Evaluation system for patients with exercise fear

Abstract

The invention discloses an assessment system for a patient with exercise fear, and relates to the technical field of medical health information. In order to solve the problems that postoperative exercise fears of some patients cannot be recognized in time, and the patients have different degrees of exercise fears after operation, so that thrombosis is easily induced, and quantification cannot be realized. Through triangular verification of physiological stress, behavior avoidance and psychological anxiety, the problem that a traditional scale is too high in subjectivity or insufficient in physiological monitoring specificity is solved, the evaluation accuracy is greatly improved, different rehabilitation stages can be automatically adapted based on a threshold adjustment mechanism of postoperative time and individual differences, misjudgment caused by unified evaluation is avoided, and the evaluation accuracy is improved. The method is suitable for personalized management of old people and patients with complex surgeries, personalized evaluation for the patients is carried out, intervention deviation caused by clinical misjudgment is reduced through multi-modal correlation analysis, passive diagnosis is turned into active prevention, and the postoperative thrombus incidence rate is greatly reduced through early recognition of the patients with movement fear.

Description

Technical Field

[0001] The present invention relates to the field of medical health information technology, and in particular to a patient assessment system for kinesia. Background Art

[0002] Kinesiophobia refers to the patient's excessive and irrational fear of movement, which leads to avoiding daily activities or physical exercise to avoid re-injury or re-damage to the body. When patients develop kinesiophobia, it is contrary to the requirements of rapid surgical recovery care, and reduced activity can easily induce thrombosis. In addition, kinesiophobia leads to increased negative emotions in patients, seriously affecting their quality of life. Traditional assessment methods mostly rely on subjective reports from patients or single physiological indicators, and patients have varying degrees of kinesiophobia after surgery, making it difficult to fully capture the complex relationship between physiological stress, behavioral avoidance, and psychological anxiety; the lack of dynamic quantitative assessment of kinesiophobia makes it impossible to adapt to individual differences and changes in rehabilitation stages, resulting in rough risk grading and insufficiently targeted intervention strategies. It is easy to miss the opportunity for early intervention, and there is a significant information asymmetry between doctors and patients, which affects rehabilitation efficiency. Summary of the Invention

[0003] The purpose of the present invention is to provide a patient assessment system for movement phobia. By integrating physiological, behavioral and psychological multidimensional data collected in real time by multi-source equipment, a full-process intelligent management system including feature extraction, dynamic risk assessment, abnormality identification and warning, and personalized intervention is constructed to achieve accurate quantification of the patient's movement phobia state, real-time warning and targeted rehabilitation management, improve clinical assessment efficiency and rehabilitation treatment effect, and solve the problems raised in the above background technology.

[0004] To achieve the above object, the present invention provides the following technical solutions: A patient assessment system for fear of movement, comprising: a kinesiophobia feature extraction module configured to acquire multidimensional data of the patient in real time based on multiple source devices, construct and process a multidimensional input data set, and extract specific features representing kinesiophobia; a dynamic risk assessment module configured to input the extracted specific features into a constructed kinesia quantification model to calculate a kinesia index, and to assess the patient's kinesia risk level based on the calculated result; An abnormality identification and early warning module is configured to identify abnormal behavior sequences and persistent avoidance patterns associated with a patient's fear of movement based on time series analysis, and to perform correlation assessments of specific features of each category based on the identification results to determine the severity of the patient's abnormality and issue an early warning; The patient assessment decision module is configured to integrate the patient's movement phobia risk level and abnormality severity assessment results, generate a comprehensive movement phobia assessment confidence level, and form a personalized rehabilitation plan based on the corresponding intervention strategy matched with the patient's basic data.

[0005] Furthermore, the motion fear feature extraction module includes: a multi-dimensional data acquisition unit configured to acquire the patient's physiological data based on the wearable device, acquire the patient's corresponding behavioral data, and simultaneously acquire the patient's psychological state data actively input; a data processing unit configured to clean, calibrate, and standardize the physiological data, behavioral data, and psychological data acquired by the multidimensional data acquisition unit to form a structured multidimensional input data set; The feature extraction unit is configured to extract specific features related to physiological stress, behavioral avoidance and psychological anxiety from a multi-dimensional input data set based on a feature extraction model, and generate a physiological stress feature subset, a behavioral avoidance feature subset and a psychological anxiety feature subset.

[0006] Furthermore, the multidimensional data acquisition unit specifically includes: Deploy a network of wearable device sensors to collect physiological data from patients; Collect the patient's motion parameter data in real time from the rehabilitation training equipment. At the same time, obtain the patient's daily activity trajectory and training environment interaction data to generate the patient's behavioral data. The patient's voice data is acquired based on voice interaction, the voice data is converted into text and subjected to sentiment analysis to generate the patient's psychological data.

