Health status assessment system applied to disease follow-up visit

Abstract

The invention relates to the technical field of disease follow-up visit, in particular to a health state assessment system applied to disease follow-up visit, which comprises a data acquisition module, a data preprocessing module, a health state assessment module, a personalized health management suggestion module, a dynamic health risk prediction module and a visualization module, the data acquisition module acquires multi-dimensional data of a patient and performs data fusion and integration, the health state evaluation module evaluates the health state of the patient based on multi-level feature extraction and a deep learning model, and the personalized health management suggestion module combines a health state evaluation result and historical data to obtain a personalized health management suggestion result. Customized health management suggestions are generated through a rule engine, and the dynamic health risk prediction module predicts future health risks of patients and provides real-time early warning by using a time sequence analysis model. The visualization module displays health assessment and management suggestion results through a chart and an interactive interface, and the accuracy and individuation degree of health management are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of disease follow-up, and in particular to a health status assessment system applied to disease follow-up. Background Art

[0002] With the continuous development of medical technology and health management needs, disease follow-up and health management have gradually become important directions in the field of medical health around the world. Especially in the areas of chronic disease management, elderly health monitoring and disease prevention, how to achieve efficient and accurate health status assessment and personalized health management has become an important issue to improve the health level of the population.

[0003] However, current health management systems generally face problems such as insufficient data integration, poor prediction accuracy, and insufficient personalization of recommendations, making it difficult to meet the needs of real applications. Although some systems attempt to use multi-dimensional data for health assessment, there are often information islands in data collection, and it is impossible to obtain the patient's comprehensive health status in real time. In addition, existing systems mostly rely on traditional rule-based health advice generation methods, which lack flexibility and adaptability, and cannot dynamically adjust recommendations based on changes in the patient's health status. In addition, existing health risk prediction systems also have certain limitations, such as the inability to effectively handle individual differences among patients, low accuracy of prediction results, and difficulty in providing personalized health warnings and intervention measures, resulting in a lack of health management effectiveness. Summary of the invention

[0004] The invention provides a health status assessment system applied to disease follow-up.

[0005] A health status assessment system for disease follow-up, comprising a data acquisition module, a data preprocessing module, a health status assessment module, a personalized health management suggestion module, a dynamic health risk prediction module and a visualization module, wherein;

[0006] The data collection module is used to collect multi-dimensional data of the patient in real time, including physiological data, clinical data, health questionnaire data, lifestyle data and environmental data, and identify and integrate different types of data through data fusion technology to form a patient's health record;

[0007] The data preprocessing module is used to perform denoising, outlier detection, missing data filling and standardization on the collected multi-dimensional data;

[0008] The health status assessment module is used to combine the collected multi-dimensional data, based on a multi-level feature extraction method, through a health status assessment model, comprehensively assess the patient's current health status, including physiological status, psychological status and living habits, and generate a patient health status assessment report;

[0009] The personalized health management suggestion module generates personalized health management suggestions through a rule engine based on the health status assessment results and the patient's historical multi-dimensional data;

[0010] The dynamic health risk prediction module predicts the patient's future health risks based on the patient's historical multi-dimensional data, real-time lifestyle data, real-time environmental data and health status assessment results, combined with the time series analysis model, and generates personalized health warnings;

[0011] The visualization module is used to display health status assessment results, personalized health management suggestions and health risk prediction results in the form of charts, graphs and text, and provide them to users (such as patients, doctors, and health managers) through an interactive interface for real-time viewing and operation.

[0012] Optionally, the data acquisition module includes:

[0013] Physiological data collection: The patient's physiological data is collected in real time through monitoring equipment, including body temperature, blood pressure, heart rate, blood oxygen, weight and blood sugar. The monitoring equipment includes smart bracelets, electronic blood pressure monitors, heart rate sensors, finger-clip pulse oximeters, weight scales and blood glucose meters;

[0014] Clinical data collection: Access medical databases through API interfaces to obtain patients' clinical data, including patients' diagnosis records, blood test results, imaging data, and treatment history;

[0015] Health questionnaire data collection: The health questionnaire data of patients are collected through a questionnaire survey platform or face-to-face interviews. The health questionnaire data includes the patient's self-reported symptoms, medical history, psychological state and quality of life assessment;

[0016] Lifestyle data collection: real-time collection of patients’ lifestyle data through smart devices or applications, including eating habits, exercise, sleep quality, smoking and drinking behavior data;

[0017] Environmental data collection: Environmental data of the patient's environment is collected through environmental sensors and smart devices. Environmental data includes air quality, temperature and humidity, noise, and pollutant concentration;

[0018] Data fusion and integration: During the data collection process, the cosine similarity algorithm is used to integrate data from different sources to form the patient's health record.

