strength, simple physical condition SPPB, 6-minute walking distance 6MWD and steps, claudication distance, maximum walking distance, ankle-brachial index ABI, leg circumference, quality of life scale, walking impairment questionnaire, hope scale, and subjective exertion score. The type of 6MWD is divided into male and female, and the male 6MWD = 1140-5.61×BMI-6.94×age, and the female 6MWD = 1017-6.24×BMI-5.83×age, with age in years.

4. The system for rehabilitation management of peripheral arterial disease patients based on whole process according to claim 1, characterized in that The matching degree of the exercise prescription is intelligently judged, and the data parameters of the exercise prescription are automatically checked and adjusted in threshold conflict.

5. The system for peripheral arterial disease patient rehabilitation management based on whole process according to claim 1, characterized in that The patient feature image includes disease diagnosis results, clinical data, a medication list, surgical records, exercise prescriptions, and patient living habits. When generating the initial health education content, each piece of content in the health education knowledge base is labeled with multi-dimensional tags, and content matching is performed based on KL divergence to select the health education content with the highest matching degree at the current time point; the matching degree is calculated according to the following formula: ； wherein, denotes the matching degree; denotes the knowledge timeliness of the current time point t, , is the decay rate, denotes the interval between the knowledge publishing date and the current time point; denotes the total number of feature dimensions of the patient feature portrait; denotes the feature dimension variable; denotes the weight coefficient of the current feature dimension; denotes the patient feature vector; denotes the i-th knowledge item; denotes the patient feature vector of the feature dimension; denotes the knowledge item of the feature dimension; denotes the intersection; denotes the union; denotes the taking value; When the matching degrees of the selected health education content are equal, the content with a high reading completion rate is preferentially selected as the initial health education content.

6. The system for peripheral arterial disease patient rehabilitation management based on whole process according to claim 1, characterized in that The initial health education content is dynamically adjusted and individually and adaptively pushed, specifically as follows: Real-time multi-dimensional features of the patient are collected, and the health education knowledge base is updated; According to the updated health education knowledge base, the knowledge weight of the initial health education content is dynamically adjusted using a time decay mechanism, and the expression is as follows: ； wherein, is the knowledge weight of the updated initial propaganda content; is the knowledge weight of the initial propaganda content before updating;n is the number of push cycles; is the number of days from the last effective reading; represents the periodic decay; represents the continuous decay; is the forgetting factor; is the event correlation degree of the ith push cycle is the indicator function. Real-time reading behavior data of the patient on the initial health education content are analyzed, and the reading behavior similarity of the patient is calculated to generate a patient recommendation list, and the expression is as follows: ； ； wherein, represents the reading behavior similarity of patient u and patient v; represents the number set of common push cycle of patient u and patient v; represents the reading situation of patient u to the i-th push cycle health education content, the value of 1 represents reading duration ≥ 30s, indicating deep reading; the value of 0.5 represents reading duration < 30s, indicating fast reading; the value of 0 represents that the content is not opened; represents the reading situation of patient v to the i-th push cycle health education content; represents the average value of reading duration of patient u to all push cycle health education content; represents the average value of reading duration of patient v to all push cycle health education content; a recommended list of patients; a recommended list of patients; a set of patients similar to patient u in reading behavior; reading of patient v on the jth push cycle health education content; According to the real-time multi-dimensional features of the patient, when an abnormal index of the patient is detected, content related to the abnormal index is searched from the health education knowledge base, inserted at the top of the patient recommendation list, and an early warning is sent to the medical side.

7. The system for peripheral arterial disease patient rehabilitation management based on whole process according to claim 1, characterized in that Further, when generating the exercise prescription, a dynamic threshold model is established and real-time feedback training is performed according to real-time biological monitoring data; the dynamic threshold model includes a maximum heart rate threshold, and the expression is as follows: Maximum heart rate threshold = 220 - patient age ± blood pressure fluctuation value on the current day.

8. The system for peripheral arterial disease patient rehabilitation management based on whole process according to claim 1, characterized in that Further, during the execution of the exercise prescription, a daily exercise compliance rate is calculated using a dynamic compliance algorithm according to the exercise prescription and real-time biological monitoring data, and the exercise results of the patient are fed back in real time; the calculation formula of the daily exercise compliance rate is as follows: Daily exercise compliance rate = Σ (daily completed amount × time decay coefficient) / exercise prescription required amount; Σ represents the sum of all exercise completed amounts on the current day; Time decay coefficient = 1 / (1 + 0.1 × delay days); The exercise results are fed back through the vascular health index (VHI) integral, and the calculation formula is as follows: VHI = 0.3 × 6MWD progress rate + 0.2 × gait symmetry improvement degree + 0.5 × prescription adherence degree; Wherein, the 6MWD progress rate represents the test results of the patient's 6-minute walk distance before and after treatment, i.e. (6MWD after treatment - 6MWD before treatment) / 6MWD before treatment × 100%; the gait symmetry improvement degree represents the improvement degree of the patient's gait symmetry, which is obtained by comparing the difference between the patient's left and right gait before and after treatment; the prescription adherence degree represents the degree of adherence of the patient to the exercise prescription prescribed by the doctor, i.e. (actual number of exercises performed / number of exercises prescribed by the prescription) × 100%.

9. The system for peripheral arterial disease patient rehabilitation management based on whole process according to claim 1, characterized in that The standardized quality control index system comprises a rehabilitation evaluation timely rate, a patient group timely rate, a movement prescription matching degree, a patient movement effective rate, a propaganda and education pushing implementation rate, a propaganda and education notification timely rate, a health education reading rate and an effective reading rate. The calculation method of the rehabilitation evaluation timely rate is (patients completing evaluation on time / total patients that should be evaluated) * 100%, and completing evaluation on time is defined as an evaluation date <= a preset date + 1 working day. The calculation method of the patient group timely rate is (patients with a system group time <= 24 hours after hospitalization / total new group patients) * 100%. The calculation method of the movement prescription matching degree is (a target rate * 80%-120% * a patient number / system group patient number) * 100%, and the target rate = (a patient actual movement amount / movement prescription requirement amount) * 100%. The calculation method of the patient movement effective rate is a main index improvement rate = (a current 6MWD-baseline 6MWD) / baseline 6MWD * 100%. The calculation method of the propaganda and education pushing implementation rate is (a patient number actually pushed by propaganda and education content / total patients that should be pushed) * 100%. The calculation method of the propaganda and education notification timely rate is (a patient number completing propaganda and education content notification within 3 days / total unread patient numbers) * 100%. The calculation method of the health education reading rate is (a patient number reading health education content >= 1 time / total received patients) * 100%. The calculation method of the effective reading rate is (a deep reading patient number / reading patient number) * 100%, and deep reading is defined as single stay >= 120 seconds and completion of a knowledge test.

Images