[0007] Furthermore, the feature extraction model specifically includes: Extract data field labels from multi-dimensional input data sets, match data types in the multi-dimensional input data according to a preset dictionary, and match corresponding standard feature templates from a preset feature sample library based on the identified data types; Generate validity constraints for data verification based on feature sample templates, filter and verify each data point in the multi-dimensional input data based on the validity constraints, and mark data points that do not meet the validity constraints as marginal data; Extract data other than marginal data from the multi-dimensional input data set as valid feature data, generate a valid feature candidate set, and extract valid feature classification labels for physiological stress, behavioral avoidance, and psychological anxiety dimensions from the valid feature candidate set; A feature extraction model for multi-dimensional input data sets is constructed based on effective feature classification labels. Features are extracted and classified based on the feature extraction model, and structured feature subsets are generated and standardized.

[0008] Furthermore, the feature extraction unit further includes: Physiological stress feature subset: performing time domain, frequency domain and time-frequency analysis on the physiological data to extract the characteristics of exercise-induced stress response and sustained stress state; Behavioral avoidance feature subset: Calculate rehabilitation training compliance index, daily activity attenuation index and temporal pattern characteristics based on the behavioral data; Psychological anxiety feature subset: Perform scale score analysis and natural language processing on the psychological data to extract psychological scale-derived features and semantic emotional features.

[0009] Furthermore, the dynamic risk assessment module includes: a model building unit configured to build a corresponding sports fear quantification model based on the feature categories of the physiological stress feature subset, the behavioral avoidance feature subset, and the psychological anxiety feature subset; a grading strategy formulation unit configured to establish risk grading standards and determine exercise fear index thresholds for physiological stress characteristics, behavioral avoidance characteristics, and psychological anxiety characteristics at each level; a real-time processing unit configured to input the multi-dimensional input data set collected in real time into a movement phobia quantification model to calculate a movement phobia index of the patient, compare the movement phobia index with a movement phobia index threshold, and determine a movement phobia risk level of the patient; The dynamic threshold adjustment unit is configured to obtain and analyze the patient's basic data, and adjust the movement fear index threshold and weight coefficient of each specific feature according to the individual differences of the patient and the postoperative stage.

[0010] Furthermore, the abnormality identification and warning module includes: a time series pattern recognition unit configured to extract time series data from the behavioral data, identify typical abnormal behavior sequences associated with motor phobia, and obtain abnormal behavior pattern features; a persistent avoidance pattern recognition unit configured to extract a feature subset time series from the behavioral avoidance feature subset, identify a potential persistent avoidance pattern based on the feature subset time series, and obtain persistent avoidance pattern features; a multimodal association assessment unit configured to perform association analysis on the identified abnormal behavior sequences and persistent avoidance patterns with a subset of physiological stress characteristics and a subset of psychological anxiety characteristics, assess the severity of the patient's abnormality, and generate corresponding warning information based on the assessed severity of the abnormality; The abnormality classification warning unit is configured with a preset three-level abnormality severity standard, combining the risk level of sports phobia and the posterior probability of abnormal behavior to trigger the corresponding warning mechanism; The warning information generation unit is configured to generate structured warning information for different levels of abnormalities and synchronize the warning information between the medical and patient ends.

[0011] Furthermore, the patient assessment and decision-making module also includes: visual display of assessment results and intervention progress based on the smart terminals on the medical and patient sides. Among them, the visual display of the smart terminal on the medical side includes: risk heat map, trend comparison chart and intervention tracking table; the visual display of the smart terminal on the patient side includes: progress dashboard, fear diary and education database.

[0012] Furthermore, based on the feature extraction model, specific features related to physiological stress, behavioral avoidance, and psychological anxiety are extracted from the multi-dimensional input dataset, including: Determine the motion state characteristics of each item of data in the multidimensional data set based on the feature extraction model, and determine the patient's response form to each item of data according to the motion state characteristics, wherein the response form includes: physiological response, behavioral response and psychological response; Classify each data item according to the patient's response form to each data item, obtain the classification results, and determine the unified state description parameters of each type of data based on the classification results; Determine the unified adaptability influencing factors of each type of data on the patient based on the unified state description parameters, and determine the impact feedback intensity of each type of data on the patient based on the unified adaptability influencing factors; Determine the average stress state index of each type of data for the patient based on the impact feedback intensity, and determine the quantitative division of the observed events based on the average stress state index and the descriptive parameters of each type of data; Based on the observed event division, each type of data is quantitatively divided into events to generate an observed event sequence, and the observed event sequence is decomposed into independent event components; Obtaining an event completion goal and an event process indicator for each independent event, and determining a physiological response state description parameter, a behavioral response state description parameter, and a psychological response state description parameter for each independent event based on the event completion goal and the event process indicator; Obtaining the adaptability of the subjective morphological characteristics of the physiological response state description parameters, the behavioral response state description parameters, and the psychological response state description parameters to the objective event development characteristics; Determine the psychological burden, behavioral limitations, and physiological resistance of each individual event to the patient based on suitability; Determine the physiological / psychological / behavioral stimulation expression level of each independent event based on the patient's psychological burden, behavioral limitation, and physiological resistance. Filter out target independent events that are greater than or equal to the preset expression level based on the physiological / psychological / behavioral stimulus expression level of each independent event; Specific features related to physiological stress, behavioral avoidance and psychological anxiety are extracted based on the event description of target independent events through the feature extraction model.