[0019] Optionally, the data preprocessing module includes:

[0020] De-noising: De-noising the collected multi-dimensional data through Kalman filtering;

[0021] Outlier detection: Use the Z-Score method to detect outliers in multi-dimensional data;

[0022] Missing data filling: fill in missing data in multi-dimensional data;

[0023] Standardization: Use the minimum-maximum standardization method to standardize multi-dimensional data;

[0024] Data consistency and comparability verification: Perform consistency verification on data that has undergone denoising, outlier detection, missing data filling, and standardization.

[0025] Optionally, the health status assessment module includes:

[0026] Multi-level feature extraction: Extract features from pre-processed multi-dimensional data, including physiological data, health questionnaire data, and lifestyle data;

[0027] Feature selection: L1 regularization method was used to perform feature selection on the extracted physiological data, health questionnaire data, and lifestyle data;

[0028] Feature fusion: features extracted from physiological data, health questionnaire data and lifestyle data are fused to obtain a unified feature vector;

[0029] Model input and evaluation: The feature vector after feature selection and feature fusion is used as input and input into the health status assessment model for health status assessment, and the patient's health assessment report is output.

[0030] Optionally, the model input and evaluation includes:

[0031] Health status assessment model construction: A health status assessment model is constructed based on a convolutional neural network. The model structure includes multiple convolutional layers, pooling layers, and fully connected layers.

[0032] Loss function and optimization: Use cross entropy loss function to optimize the parameters of the health status assessment model;

[0033] Model training: The feature vectors after feature selection and feature fusion are divided into training set and validation set. The training set data is used to train the health status assessment model, and the validation set data is used to verify the parameters of the health status assessment model.

[0034] Health status assessment results: The trained health status assessment model outputs the patient's health status assessment results based on the input multi-dimensional data, including a comprehensive assessment of the physiological state, psychological state and living habits, and generates a personalized health report.

[0035] Optionally, the personalized health management suggestion module generates personalized health management suggestions through a rule engine and a machine learning algorithm based on the health status assessment results and the patient's historical multi-dimensional data, specifically including:

[0036] Integration of health status assessment results and historical data: The personalized health management suggestion module integrates the health status assessment results (including physiological status, psychological status and behavioral habit assessment) generated by the health status assessment module with the patient's historical multi-dimensional data (including historical physiological data, historical health questionnaire data, historical lifestyle data, etc.). The integrated data forms the patient's complete health record, providing basic data for generating personalized health management suggestions;

[0037] Rule engine application: The personalized health management suggestion module analyzes the integrated complete health records through the rule engine and generates preliminary health management suggestions based on the preset health management rules (such as the standard range of physiological indicators such as blood sugar value and blood pressure, the suitability of lifestyle, etc.);

[0038] Personalized recommendation generation: Based on health status assessment results and historical multi-dimensional data, the rule engine automatically generates personalized health management recommendations, including diet recommendations, exercise recommendations, medication recommendations, and regular check-up recommendations.

[0039] Optionally, the dynamic health risk prediction module includes:

[0040] Data reception: The dynamic health risk prediction module receives the patient's real-time lifestyle data, real-time environmental data, and the patient's historical multi-dimensional data;

[0041] Time series analysis model construction and training: Based on the patient's real-time lifestyle data, real-time environmental data, and historical multi-dimensional data of the patient, a time series analysis model is constructed through a long short-term memory network, and the time series analysis model is trained through the patient's historical multi-dimensional data;

[0042] Health risk prediction: Through the trained time series analysis model, the dynamic health risk prediction module predicts the patient's future health risks (such as cardiovascular disease risk, diabetes risk, etc.) based on real-time lifestyle data and real-time environmental data, and generates corresponding health risk scores.

[0043] Optionally, the dynamic health risk prediction module further includes:

[0044] Health risk assessment and warning generation: Compare the prediction results of the time series analysis model with the preset health risk threshold to generate personalized health risk warnings;

[0045] Risk prediction result display: The prediction results are displayed through the visualization module, including health risk score, risk level and related health warning information.

[0046] Risk level classification: Based on the health risk prediction results, health risks are divided into low risk, medium risk and high risk.

[0047] Optionally, the visualization module includes:

[0048] Health status assessment result display: The visualization module displays the health status assessment results generated by the health status assessment module in the form of charts, graphs and text;

[0049] Personalized health management suggestion display: The visualization module displays the patient's diet, exercise, medication and regular check-up suggestions based on the personalized health management suggestion content generated by the personalized health management suggestion module;

[0050] Health risk prediction result display: The visualization module displays the prediction results of the dynamic health risk prediction module in the form of charts and text;

[0051] Interactive interface and user operation: The visualization module provides real-time viewing and operation functions through an interactive interface.

