Hemodialysis patient home management system based on informatization technology

Abstract

The invention discloses a hemodialysis patient home management system based on an informatization technology, and relates to the technical field of medical informatization, and the system comprises the following components: a data collection and analysis module which collects the information of vital signs, dialysis data, diet intake and living habits of a patient in real time through a wearable device and an intelligent home sensor, according to the method, a decision tree algorithm is utilized to predict dialysis requirements and complication risks, a linear regression model is combined to predict dialysis effects, scientificity and effectiveness of a dialysis scheme are effectively improved, a Q-learning algorithm is introduced to dynamically adjust the dialysis scheme, and the dialysis efficiency is improved. The treatment strategy is continuously optimized according to the real-time feedback of the patient, the comfort of the patient is improved, the long-term effect of dialysis treatment is remarkably improved, the workload of medical staff is relieved, and the self-management ability of the patient is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of medical informatization, and particularly to a home management system for hemodialysis patients based on informatization technology. Background Art

[0002] With the continuous progress of medical technology and the wide application of hemodialysis treatment in the treatment of kidney diseases, the home management of hemodialysis patients has become a difficult problem to solve. As the main alternative treatment method for end-stage kidney diseases, the treatment effect of hemodialysis not only depends on professional treatment in the hospital, but more on the self-management and living habits of patients at home.

[0003] In traditional technologies, there are significant limitations in formulating personalized dialysis plans. Traditional methods often rely on limited health data and doctors' empirical judgments, making it difficult to comprehensively and deeply analyze the individual differences of patients, resulting in the lack of pertinence and scientificity of dialysis plans. This not only limits the improvement of dialysis effects, but also may increase the risk of complications. At the same time, traditional technologies lack a dynamic adjustment mechanism and cannot respond in a timely manner to changes in patients' health conditions, making it difficult for dialysis plans to meet the actual needs of patients.

[0004] Therefore, developing a home management system for hemodialysis patients based on informatization technology to solve the above-mentioned problems can help patients reduce medical burdens and economic pressures, and improve the quality and efficiency of overall medical services. Summary of the Invention

[0005] The purpose of the present invention is to make up for the deficiencies of the existing technology and provide a home management system for hemodialysis patients based on informatization technology. It can integrate modules such as data collection, personalized plan generation, patient feedback adjustment, and remote monitoring and warning, realizing the intelligence and personalization of dialysis treatment. Using advanced algorithms to accurately predict dialysis needs and effects, dynamically optimize treatment strategies, significantly improving treatment effects and patient comfort, enhancing patients' self-management ability, reducing the burden on medical staff, and providing a comprehensive and efficient solution for the home management of hemodialysis patients, which is an important innovation in the field of medical informatization.

[0006] To solve the above technical problems, the present invention provides the following technical solution: A home management system for hemodialysis patients based on informatization technology, which includes the following components:

[0007] Data collection and analysis module: Real-time collect information on patients' vital signs, dialysis data, dietary intake, and living habits through wearable devices and smart home sensors, and use big data analysis technology to form a personal health record of patients in combination with patients' historical data;

[0008] Personalized Dialysis Plan Generation Module: Deeply analyze the patient's personal health record, identify the key factors affecting the patient's dialysis effect, and accordingly customize a dialysis plan for the patient to maximize the satisfaction of the patient's individual needs, improve the dialysis effect, and reduce the risk of complications. Utilize the continuous optimization and learning ability of machine learning algorithms to enable the dialysis plan to be automatically adjusted as the patient's health condition changes;

[0009] Patient Feedback and Plan Adjustment Module: Support patients to feedback their feelings and effects during dialysis through the user interface and mobile application. According to the patient's feedback and real-time monitoring data, automatically combine the doctor's opinions to continuously optimize and adjust the dialysis plan;

[0010] Remote Monitoring and Early Warning Module: Utilize Internet of Things technology to real-time monitor the patient's health condition and dialysis process, and immediately notify the patient and the medical team when abnormalities are found to achieve early intervention and early warning.

[0011] First, the Data Collection and Analysis Module real-time collects information on the patient's vital signs, dialysis data, dietary intake, and living habits through wearable devices and smart home sensors. Dedup, denoise, and fill in missing values for the received raw data, convert data from different sources into a unified format to form a complete health data set. Apply K-means clustering analysis to deeply mine the health data set, discover the correlations between health indicators, compare the clustering results with the patient's historical data, identify the trends and abnormal points of changes in health indicators, provide a basis for formulating personalized dialysis plans, and fuse the real-time collected data with the patient's historical data to form a personal health record.