[0013] Furthermore, the patient's motion parameter data is collected in real time from the rehabilitation training equipment, and at the same time, the patient's daily activity trajectory and training environment interaction data are obtained to generate the patient's behavioral data, including: Collecting the patient's real-time exercise project parameters from the rehabilitation training equipment, and determining the patient's exercise form based on the real-time exercise project parameters; Determine the associated activity mode and environmental interaction mode based on the patient's movement form, and determine the behavior matching identification index based on the associated activity mode and environmental interaction mode; Obtain the matching factor of the behavioral matching identification indicator. At the same time, use millimeter-level radar and sensors to obtain the patient's daily activity trajectory in each sport and the training environment interaction data; The patient's daily activity trajectory and training environment interaction data are evaluated for sports project matching through behavioral matching identification indicators: in, It is represented by the daily activity trajectory of the i-th sport and the matching degree between the training environment interaction data and the sport. It is expressed as the exercise relaxation of the i-th exercise, It is represented by the patient's daily activity trajectory and the number of training environment interaction data in the i-th sport, j is represented by the j-th daily activity trajectory and training environment interaction data, It is expressed as the maximum matrix factor of the motion observation matrix of the j-th daily activity trajectory and the training environment interaction data, It is expressed as the matching factor of the behavioral matching identification index, ln is expressed as the natural logarithm, It is represented as the sports target identification intensity value of the i-th sports event, It is represented by the patient's daily activity trajectory of the i-th exercise and the exercise feedback intensity value shown in the training environment interaction data; Filter out qualified daily activity trajectory and training environment interaction data sets for each sport based on the matching degree between the daily activity trajectory and training environment interaction data of each sport and the sport; Generate patient behavioral data based on qualified daily activity trajectories and training environment interaction datasets.

[0014] Compared with the prior art, the present invention has the following beneficial effects: Through triangular verification of physiological stress, behavioral avoidance, and psychological anxiety, the problems of excessive subjectivity of traditional scales or insufficient specificity of physiological monitoring are solved, and the assessment accuracy is greatly improved. The threshold adjustment mechanism based on postoperative time and individual differences can automatically adapt to different rehabilitation stages and avoid misjudgments caused by unified assessments. It is suitable for personalized management of elderly and complex surgical patients and conducts personalized assessments for patients. At the same time, through multimodal association analysis, it reduces intervention bias caused by clinical misjudgments, shifts from passive diagnosis to active prevention, and greatly reduces the incidence of postoperative thrombosis through early identification of patients with movement phobia, shortens the average length of hospital stay, and improves compliance with rehabilitation treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a module diagram of the patient assessment system for kinesia according to the present invention. DETAILED DESCRIPTION

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] In order to solve the technical problem of not being able to identify some patients with postoperative fear of movement in time, and patients with different degrees of fear of movement after surgery, which is easy to induce thrombosis and cannot be quantified, please refer to Figure 1 , this embodiment provides the following technical solutions: A patient assessment system for fear of movement, comprising: A kinesiophobia feature extraction module is configured to acquire multidimensional patient data in real time based on multiple source devices, including physiological, behavioral, and psychological data related to kinesiophobia, construct and process a multidimensional input dataset, and extract specific features that characterize kinesiophobia, which are categorized into three dimensions: physiological stress, behavioral avoidance, and psychological anxiety. a dynamic risk assessment module configured to input the extracted specific features into a constructed kinesia quantification model to calculate a kinesia index, and to assess the patient's kinesia risk level based on the calculated result; An abnormality identification and early warning module is configured to identify abnormal behavior sequences and persistent avoidance patterns associated with a patient's fear of movement based on time series analysis, and to perform correlation assessments of specific features of each category based on the identification results to determine the severity of the patient's abnormality and issue an early warning; The patient assessment decision module is configured to integrate the patient's movement fear risk level and abnormal severity assessment results, generate a comprehensive movement fear assessment confidence level, and form a personalized rehabilitation plan based on the corresponding intervention strategy matching the assessment level with the patient's basic data. In this embodiment, the patient's basic data includes: basic information such as age, gender, height, weight, surgery type, surgery time, postoperative recovery stage, overall health level, physical condition, daily activity level, exercise habits, etc.; In this embodiment, the intervention strategy includes clinical pathway rules, which are divided into mild risk, moderate risk, and severe risk according to the risk level of kinesia; In this embodiment, the evaluation results and intervention progress are visualized based on the smart terminals on the medical and patient sides. The visualization of the smart terminals on the medical and patient sides includes: Risk heat map: This uses a color gradient to show the contribution of each dimension to the current fear level (e.g., if psychological factors account for 60%, they are highlighted in red); Trend comparison chart: Overlays the weekly change curves of fear index, activity level, and psychological score, and marks key intervention nodes (such as the turning point of index decline after the first psychological intervention); Intervention tracking sheet: Real-time update of the completion status of intervention measures at each stage (such as rehabilitation training execution rate, psychotherapy compliance), and automatic calculation of intervention effect indicators (such as fear index reduction rate = (pre-intervention - post-intervention) / pre-intervention × 100%); The patient-side smart terminal visualization display includes: Progress dashboard: A cartoon-like interface shows the completion of daily activity goals (e.g., "Today's time out of bed: 35 minutes / goal 40 minutes"); Fear Diary: Patients can record their feelings during exercise, and the system automatically generates a correlation analysis chart between mood swings and activity levels (e.g., "When fear of injury arises, activity levels decrease by an average of 40%"); Educational database: Push customized popular science content based on the evaluation results (such as "thrombosis prevention tips" and "safe exercise posture animations" for patients with severe fear).