[0052] Beneficial effects of the present invention:

[0053] The present invention, by combining the patient's historical multi-dimensional data, real-time lifestyle data, real-time environmental data and health status assessment results, adopts a multi-level feature extraction method and a deep learning model to comprehensively and accurately assess the patient's health status. The health status assessment module provides patients with a more personalized health management plan through a comprehensive analysis of multi-dimensional indicators such as physiology, psychology and behavioral habits, ensuring that health management recommendations meet the specific needs of each patient, thereby greatly improving the accuracy and personalization of health management and avoiding the traditional one-size-fits-all health recommendation model.

[0054] The present invention, by using a time series analysis model, the dynamic health risk prediction module can track the patient's health status changes in real time, combine historical data with real-time collected data, and predict the patient's future health risks. This dynamic prediction capability based on real-time data streams and historical data enables the present invention to promptly discover potential health risks and generate personalized health warnings in advance, thereby providing patients with opportunities for early intervention and greatly reducing the probability of health events, especially in the fields of chronic disease management and elderly patient monitoring.

[0055] In the present invention, the visualization module presents the health status assessment results, personalized health management suggestions and health risk prediction results to patients, doctors and health managers in the form of charts, graphics and text. This visualization display method not only allows users to intuitively understand complex health data, but also improves the interactivity and participation of the health management process. Patients can view and operate their health records in real time through the interactive interface, and doctors and health managers can make decisions and intervene more efficiently. In addition, the display of the visualization module helps patients better track their own health changes and enhances their self-awareness and self-management ability of health management. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0057] Figure 1 A schematic diagram of a system flow of an embodiment of the present invention;

[0058] Figure 2 Schematic diagram of the dynamic health risk prediction module flow in an embodiment of the present invention. DETAILED DESCRIPTION

[0059] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0060] like Figure 1-Figure 2 As shown, a health status assessment system for disease follow-up includes a data acquisition module, a data preprocessing module, a health status assessment module, a personalized health management suggestion module, a dynamic health risk prediction module and a visualization module, wherein;

[0061] The data collection module is used to collect multi-dimensional data of patients in real time. The multi-dimensional data includes physiological data, clinical data, health questionnaire data, lifestyle data and environmental data. Through data fusion technology, it identifies and integrates different types of data to form the patient's health record;

[0062] The data preprocessing module is used to perform denoising, outlier detection, missing data filling and standardization on the collected multi-dimensional data to ensure consistency and comparability between different data sources;

[0063] The health status assessment module is used to combine the collected multi-dimensional data, based on the multi-level feature extraction method, through the health status assessment model, comprehensively assess the patient's current health status, including physiological status, psychological status and living habits, and generate a patient health status assessment report;

[0064] The personalized health management suggestion module generates personalized health management suggestions, including diet, exercise, medication and regular check-ups, based on the health status assessment results and the patient's historical multi-dimensional data through the rule engine;

[0065] The dynamic health risk prediction module predicts the patient's future health risks based on the patient's historical multi-dimensional data, real-time lifestyle data, real-time environmental data and health status assessment results, combined with the time series analysis model, and generates personalized health warnings;

[0066] The visualization module is used to display health status assessment results, personalized health management recommendations and health risk prediction results in the form of charts, graphs and text, and provide them to users (such as patients, doctors, and health managers) through an interactive interface for real-time viewing and operation, so that users can intuitively understand and track their health status and corresponding management strategies.

[0067] The data acquisition module includes:

[0068] Physiological data collection: The patient's physiological data is collected in real time through monitoring equipment, including body temperature, blood pressure, heart rate, blood oxygen, weight and blood sugar. The monitoring equipment includes smart bracelets, electronic blood pressure monitors, heart rate sensors, finger-clip pulse oximeters, weight scales and blood glucose meters, among which;

[0069] Smart bracelets are used to collect patient temperature data;

[0070] The electronic sphygmomanometer is used to collect blood pressure data of patients;

[0071] The heart rate sensor is used to collect the patient's heart rate data;

[0072] Finger-clip pulse oximeter is used to collect patient blood oxygen data;

[0073] The weighing scale is used to collect the patient's weight data;

[0074] The blood glucose meter is used to collect the patient's blood glucose data;

[0075] Clinical data collection: Access medical databases through API interfaces to obtain patients' clinical data, including patients' diagnosis records, blood test results, imaging data, and treatment history;

[0076] Health questionnaire data collection: Health questionnaire data from patients are collected through a questionnaire survey platform or face-to-face interviews. The health questionnaire data includes self-reported symptoms, medical history, psychological state, and quality of life assessment. Patients can fill out standardized questionnaires through mobile devices (such as smartphones and tablets), or have health management personnel conduct face-to-face interviews to input data;

[0077] Lifestyle data collection: real-time collection of patients’ lifestyle data through smart devices or applications, including eating habits, exercise, sleep quality, smoking and drinking behavior data;