[0012] Second, the Data Collection and Analysis Module applies K-means clustering analysis to deeply mine the health data set, discovers the correlations between health indicators, selects health indicators that are meaningful for clustering analysis from the data set as features, uses the elbow method to determine the number of K, and randomly selects K data points as the initial centroids μ1, μ2, …, μ K , for each data point x, use Euclidean distance to calculate its distance to each centroid. The Euclidean distance formula is: where: n is the number of features, x j is the j-th feature value of the data point x, μ ij is the j-th feature value of the centroid μ i , assign the data point to the cluster to which the nearest centroid belongs, and recalculate the centroid of each cluster. The centroid is the mean of all data points in the cluster. The mean calculation formula is: where: |C i | is the cluster C iThe number of data points in, repeat the above steps until the centroid no longer changes significantly, analyze the final clustering results, and observe the distribution and characteristics of health indicators in different clusters.

[0013] Again, the personalized dialysis plan generation module selects features that have an important impact on the generation of the dialysis plan from the patient's health record, performs standardization and normalization processing on the extracted features, and based on historical data and the current health status, uses a decision tree to predict the patient's dialysis needs and complication risks, and uses a linear regression equation to predict the expected effect and physiological index changes after the patient's dialysis, so as to guide the formulation of the dialysis plan. According to the patient's comfort level and symptom improvement, the dialysis plan is dynamically adjusted through the Q-learning algorithm to optimize the long-term treatment effect. According to the prediction results of the decision tree and the linear regression model, combined with the optimal strategy found by the Q-learning algorithm, a personalized dialysis plan is generated.

[0014] The personalized dialysis plan generation module uses a decision tree to predict the patient's dialysis needs and complication risks, screens out feature data that is highly correlated with the dialysis needs and complication risks from numerous features, and uses the Gini index indicator to select the best feature as the root node. The calculation formula of the Gini index is: where: y is the set of categories, p k is the probability that category k appears in the dataset. According to the selected feature, the dataset is divided into multiple sub-datasets, and the above steps are repeated for each sub-dataset until the number of samples in the node is too small, then the decision tree model is constructed. The decision tree model is applied to new patient data to predict their dialysis needs and complication risks, providing support for medical decision-making.

[0015] Furthermore, the personalized dialysis plan generation module uses a linear regression equation to predict the expected effect and physiological index changes after the patient's dialysis, extracts features that are associated with the expected effect and physiological index changes after dialysis from the collected data. The general form of the linear regression equation is: y = β0 + β1x1 + β2x2 + … + β n x n + ∈, where: y is the predicted expected effect and physiological index changes after dialysis, x1, x2, …, x n are the selected features, β0 is the intercept, β1, β2, …, β n are the regression coefficients, ∈ is the error term. The regression coefficients are estimated by the least squares method. The linear regression model is applied to new patient data to minimize the sum of the squared errors between the predicted value and the actual value. By inputting the relevant features before dialysis into the linear regression model, the expected effect and physiological index changes after dialysis are predicted, providing a reference for medical decision-making.

[0016] Furthermore, the personalized dialysis plan generation module uses the Q-learning algorithm to dynamically adjust the dialysis plan to clarify the environment of dialysis treatment. The mathematical vector S represents the combination of the patient's current comfort level, various symptom indicators, and relevant parameters of the dialysis plan. The mathematical vector A represents the adjustment operations of the dialysis plan, such as increasing dialysis time, decreasing dialysis time, and adjusting the dialysis fluid flow rate. The reward function R(S, A) is designed according to the patient's comfort level and symptom improvement. If the patient's comfort level improves and the symptoms are significantly improved, a positive reward is given; if the situation deteriorates or there is no obvious change, a negative reward is given. The Q-value table Q(S, A) is initialized to 0, and the update formula of Q-learning is as follows: where: S t and A t represent the state at time t and the action taken respectively. α is the learning rate, and γ is used to weigh the importance of future rewards. R t+1 is the immediate reward obtained after taking action A t in state S t , is the maximum Q-value that can be obtained by taking all possible actions in the subsequent state S t+1 . After each dialysis treatment, according to the patient's new state and reward, the update formula is used to update the Q-value table. Through continuous iteration, the Q-value table gradually converges to find the optimal strategy.