[0018] In this embodiment, multi-dimensional physiological, behavioral, and psychological data are collected in real time through multi-source equipment. After cleaning and standardization, the feature extraction model is used to generate three feature subsets of physiological stress, behavioral avoidance, and psychological anxiety, so as to achieve a multi-dimensional and accurate characterization of movement phobia. The dynamic risk assessment module constructs a quantitative model based on the feature subset, dynamically adjusts the threshold and weight based on individual differences, calculates the fear index in real time, and assesses the risk level in a graded manner; the abnormal recognition and warning module identifies abnormal behaviors and continuous avoidance patterns through time series analysis and multimodal association, and issues graded warnings to achieve early intervention of risk events; the patient assessment and decision-making module integrates the assessment results to generate a personalized rehabilitation plan, synchronizes information through visual interactive tools on the medical and patient sides, improves the scientific nature of diagnosis and targeted intervention, optimizes the efficiency of doctor-patient collaboration, and provides an intelligent solution for the precise rehabilitation of patients with movement phobia.

[0019] In this embodiment, the motion fear feature extraction module includes: The multi-dimensional data acquisition unit is configured to acquire the patient's physiological data based on the wearable device, collect the patient's corresponding behavioral data, and simultaneously acquire the patient's actively input psychological state data, specifically including: Deploy a wearable device sensor network, including wrist-worn physiological monitoring devices (integrated with PPG heart rate sensors, triaxial accelerometers, galvanic skin response electrodes, etc.) to collect patients' physiological data, including heart rate (HR), heart rate variability (HRV), skin conductance (SC), and body movement intensity (acceleration values) at rest and during exercise; The built-in pressure and angle sensors in rehabilitation training equipment, such as gait trainers and range of motion meters, collect real-time motion parameter data from patients, including range of motion (ROM), muscle exertion duration (e.g., isometric contraction duration), and movement completion (actual number of movements / preset number of movements). At the same time, the patient's daily activity trajectory (e.g., time out of bed, walking speed, gait symmetry) and training environment interaction data (e.g., duration in the training area, frequency of device contact) are collected to generate patient behavioral data. Acquire the patient's voice data based on voice interaction, including the patient's spoken answers to exercise-related questions, convert the voice data into text and perform sentiment analysis to generate the patient's psychological data; a data processing unit configured to clean, calibrate, and standardize the physiological data, behavioral data, and psychological data acquired by the multidimensional data acquisition unit to form a structured multidimensional input data set; The feature extraction unit is configured to extract specific features related to physiological stress, behavioral avoidance and psychological anxiety from a multi-dimensional input data set based on a feature extraction model, and generate a physiological stress feature subset, a behavioral avoidance feature subset and a psychological anxiety feature subset.

[0020] In this embodiment, physiological, behavioral, and psychological three-dimensional data are collected in real time through multi-source devices such as wearable devices, rehabilitation training equipment, and voice interaction. Combined with time domain, frequency domain analysis, and temporal pattern recognition technologies, a structured feature subset is constructed to comprehensively capture the physiological stress response, behavioral avoidance patterns, and psychological anxiety characteristics related to sports phobia, thereby avoiding the one-sidedness of single-dimensional evaluation and significantly improving the accuracy and reliability of the evaluation results.