[0078] The patient's dietary data is collected through the application. The application records the patient's daily diet through manual input or taking photos, and analyzes the calories and nutritional content of the food through AI algorithms;

[0079] Monitor the patient's exercise data in real time through a smart bracelet;

[0080] Monitor your sleep quality with a smart bracelet or smartwatch (such as the Apple Watch);

[0081] Record patients’ smoking and drinking habits through health management applications, and track the frequency, quantity, and duration of smoking and drinking through patient self-reporting or;

[0082] Environmental data collection: Environmental data of the patient's environment is collected through environmental sensors and smart devices. Environmental data includes air quality, temperature and humidity, noise, and pollutant concentration;

[0083] Data fusion and integration: During the data collection process, the cosine similarity algorithm is used to integrate data from different sources to form the patient's health record. The specific steps are as follows:

[0084] Format conversion and standardization: Format conversion from different data sources (such as physiological data, clinical data, health questionnaire data, lifestyle data, and environmental data) to ensure that all data are in a standard format (such as JSON, XML, or CSV). At the same time, the data is standardized so that different data dimensions have the same metric to ensure data comparability;

[0085] Timestamp synchronization: Real-time collected data is synchronized through timestamps to ensure that all types of data are arranged in chronological order, solving the time inconsistency problem caused by real-time collection from different data sources. Real-time data and historical data will be matched and integrated through a unified timeline;

[0086] Matching algorithm: The cosine similarity algorithm is used to match and integrate different types of data. Cosine similarity calculates the similarity of two sets of data and evaluates their similarity in multi-dimensional space, thereby realizing the association and integration of different types of data. The calculation formula is as follows:

[0087] ;

[0088] in, and are two vectors, representing the feature vectors of the two data sets respectively. is the vector inner product, and is the magnitude of the vector;

[0089] By calculating the similarity between various data sources, the multi-dimensional data of patients can be effectively matched and associated to ensure the accuracy of data fusion;

[0090] The data collection module is designed in multiple steps to collect all kinds of multi-dimensional data of patients in real time and comprehensively, including physiological data, clinical data, health questionnaire data, lifestyle data and environmental data. Through data fusion technology, the system can effectively integrate data from different sources to form a complete patient health record, which can comprehensively and accurately capture the patient's health information and provide complete basic data for subsequent health status assessment and personalized management. At the same time, real-time data collection and intelligent fusion technology also improve the flexibility and adaptability of the system.

[0091] The data preprocessing module includes:

[0092] Denoising: Denoise the collected multi-dimensional data through Kalman filtering. The calculation formula is expressed as:

[0093] ;

[0094] in, is the estimated current state, is the Kalman gain, is the measured value;

[0095] Outlier detection: The Z-Score method is used to detect outliers in multi-dimensional data, expressed as:

[0096] ;

[0097] in, For the data points, is the mean of the data set, is the standard deviation;

[0098] Missing data filling: fill in missing data in multi-dimensional data, including:

[0099] Mean filling: For numerical data, fill in the data by calculating the mean of the dimension data;

[0100] Interpolation method: For time series data, use linear interpolation or spline interpolation to fill missing values. Linear interpolation is expressed as:

[0101] ;

[0102] in, is a known point, is the value corresponding to the known point, is the position to be filled;

[0103] Standardization: The minimum-maximum standardization method is used to standardize multi-dimensional data to ensure that different data dimensions have the same dimensions and make them comparable. The minimum-maximum standardization is to scale the data to a specified range (usually 0 to 1), expressed as:

[0104] ;

[0105] in, and are the minimum and maximum values ​​in the data set, respectively. is the value to be standardized;

[0106] Data consistency and comparability verification: Perform consistency verification on data that has undergone denoising, outlier detection, missing data filling, and standardization to ensure that different data sources have the same data structure and data type, and that the timeliness of the data is guaranteed. By verifying data types, timestamp alignment, and data format conversion, ensure that all data sets can be compared and integrated in the same analysis model;

[0107] The data preprocessing module effectively improves the quality of collected data through steps such as denoising, outlier detection, missing data filling and standardization, eliminates noise, identifies and corrects outliers, fills missing data, and ensures consistency and comparability between different data sources. The preprocessing effect of this module provides accurate and reliable basic data for subsequent health status assessment, personalized health management recommendations and health risk prediction.

[0108] The health status assessment module includes:

[0109] Multi-level feature extraction: Extract features from pre-processed multi-dimensional data, physiological data, health questionnaire data, and lifestyle data, including:

[0110] Feature extraction of physiological data: For the collected physiological data (such as body temperature, blood pressure, heart rate, blood oxygen, weight, blood sugar, etc.), the following methods are used for feature extraction:

[0111] (1) Statistical feature extraction: Calculate the mean, standard deviation, maximum, minimum, volatility and other statistical features of physiological data to reveal the patient's basic physiological state.