[0017] Furthermore, the patient feedback and plan adjustment module feeds back the feelings and effects during dialysis through the user interface and mobile application, and obtains the patient monitoring data in real time through the connected hemodialysis equipment. The feedback information is analyzed and quantified, and the patient's state and treatment effect are evaluated in combination with the monitoring data. A viewing interface is provided for doctors. Based on this information and their own professional knowledge and experience, doctors give personalized adjustment suggestions. Based on the patient's feedback, monitoring data, and doctors' adjustment suggestions, the initial plan generated by the personalized dialysis plan generation module is evaluated and verified, and the medical team is organized to discuss and determine the final plan, which is communicated to the treatment team, and the feedback and data after the implementation of the new plan are continuously tracked.

[0018] Furthermore, the remote monitoring and warning module of the remote monitoring device collects the patient's vital signs and dialysis parameters in real time, encrypts and transmits them to the cloud, uses the threshold detection algorithm to identify abnormal data, and combines the clinical warning rules to achieve immediate abnormal detection. The abnormal data points are notified to the patient, family members, and medical team in real time, and preliminary treatment suggestions are provided. The patient's state is continuously monitored, and the warning rules and intervention strategies are adjusted according to the feedback to optimize the treatment effect. New technologies are introduced regularly to improve the system intelligence and user experience.

[0019] Finally, the remote monitoring and early warning module uses a threshold detection algorithm to identify abnormal data. By calculating the mean μ and standard deviation σ of the processed data, the data is analyzed. The formula for the mean is: where: x i is the data point, N is the total number of data points, and the formula for the standard deviation is: Based on the data distribution, set the upper threshold as: T upper = μ + 2σ, and the lower threshold as: T lonver = μ - 2σ. For each newly arrived real-time data point x i , compare it with the set threshold. When x > T upper or x < T lower occurs, then this data point is considered an abnormal data point. Mark and record the data points identified as abnormal, and issue an alarm.

[0020] Compared with the prior art, this hemodialysis patient home management system based on information technology has the following beneficial effects:

[0021] First, by using the decision tree algorithm to predict dialysis needs and complication risks, and combining with the linear regression model to predict dialysis effects, the present invention effectively improves the scientificity and effectiveness of the dialysis plan. Introducing the Q-learning algorithm to dynamically adjust the dialysis plan, and continuously optimizing the treatment strategy according to the patient's real-time feedback. For patients, it not only improves the patient's comfort, enhances the patient's self-management ability, significantly improves the long-term effect of dialysis treatment, and reduces the incidence of complications; for hospitals, it reduces the workload of medical staff, improves the overall quality of medical services and patient satisfaction; for society, it reduces the hospitalization rate of hemodialysis patients, reduces medical expenses, saves medical resources, and brings a revolutionary change to the home management of hemodialysis patients.

[0022] Second, by collecting patients' health data in real time and accurately, using advanced algorithms for in-depth mining and analysis, the present invention provides a scientific basis for the formulation of personalized dialysis plans. It also has a dynamic adjustment mechanism to continuously optimize the treatment plan according to the patient's real-time feedback and monitoring data to ensure the maximization of treatment effects. At the same time, the remote monitoring and early warning function effectively improves the timeliness and accuracy of medical services, ensures patient safety, reduces the burden on family members and the medical team, significantly improves the quality of life of hemodialysis patients, reduces the risk of complications, and provides a comprehensive and intelligent solution for the home management of hemodialysis patients.

[0023] Other advantages, objectives and features of the present invention will be set forth to some extent in the following description, and to some extent, will be apparent to those skilled in the art based on the study of the following text, or can be inspired from the practice of the present invention. Description of the Drawings

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0025] Figure 1 It is a flow operation diagram of a home management system for hemodialysis patients based on information technology;

[0026] Figure 2 It is a flowchart of a home management system for hemodialysis patients based on information technology. Detailed Embodiments

[0027] The following will clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0028] Embodiment 1:

[0029] This embodiment provides a home management system for hemodialysis patients based on information technology. The system realizes the comprehensive intelligent management of the home treatment of hemodialysis patients by integrating multiple core modules such as data collection, generation of personalized dialysis plans, real-time feedback and dynamic adjustment of patients, and remote monitoring and early warning. Using advanced algorithms and data analysis technologies, it accurately predicts the dialysis needs and complication risks of patients, scientifically formulates and dynamically optimizes personalized dialysis plans, greatly improves the treatment effect and patient comfort, reduces the workload of medical staff by enhancing the self-management ability of patients, improves the overall medical service quality, and the remote monitoring and early warning function ensures patient safety, providing a reliable guarantee for the home management of hemodialysis patients.