[0021] In this embodiment, the feature extraction model specifically includes: Extract data field labels from multi-dimensional input data sets, match data types in the multi-dimensional input data according to a preset dictionary, and match corresponding standard feature templates from a preset feature sample library based on the identified data types; Generate validity constraints for data verification based on feature sample templates, filter and verify each data point in the multi-dimensional input data based on the validity constraints, and mark data points that do not meet the validity constraints as marginal data; Extract data other than marginal data from the multi-dimensional input data set as valid feature data, generate a valid feature candidate set, and extract valid feature classification labels for physiological stress, behavioral avoidance, and psychological anxiety dimensions from the valid feature candidate set; Build a feature extraction model for multi-dimensional input data sets based on valid feature classification labels, extract and classify features based on the feature extraction model, generate structured feature subsets, and perform standardization processing; In this embodiment, the feature extraction unit further includes: Physiological stress feature subset: Perform time domain, frequency domain, and time-frequency analysis on the physiological data to extract exercise-induced stress response characteristics and sustained stress state characteristics. The exercise-induced stress response characteristics include the heart rate increase, the rising slope of the galvanic skin response, and the standard deviation of body acceleration within a preset time after the start of exercise. The sustained stress state characteristics include the ratio of low-frequency to high-frequency components of heart rate variability in the resting state and the proportion of abnormal stress physiological signal duration within 24 hours. Behavioral avoidance feature subset: Based on the behavioral data, the rehabilitation training compliance index, daily activity attenuation index and temporal pattern characteristics are calculated. The rehabilitation training compliance index includes the training plan completion rate, the frequency of active interruptions, and the training start delay time. The daily activity attenuation index includes the deviation rate between the daily activity volume and the expected amount of the rehabilitation plan and the gait asymmetry. The temporal pattern characteristics include the proportion of consecutive days with activity volume below the preset threshold; Psychological anxiety feature subset: The psychological data is subjected to scale score analysis and natural language processing to extract psychological scale-derived features and semantic-emotional features. The psychological scale-derived features include the scores of the movement fear-related dimensions in the standardized psychological scale, the negative correlation coefficient between the visual analog score and the actual activity amount, and the semantic-emotional features include the word frequency and emotional polarity score of fear-related words in the self-report text.

[0022] In this embodiment, the construction of three-dimensional feature subsets of physiological, behavioral, and psychological dimensions enables a comprehensive and multi-dimensional assessment of patients, providing a scientific basis for personalized rehabilitation plans. The structured feature processing and standardized process of the model enhance the consistency and comparability of the data, facilitate dynamic risk monitoring and abnormal warning, and improve the efficiency and accuracy of movement phobia diagnosis.

[0023] In this embodiment, the dynamic risk assessment module includes: a model building unit configured to build a corresponding sports fear quantification model based on the feature categories of the physiological stress feature subset, the behavioral avoidance feature subset, and the psychological anxiety feature subset; a grading strategy formulation unit configured to establish risk grading standards and determine exercise fear index thresholds for physiological stress characteristics, behavioral avoidance characteristics, and psychological anxiety characteristics at each level; a real-time processing unit configured to input the multi-dimensional input data set collected in real time into a movement phobia quantification model to calculate a movement phobia index of the patient, compare the movement phobia index with a movement phobia index threshold, and determine a movement phobia risk level of the patient; The dynamic threshold adjustment unit is configured to obtain and analyze the patient's basic data (such as age, gender, surgery type, postoperative time, etc.), and adjust the movement fear index threshold and weight coefficient of each specific characteristic according to the individual differences of the patient and the postoperative stage.

[0024] In this embodiment, a sports fear quantification model is constructed based on multi-dimensional features, the fear index is calculated in real time, and the threshold and weight are dynamically adjusted to adapt to individual differences in patients and changes in the postoperative stage, so as to achieve accurate classification of risk levels; at the same time, the risk level and the severity of the abnormality are integrated to generate a comprehensive assessment confidence, and personalized intervention strategies are matched to form an "assessment-intervention-tracking" closed loop, thereby improving the pertinence and effectiveness of the rehabilitation program.

[0025] In this embodiment, the abnormality identification and warning module includes: A temporal pattern recognition unit is configured to extract time series data from behavioral data, including daily activity trajectories and training environment interaction data, identify typical abnormal behavior sequences associated with kinesia, such as a persistently elevated heart rate one hour before training accompanied by avoidance of the training area, a 0% completion rate for rehabilitation training for three consecutive days, frequent interruptions in training, and prolonged avoidance of specific movements, and obtain characteristics of the abnormal behavior patterns, including the time of occurrence and duration. a persistent avoidance pattern recognition unit configured to extract a feature subset time series from the behavioral avoidance feature subset, identify potential persistent avoidance patterns based on the feature subset time series, such as low activity levels for multiple consecutive days, frequent active interruptions to training, etc., and obtain persistent avoidance pattern characteristics, including duration, intensity, etc.; a multimodal correlation assessment unit configured to perform correlation analysis on the identified abnormal behavior sequences and persistent avoidance patterns with a subset of physiological stress characteristics and a subset of psychological anxiety characteristics, assess the severity of the patient's abnormality (mild, moderate, or severe), and generate corresponding warning information based on the assessed severity of the abnormality; The abnormality classification warning unit is configured with a preset three-level abnormality severity standard (mild, moderate, and severe), combining the risk level of sports phobia and the posterior probability of abnormal behavior to trigger the corresponding warning mechanism; The early warning information generation unit is configured to generate structured early warning information for different levels of abnormalities, including the time of abnormality occurrence, characteristic evidence chain (such as physiological signal fluctuation curve, behavioral trajectory heat map, psychological scale score trend), and recommended intervention time window, and synchronize the early warning information between the medical and patient ends.