[0112] (2) Time series feature extraction: Perform Fourier transform on the time series signals of physiological data (such as heart rate, blood pressure, etc.), extract frequency domain features, and identify periodic change patterns, such as heart rate variability and blood sugar fluctuation period. Use Wavelet transform method to capture local features of non-stationary signals, such as blood sugar fluctuation and short-term fluctuation of body temperature.

[0113] Feature extraction of health questionnaire data: By performing natural language processing on the health questionnaire data filled out by patients, information such as mental health and mood swings can be extracted, including:

[0114] (1) Sentiment analysis: Conduct sentiment analysis on open-ended questions in health questionnaires to identify the patient’s psychological state (such as anxiety, depression, etc.) and quantify it into a sentiment index.

[0115] (2) Feature engineering: Combine the scale data in the mental health questionnaire (such as PHQ-9, GAD-7, etc.) to extract the patient's mental health score and its changing trend.

[0116] Feature extraction of lifestyle data: Lifestyle data mainly includes the patient's eating habits, exercise, sleep quality, smoking and drinking behavior. The extraction steps include:

[0117] (1) Eating habits analysis: Based on the dietary data collected by smart devices (such as smart plates and health management applications), information such as dietary calories, nutrients, and food categories is extracted.

[0118] (2) Exercise volume analysis: Through smart watches, bracelets and other devices, indicators such as exercise duration, intensity (such as number of steps, exercise type, etc.) and exercise calories consumed are extracted.

[0119] (3) Sleep quality analysis: Through sleep monitoring devices (such as Oura Ring, Fitbit, etc.), key characteristics such as the patient's sleep duration, deep sleep duration, sleep efficiency, etc. are extracted to evaluate sleep quality.

[0120] (4) Analysis of smoking and drinking behavior: extract information such as smoking frequency and drinking amount through the patient's behavior records or health questionnaires;

[0121] Feature selection: The L1 regularization method is used to select features from the extracted physiological data, health questionnaire data, and lifestyle data. L1 regularization adds an L1 penalty term to the regression model to force the coefficients of certain features to be zero, thereby removing redundant features and retaining the features that contribute most to health status assessment. The calculation is:

[0122] ;

[0123] in, It is The true value of the samples, It is The feature vector of the samples, is the regression coefficient, is a regularization parameter that controls the strength of feature selection. Through the L1 regularization method, the features that have the greatest impact on health status assessment are selected to reduce the risk of overfitting;

[0124] Feature fusion: The features extracted from physiological data, health questionnaire data and lifestyle data are fused to obtain a unified feature vector. The specific method is as follows:

[0125] Weighted fusion method: According to the weights of different data sources, multiple features are weighted and merged. The weights are verified by experiments or determined based on the importance of the data source. The merged feature vector expresses the comprehensive information of various types of data. The merging formula is expressed as:

[0126] ;

[0127] in, is the fused feature vector, It is The feature vector of the data source, It is The weight of the data source, ;

[0128] Principal component analysis: If the data dimension is too high, use the PCA method to reduce the dimension of the features to reduce redundant data and retain the most important health information. PCA calculates the covariance matrix and extracts its eigenvectors, and selects the principal component that explains the maximum variance for feature representation. The principal component analysis is expressed as

[0129] ;

[0130] in, is the original data matrix, are the eigenvectors of the data matrix, is the feature matrix after dimensionality reduction;

[0131] Model input and evaluation: The feature vector after feature selection and feature fusion is used as input and input into the health status assessment model for health status assessment, and the patient's health assessment report is output.

[0132] Model inputs and evaluation include:

[0133] Health status assessment model construction: A health status assessment model is constructed based on a convolutional neural network. The model structure includes multiple convolutional layers, pooling layers, and fully connected layers. The specific structure is as follows:

[0134] Convolutional layer: Convolution operation is performed on the input data through multiple convolution kernels to extract local features in the data. With input data Perform convolution calculation and output feature map , expressed as:

[0135] ;

[0136] in, is the convolution kernel, are the elements of the input data, is the bias, It is the feature map after convolution operation;

[0137] Pooling layer: Use maximum pooling or average pooling operations to reduce data dimensions and retain key information. The pooling operation can be calculated using the following formula, assuming that maximum pooling is used:

[0138] ;

[0139] in, is the pooling result, is the element of the pooling area;

[0140] Fully connected layer: The pooled features are further processed by the fully connected layer and the health status prediction results are output. The calculation formula of the fully connected layer is:

[0141] ;

[0142] in, is the weight matrix of the fully connected layer, is the feature map after pooling, is the bias term, is the final output;

[0143] Loss function and optimization: The cross entropy loss function is used to optimize the health status assessment model parameters, expressed as:

[0144] ;

[0145] in, is the sample size, is the true label, is the model prediction output, is the loss function;

[0146] Model training: The feature vectors after feature selection and feature fusion are divided into training set and validation set. The training set data is used to train the health status assessment model, and the validation set data is used to verify the parameters of the health status assessment model.