[0026] In this embodiment, abnormal behavior sequences and continuous avoidance patterns are identified through time series analysis, and multimodal data is associated to assess the severity of abnormalities and provide graded warnings to achieve early intervention in risk events. With the help of visualization tools such as risk heat maps, trend comparison charts on the medical side, and progress dashboards on the patient side, the synchronization of medical and patient information is promoted, and the intelligence and collaborative efficiency of the diagnosis and treatment process are improved.

[0027] In one embodiment, specific features related to physiological stress, behavioral avoidance, and psychological anxiety are extracted from a multi-dimensional input dataset based on a feature extraction model, including: Determine the motion state characteristics of each item of data in the multidimensional data set based on the feature extraction model, and determine the patient's response form to each item of data according to the motion state characteristics, wherein the response form includes: physiological response, behavioral response and psychological response; Classify each data item according to the patient's response form to each data item, obtain the classification results, and determine the unified state description parameters of each type of data based on the classification results; Determine the unified adaptability influencing factors of each type of data on the patient based on the unified state description parameters, and determine the impact feedback intensity of each type of data on the patient based on the unified adaptability influencing factors; Determine the average stress state index of each type of data for the patient based on the impact feedback intensity, and determine the quantitative division of the observed events based on the average stress state index and the descriptive parameters of each type of data; Based on the observed event division, each type of data is quantitatively divided into events to generate an observed event sequence, and the observed event sequence is decomposed into independent event components; Obtaining an event completion goal and an event process indicator for each independent event, and determining a physiological response state description parameter, a behavioral response state description parameter, and a psychological response state description parameter for each independent event based on the event completion goal and the event process indicator; Obtaining the adaptability of the subjective morphological characteristics of the physiological response state description parameters, the behavioral response state description parameters, and the psychological response state description parameters to the objective event development characteristics; Determine the psychological burden, behavioral limitations, and physiological resistance of each individual event to the patient based on suitability; Determine the physiological / psychological / behavioral stimulation expression level of each independent event based on the patient's psychological burden, behavioral limitation, and physiological resistance. Filter out target independent events that are greater than or equal to the preset expression level based on the physiological / psychological / behavioral stimulus expression level of each independent event; Specific features related to physiological stress, behavioral avoidance and psychological anxiety are extracted based on the event description of target independent events through the feature extraction model.

[0028] The above technical solution has the beneficial effect of accurately identifying a user's specific fear events and behaviors by splitting multi-dimensional input data into independent events and then evaluating them, thereby extracting specific features based on the event description of the specific fear event. This improves the reliability and accuracy of feature extraction.

[0029] In one embodiment, the real-time collection of the patient's motion parameter data from the rehabilitation training device, and the acquisition of the patient's daily activity trajectory and training environment interaction data to generate the patient's behavioral data, include: Collecting the patient's real-time exercise project parameters from the rehabilitation training equipment, and determining the patient's exercise form based on the real-time exercise project parameters; Determine the associated activity mode and environmental interaction mode based on the patient's movement form, and determine the behavior matching identification index based on the associated activity mode and environmental interaction mode; Obtain the matching factor of the behavioral matching identification indicator. At the same time, use millimeter-level radar and sensors to obtain the patient's daily activity trajectory in each sport and the training environment interaction data; The patient's daily activity trajectory and training environment interaction data are evaluated for sports project matching through behavioral matching identification indicators: in, It is represented by the daily activity trajectory of the i-th sport and the matching degree between the training environment interaction data and the sport. It is expressed as the exercise relaxation of the i-th exercise, It is represented by the patient's daily activity trajectory and the number of training environment interaction data in the i-th sport, j is represented by the j-th daily activity trajectory and training environment interaction data, It is expressed as the maximum matrix factor of the motion observation matrix of the j-th daily activity trajectory and the training environment interaction data, It is expressed as the matching factor of the behavioral matching identification index, ln is expressed as the natural logarithm, It is represented as the sports target identification intensity value of the i-th sports event, It is represented by the patient's daily activity trajectory of the i-th exercise and the exercise feedback intensity value shown in the training environment interaction data; Filter out qualified daily activity trajectory and training environment interaction data sets for each sport based on the matching degree between the daily activity trajectory and training environment interaction data of each sport and the sport; Generate patient behavioral data based on qualified daily activity trajectories and training environment interaction datasets.

[0030] The beneficial effects of the above technical solution are: by calculating the matching degree between the daily activity trajectory of each sport and the training environment interaction data and the sport, the motion matching factor can be determined based on the sport relaxation of each sport, and then the motion observation matrix of the daily activity trajectory and the training environment interaction data can be evaluated to screen the highly correlated behavioral data of each sport, thereby ensuring the rationality and reliability of the behavioral data and its correlation with each sport, laying the foundation for the subsequent assessment of sports fear in patients and improving practicality.

[0031] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A patient assessment system for kinesia, characterized in that: include: a kinesiophobia feature extraction module configured to acquire multidimensional data of the patient in real time based on multiple source devices, construct and process a multidimensional input data set, and extract specific features representing kinesiophobia; a dynamic risk assessment module configured to input the extracted specific features into a constructed kinesia quantification model to calculate a kinesia index, and to assess the patient's kinesia risk level based on the calculated result; An abnormality identification and warning module is configured to identify abnormal behavior sequences and persistent avoidance patterns related to patients' fear of movement based on time series analysis, and to perform correlation assessments on specific features of each category based on the identification results to determine the severity of the patient's abnormality and issue a warning; The patient assessment decision module is configured to integrate the patient's movement phobia risk level and abnormality severity assessment results, generate a comprehensive movement phobia assessment confidence level, and form a personalized rehabilitation plan based on the corresponding intervention strategy matched with the patient's basic data.

2. A patient assessment system for kinesiophobia according to claim 1, characterized in that: Movement fear feature extraction module, including: a multi-dimensional data acquisition unit configured to acquire the patient's physiological data based on the wearable device, acquire the patient's corresponding behavioral data, and simultaneously acquire the patient's psychological state data actively input; a data processing unit configured to clean, calibrate, and standardize the physiological data, behavioral data, and psychological data acquired by the multidimensional data acquisition unit to form a structured multidimensional input data set; The feature extraction unit is configured to extract specific features related to physiological stress, behavioral avoidance and psychological anxiety from a multi-dimensional input data set based on a feature extraction model, and generate a physiological stress feature subset, a behavioral avoidance feature subset and a psychological anxiety feature subset.

3. A patient assessment system for kinesiophobia according to claim 2, characterized in that: Based on the feature extraction model, specific features related to physiological stress, behavioral avoidance, and psychological anxiety are extracted from the multi-dimensional input dataset, including: Determine the motion state characteristics of each item of data in the multidimensional data set based on the feature extraction model, and determine the patient's response form to each item of data according to the motion state characteristics, wherein the response form includes: physiological response, behavioral response and psychological response; Classify each data item according to the patient's response form to each data item, obtain the classification results, and determine the unified state description parameters of each type of data based on the classification results; Determine the unified adaptability influencing factors of each type of data on the patient based on the unified state description parameters, and determine the impact feedback intensity of each type of data on the patient based on the unified adaptability influencing factors; Determine the average stress state index of each type of data for the patient based on the impact feedback intensity, and determine the quantitative division of the observed events based on the average stress state index and the descriptive parameters of each type of data; Based on the observed event division, each type of data is quantitatively divided into events to generate an observed event sequence, and the observed event sequence is decomposed into independent event components; Obtaining an event completion goal and an event process indicator for each independent event, and determining a physiological response state description parameter, a behavioral response state description parameter, and a psychological response state description parameter for each independent event based on the event completion goal and the event process indicator; Obtaining the adaptability of the subjective morphological characteristics of the physiological response state description parameters, the behavioral response state description parameters, and the psychological response state description parameters to the objective event development characteristics; Determine the psychological burden, behavioral limitations, and physiological resistance of each individual event to the patient based on suitability; Determine the physiological / psychological / behavioral stimulation expression level of each independent event based on the patient's psychological burden, behavioral limitation, and physiological resistance. Filter out target independent events that are greater than or equal to the preset expression level based on the physiological / psychological / behavioral stimulus expression level of each independent event; Specific features related to physiological stress, behavioral avoidance and psychological anxiety are extracted based on the event description of target independent events through the feature extraction model.

4. The patient assessment system for kinesiophobia according to claim 2, wherein: Multi-dimensional data acquisition unit, specifically including: Deploy a network of wearable sensor devices to collect physiological data from patients; Collect the patient's motion parameter data in real time from the rehabilitation training equipment. At the same time, obtain the patient's daily activity trajectory and training environment interaction data to generate the patient's behavioral data. The patient's voice data is acquired based on voice interaction, the voice data is converted into text and subjected to sentiment analysis to generate the patient's psychological data.

5. The patient assessment system for kinesiophobia according to claim 4, wherein: The feature extraction model specifically includes: Extract data field labels from multi-dimensional input data sets, match data types in the multi-dimensional input data according to a preset dictionary, and match corresponding standard feature templates from a preset feature sample library based on the identified data types; Generate validity constraints for data verification based on feature sample templates, filter and verify each data point in the multi-dimensional input data based on the validity constraints, and mark data points that do not meet the validity constraints as marginal data; Extract data other than marginal data from the multi-dimensional input data set as valid feature data, generate a valid feature candidate set, and extract valid feature classification labels for physiological stress, behavioral avoidance, and psychological anxiety dimensions from the valid feature candidate set; A feature extraction model for multi-dimensional input data sets is constructed based on effective feature classification labels. Features are extracted and classified based on the feature extraction model, and structured feature subsets are generated and standardized.