[0147] Health status assessment results: The trained health status assessment model outputs the patient's health status assessment results based on the input multi-dimensional data, including a comprehensive assessment of the physiological state, psychological state and living habits, and generates a personalized health report.

[0148] The personalized health management suggestion module generates personalized health management suggestions based on the health status assessment results and the patient's historical multi-dimensional data through the rule engine and machine learning algorithm, including:

[0149] Integration of health status assessment results and historical data: The personalized health management suggestion module integrates the health status assessment results (including physiological status, psychological status and behavioral habit assessment) generated by the health status assessment module with the patient's historical multi-dimensional data (including historical physiological data, historical health questionnaire data, historical lifestyle data, etc.). The integrated data forms the patient's complete health record, providing basic data for generating personalized health management suggestions;

[0150] Rule engine application: The personalized health management suggestion module analyzes the integrated complete health records through the rule engine and generates preliminary health management suggestions based on the preset health management rules (such as the standard range of physiological indicators such as blood sugar value and blood pressure, the suitability of lifestyle, etc.);

[0151] Rule engine operation: The rule engine uses a series of logical rules based on expert experience and medical standards (for example, if the blood sugar value is > 7 mmol / L, it is recommended to reduce sugar intake) to analyze the patient's multi-dimensional data in real time and give corresponding health management suggestions. The rule engine will automatically adjust the suggestions according to different health status or data changes;

[0152] Personalized recommendation generation: Based on health status assessment results and historical multi-dimensional data, the rule engine automatically generates personalized health management recommendations, including diet recommendations, exercise recommendations, medication recommendations, and regular check-up recommendations, including:

[0153] Dietary advice: Provide personalized dietary advice (such as controlling sugar intake, increasing fiber intake, etc.) based on the patient's health status (such as blood sugar level, weight, etc.).

[0154] Exercise recommendations: Provide appropriate exercise plans based on the patient's physiological condition, lifestyle, etc. (such as recommending low-intensity aerobic exercise, stretching exercises, etc.).

[0155] Medication recommendations: Provide appropriate medication dosage and frequency recommendations based on the patient's health status and medication history, combined with medical standards and guidelines.

[0156] Regular check-up recommendations: Automatically recommend regular check-up items and check-up frequency (such as blood sugar monitoring, blood pressure testing, etc.) based on the patient's health status.

[0157] The dynamic health risk prediction module includes:

[0158] Data reception: The dynamic health risk prediction module receives the patient's real-time lifestyle data, real-time environmental data, and the patient's historical multi-dimensional data (including historical physiological data, historical health questionnaire data, historical lifestyle data, and historical environmental data);

[0159] Time series analysis model construction and training: Based on the patient's real-time lifestyle data, real-time environmental data, and historical multi-dimensional data of the patient, a time series analysis model is constructed through a long short-term memory network, and the time series analysis model is trained through the patient's historical multi-dimensional data;

[0160] The time series analysis model models time series data using the following formula:

[0161] ;

[0162] in, is the hidden state at the current moment, For input data (including patients' historical physiological data, real-time lifestyle data, etc.), is the weight matrix, is the weight matrix of the hidden state at the previous moment, is the bias term, As the activation function, by training the LSTM model, the trend of patients’ future health risk changes can be predicted;

[0163] Health risk prediction: Through the trained time series analysis model, the dynamic health risk prediction module predicts the patient's future health risks (such as cardiovascular disease risk, diabetes risk, etc.) based on real-time lifestyle data and real-time environmental data, and generates corresponding health risk scores.

[0164] The dynamic health risk prediction module also includes:

[0165] Health risk assessment and warning generation: Compare the prediction results of the time series analysis model with the preset health risk threshold to generate personalized health risk warnings;

[0166] For each patient, the health risk score output by the model will be compared with the preset health risk threshold. If the score exceeds the set threshold, the system will trigger a health risk warning. For example, if the patient's cardiovascular risk score is higher than the set threshold, the system will issue a health warning, prompting the patient to take necessary interventions or consult a professional doctor;

[0167] Risk prediction result display: The prediction results are displayed through the visualization module, including health risk score, risk level and related health warning information.

[0168] Risk level classification: Based on the health risk prediction results, health risks are classified into low risk, medium risk and high risk;

[0169] Low risk: The health risk score predicted by the model is less than 0.3, and the patient is classified as a low-risk state;

[0170] Medium risk: The model-predicted health risk score is between 0.3 and 0.7, and the patient is classified as a medium-risk state;

[0171] High Risk: The model-predicted health risk score is higher than 0.7 and the patient is classified as a high-risk state.

[0172] The visualization modules include:

[0173] Health status assessment result display: The visualization module displays the health status assessment results generated by the health status assessment module in the form of charts, graphics and text, including:

[0174] Display content: The display content of the health status assessment report includes the patient's physiological status (such as blood sugar, blood pressure, weight, etc.), psychological status (such as emotional assessment, anxiety level, etc.), and behavioral habits (such as exercise volume, eating habits, etc.);

[0175] Display method: Use radar charts to display various indicators of physiological status, psychological status and behavioral habits, bar charts to display the changing trends of different indicators, and text labels to provide specific values ​​and standard range descriptions, so that users can intuitively understand each evaluation result;

[0176] Personalized health management suggestion display: The visualization module displays the patient's diet, exercise, medication and regular check-up suggestions based on the personalized health management suggestion content generated by the personalized health management suggestion module, including:

[0177] Display content: Specific dietary recommendations (e.g., increasing fruit and vegetable intake), exercise recommendations (e.g., daily walking goals, recommended types of exercise), medication recommendations (e.g., adjusting medication dosages), and regular checkup recommendations (e.g., annual physicals);

[0178] Presentation method: Use a list to present the specific content of each suggestion, and use icons (such as exercise icons, medicine bottle icons) to improve operability; at the same time, provide a clickable button next to each suggestion to allow users to set reminders or track execution status;

[0179] Health risk prediction result display: The visualization module displays the prediction results of the dynamic health risk prediction module in the form of charts and text, including:

[0180] Display content: Display the patient's future health risk level (such as low risk, medium risk, high risk) and its changing trend (based on time series prediction).

[0181] Display method: Use a line chart to show the trend of health risk prediction scores over time, and use a bar chart to show the intervals of different health risk levels (such as score < 0.3 for low risk, 0.3 ≤ score < 0.7 for medium risk, and score ≥ 0.7 for high risk). At the same time, color coding (for example, green for low risk, yellow for medium risk, and red for high risk) is used to help users quickly identify risk levels;

[0182] Interactive interface and user operation: The visualization module provides real-time viewing and operation functions through an interactive interface, including:

[0183] Viewing function: Users (such as patients, doctors, and health managers) can view health status assessment reports, personalized health management recommendations, and health risk prediction results through the interface. The interface supports sliding to switch between different modules and zooming in and out of certain detail charts to ensure the readability and visibility of information.

[0184] Operational functions: Users can operate health management suggestions in the interface, such as setting health reminders (such as diet reminders, exercise goal reminders, etc.), recording daily health data (such as exercise volume, blood sugar level, etc.), and synchronizing the recorded data with the system in real time. At the same time, doctors and health managers can add notes or adjust health management strategies through the interface.

[0185] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0186] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A health status assessment system for disease follow-up, characterized in that: It includes data acquisition module, data preprocessing module, health status assessment module, personalized health management suggestion module, dynamic health risk prediction module and visualization module, among which; The data collection module is used to collect multi-dimensional data of the patient in real time, including physiological data, clinical data, health questionnaire data, lifestyle data and environmental data, and identify and integrate different types of data through data fusion technology to form a patient's health record; The data preprocessing module is used to perform denoising, outlier detection, missing data filling and standardization on the collected multi-dimensional data; The health status assessment module is used to combine the collected multi-dimensional data, based on a multi-level feature extraction method, through a health status assessment model, comprehensively assess the patient's current health status, including physiological status, psychological status and living habits, and generate a patient health status assessment report; The personalized health management suggestion module generates personalized health management suggestions through a rule engine based on the health status assessment results and the patient's historical multi-dimensional data; The dynamic health risk prediction module predicts the patient's future health risks based on the patient's historical multi-dimensional data, real-time lifestyle data, real-time environmental data and health status assessment results, combined with the time series analysis model, and generates personalized health warnings; The visualization module is used to display the health status assessment results, personalized health management suggestions and health risk prediction results in the form of charts, graphs and text, and provide them to users for real-time viewing and operation through an interactive interface.

2. A health status assessment system for disease follow-up according to claim 1, characterized in that: The data acquisition module comprises: Physiological data collection: The patient's physiological data is collected in real time through monitoring equipment, including body temperature, blood pressure, heart rate, blood oxygen, weight and blood sugar. The monitoring equipment includes smart bracelets, electronic blood pressure monitors, heart rate sensors, finger-clip pulse oximeters, weight scales and blood glucose meters; Clinical data collection: Access medical databases through API interfaces to obtain patients' clinical data, including patients' diagnosis records, blood test results, imaging data, and treatment history; Health questionnaire data collection: The health questionnaire data of patients are collected through a questionnaire survey platform or face-to-face interviews. The health questionnaire data includes the patient's self-reported symptoms, medical history, psychological state and quality of life assessment; Lifestyle data collection: real-time collection of patients’ lifestyle data through smart devices or applications, including eating habits, exercise, sleep quality, smoking and drinking behavior data; Environmental data collection: Environmental data of the patient's environment is collected through environmental sensors and smart devices. Environmental data includes air quality, temperature and humidity, noise, and pollutant concentration; Data fusion and integration: During the data collection process, the cosine similarity algorithm is used to integrate data from different sources to form the patient's health record.