6. A patient assessment system for kinesiophobia according to claim 5, characterized in that: The feature extraction unit also includes: Physiological stress feature subset: performing time domain, frequency domain and time-frequency analysis on the physiological data to extract the characteristics of exercise-induced stress response and sustained stress state; Behavioral avoidance feature subset: Calculate rehabilitation training compliance index, daily activity attenuation index and temporal pattern characteristics based on the behavioral data; Psychological anxiety feature subset: Perform scale score analysis and natural language processing on the psychological data to extract psychological scale-derived features and semantic emotional features.

7. The patient assessment system for kinesiophobia according to claim 6, wherein: Dynamic risk assessment module, including: a model building unit configured to build a corresponding sports fear quantification model based on the feature categories of the physiological stress feature subset, the behavioral avoidance feature subset, and the psychological anxiety feature subset; a grading strategy formulation unit configured to establish risk grading standards and determine exercise fear index thresholds for physiological stress characteristics, behavioral avoidance characteristics, and psychological anxiety characteristics at each level; a real-time processing unit configured to input the multi-dimensional input data set collected in real time into a movement phobia quantification model to calculate a movement phobia index of the patient, compare the movement phobia index with a movement phobia index threshold, and determine a movement phobia risk level of the patient; The dynamic threshold adjustment unit is configured to obtain and analyze the patient's basic data, and adjust the movement fear index threshold and weight coefficient of each specific feature according to the individual differences of the patient and the postoperative stage.

8. The patient assessment system for kinesiophobia according to claim 7, wherein: Abnormal identification and warning module, including: a time series pattern recognition unit configured to extract time series data from the behavioral data, identify typical abnormal behavior sequences associated with motor phobia, and obtain abnormal behavior pattern features; a persistent avoidance pattern recognition unit configured to extract a feature subset time series from the behavioral avoidance feature subset, identify a potential persistent avoidance pattern based on the feature subset time series, and obtain persistent avoidance pattern features; a multimodal correlation assessment unit configured to perform correlation analysis on the identified abnormal behavior sequences and persistent avoidance patterns with a subset of physiological stress characteristics and a subset of psychological anxiety characteristics, assess the severity of the patient's abnormality, and generate corresponding warning information based on the assessed severity of the abnormality; The abnormality classification warning unit is configured with a preset three-level abnormality severity standard, combining the risk level of sports phobia and the posterior probability of abnormal behavior to trigger the corresponding warning mechanism; The warning information generation unit is configured to generate structured warning information for different levels of abnormalities and synchronize the warning information between the medical and patient ends.

9. The patient assessment system for kinesiophobia according to claim 8, wherein: The patient assessment and decision-making module also includes: visual display of assessment results and intervention progress based on the smart terminals on the medical and patient sides. Among them, the visual display of the smart terminal on the medical side includes: risk heat map, trend comparison chart and intervention tracking table; the visual display of the smart terminal on the patient side includes: progress dashboard, fear diary and education database.

10. The patient assessment system for kinesiophobia according to claim 4, wherein: The patient's motion parameter data is collected in real time from the rehabilitation training equipment, and at the same time, the patient's daily activity trajectory and training environment interaction data are obtained to generate the patient's behavior data, including: Collecting the patient's real-time exercise project parameters from the rehabilitation training equipment, and determining the patient's exercise form based on the real-time exercise project parameters; Determine the associated activity mode and environmental interaction mode based on the patient's movement form, and determine the behavior matching identification index based on the associated activity mode and environmental interaction mode; Obtain the matching factor of the behavioral matching identification indicator. At the same time, use millimeter-level radar and sensors to obtain the patient's daily activity trajectory in each sport and the training environment interaction data; The patient's daily activity trajectory and training environment interaction data are evaluated for sports project matching through behavioral matching identification indicators: in, It is represented by the daily activity trajectory of the i-th sport and the matching degree between the training environment interaction data and the sport. Expressed as the exercise relaxation of the i-th exercise, It is represented by the patient's daily activity trajectory and the number of training environment interaction data in the i-th sport, j is represented by the j-th daily activity trajectory and training environment interaction data, It is expressed as the maximum matrix factor of the motion observation matrix of the j-th daily activity trajectory and the training environment interaction data, It is expressed as the matching factor of the behavior matching identification index, ln is expressed as the natural logarithm, It is represented as the sports target identification intensity value of the i-th sports event, It is represented by the patient's daily activity trajectory of the i-th sport and the exercise feedback intensity value shown in the training environment interaction data; Filter out qualified daily activity trajectory and training environment interaction data sets for each sport based on the matching degree between the daily activity trajectory and training environment interaction data of each sport and the sport; Generate patient behavioral data based on qualified daily activity trajectories and training environment interaction datasets.

Images