3. A health status assessment system for disease follow-up according to claim 2, characterized in that: The data preprocessing module comprises: De-noising: De-noising the collected multi-dimensional data through Kalman filtering; Outlier detection: Use the Z-Score method to detect outliers in multi-dimensional data; Missing data filling: fill in missing data in multi-dimensional data; Standardization: Use the minimum-maximum standardization method to standardize multi-dimensional data; Data consistency and comparability verification: Perform consistency verification on data that has undergone denoising, outlier detection, missing data filling, and standardization.

4. A health status assessment system for disease follow-up according to claim 3, characterized in that: The health status assessment module includes: Multi-level feature extraction: Extract features from pre-processed multi-dimensional data, including physiological data, health questionnaire data, and lifestyle data; Feature selection: L1 regularization method was used to perform feature selection on the extracted physiological data, health questionnaire data, and lifestyle data; Feature fusion: features extracted from physiological data, health questionnaire data and lifestyle data are fused to obtain a unified feature vector; Model input and evaluation: The feature vector after feature selection and feature fusion is used as input and input into the health status assessment model for health status assessment, and the patient's health assessment report is output.

5. A health status assessment system for disease follow-up according to claim 4, characterized in that: The model input and evaluation include: Health status assessment model construction: A health status assessment model is constructed based on a convolutional neural network. The model structure includes multiple convolutional layers, pooling layers, and fully connected layers. Loss function and optimization: Use cross entropy loss function to optimize the parameters of the health status assessment model; Model training: The feature vectors after feature selection and feature fusion are divided into training set and validation set. The training set data is used to train the health status assessment model, and the validation set data is used to verify the parameters of the health status assessment model. Health status assessment results: The trained health status assessment model outputs the patient's health status assessment results based on the input multi-dimensional data, including a comprehensive assessment of the physiological state, psychological state and living habits, and generates a personalized health report.

6. A health status assessment system for disease follow-up according to claim 5, characterized in that: The personalized health management suggestion module generates personalized health management suggestions based on the health status assessment results and the patient's historical multi-dimensional data through a rule engine and a machine learning algorithm, specifically including: Integration of health status assessment results and historical data: The personalized health management suggestion module integrates the health status assessment results generated by the health status assessment module and the patient's historical multi-dimensional data. The integrated data forms a complete health record of the patient; Rule engine application: The personalized health management suggestion module analyzes the integrated complete health records through the rule engine and generates preliminary health management suggestions based on the preset health management rules; Personalized recommendation generation: Based on health status assessment results and historical multi-dimensional data, the rule engine automatically generates personalized health management recommendations, including diet recommendations, exercise recommendations, medication recommendations, and regular check-up recommendations.

7. A health status assessment system for disease follow-up according to claim 6, characterized in that: The dynamic health risk prediction module includes: Data reception: The dynamic health risk prediction module receives the patient's real-time lifestyle data, real-time environmental data, and the patient's historical multi-dimensional data; Time series analysis model construction and training: Based on the patient's real-time lifestyle data, real-time environmental data, and historical multi-dimensional data of the patient, a time series analysis model is constructed through a long short-term memory network, and the time series analysis model is trained using the patient's historical multi-dimensional data; Health risk prediction: Through the trained time series analysis model, the dynamic health risk prediction module predicts the patient's future health risks based on real-time lifestyle data and real-time environmental data, and generates corresponding health risk scores.

8. A health status assessment system for disease follow-up according to claim 7, characterized in that: The dynamic health risk prediction module also includes: Health risk assessment and warning generation: Compare the prediction results of the time series analysis model with the preset health risk threshold to generate personalized health risk warnings; Risk prediction result display: The prediction results are displayed through the visualization module, including health risk score, risk level and related health warning information. Risk level classification: Based on the health risk prediction results, health risks are divided into low risk, medium risk and high risk.

9. A health status assessment system for disease follow-up according to claim 8, characterized in that: The visualization module comprises: Health status assessment result display: The visualization module displays the health status assessment results generated by the health status assessment module in the form of charts, graphs and text; Personalized health management suggestion display: The visualization module displays the patient's diet, exercise, medication and regular check-up suggestions based on the personalized health management suggestion content generated by the personalized health management suggestion module; Health risk prediction result display: The visualization module displays the prediction results of the dynamic health risk prediction module in the form of charts and text; Interactive interface and user operation: The visualization module provides real-time viewing and operation functions through an interactive interface.

Images