[0030] Data collection and analysis module: Patients wear wearable devices and install smart home sensors at home. These devices collect information on patients' vital signs, dialysis data, dietary intake, and lifestyle habits in real time. After receiving the raw data, duplicate and noisy data are removed first to ensure the accuracy and reliability of the data. Missing values are filled, and interpolation method is used for estimation. Data from different sources are converted into a unified format to form a complete health dataset. K-means clustering analysis is applied to deeply mine the health dataset to discover the correlations between health indicators. Meaningful health indicators for clustering analysis are selected as features from the preprocessed dataset. The elbow method is used to determine the number of K, and K data points are randomly selected as the initial centroids, μ1, μ2, …, μ K , for each data point in the dataset, the Euclidean distance formula is used to calculate its distance to each centroid. The Euclidean distance formula is: And the data point is assigned to the cluster to which the nearest centroid belongs. The centroid of each cluster is recalculated, and the centroid is updated to the mean value of all data points in the cluster. The mean value calculation formula is: Repeat the above steps until the position of the centroid no longer changes significantly. At this time, the clustering is completed. Analyze the final clustering results and observe the distribution and characteristics of health indicators in different clusters. For example, if it is found that the heart rate and blood pressure of a certain type of patients are generally high while the dialysis frequency is low, this indicates that this type of patients may need to increase the dialysis frequency or adjust their diet to reduce water and salt intake. Compare the clustering results with the patients' historical data to identify the trends and abnormal points of changes in health indicators, providing a basis for formulating personalized dialysis plans. Integrate the real-time collected data with the patients' historical data to form a comprehensive personal health record, which not only includes the patients' health data but also the clustering analysis results and data-based health suggestions.

[0031] Personalized dialysis plan generation module: Collect various types of relevant data from the patients' health records, clean and standardize, normalize the data. Select and extract features that are highly relevant to dialysis plan generation, and screen and optimize them using statistical algorithms. Use historical data to build a decision tree model, select split nodes and prune based on the Gini index to predict the dialysis needs and complication risks of new patients. Build a linear regression model, select relevant features and estimate parameters using optimization algorithms to predict the effects after dialysis and changes in physiological indicators. Define the dialysis treatment environment, design a reward function, initialize the Q-table and set relevant factors. Select actions according to the state and record them in each treatment. Use the update formula to find the optimal strategy. Generate a personalized dialysis plan by integrating the results of the above three models, covering parameter settings such as dialysis time, flow rate, and frequency, and feedback it to the patients and medical staff for implementation to achieve precise and effective dialysis treatment.

[0032] Patient Feedback and Plan Adjustment Module: Patients provide feedback on their feelings during dialysis through the user interface and mobile application, including comfort level, pain degree, and symptom changes. Physiological index data of patients, such as blood pressure, heart rate, and dialysis fluid flow, are obtained in real time through connected hemodialysis devices. Text analysis is performed on the feedback information of patients to extract keywords and emotion tags, quantifying subjective feelings into analyzable indicators. Combining monitoring data, statistical and machine learning algorithms are used to quantitatively evaluate the patient's status and treatment effect. A dedicated viewing interface is provided for doctors to display the patient's feedback content, monitoring data, and analysis results. Based on this information and their professional knowledge and experience, doctors give personalized adjustment suggestions. Based on the patient's feedback, monitoring data, and doctors' adjustment suggestions, the initial plan generated by the personalized dialysis plan generation module is evaluated and verified. The medical team is organized for discussion, and multiple opinions are integrated to further optimize and adjust the plan. The finally determined and approved dialysis plan will be used as a new treatment plan. The approved dialysis plan is promptly communicated to the treatment team, and the treatment team implements it according to the new plan while maintaining communication with the patient to ensure the smooth execution of the plan. After the implementation of the new plan, the patient's feedback and monitoring data are continuously tracked to evaluate the treatment effect. If the treatment effect fails to meet the expectations, the steps of patient feedback collection, monitoring data acquisition, and analysis and quantification are restarted, and the dialysis plan is re - formulated and adjusted according to the new situation. Through this continuous feedback loop, it is ensured that the dialysis plan always adapts to the actual situation and needs of the patient.

[0033] Remote Monitoring and Warning Module: Real - time collection of patients' vital sign data and dialysis parameters is carried out through remote monitoring devices such as wearable devices and hemodialysis machines, and transmitted to the cloud server through a secure network channel. After receiving the data, the cloud server cleans and pre - processes the data, removing abnormal data points such as noise and missing values to improve data quality. Threshold detection algorithms are applied to identify abnormal data, and statistical analysis is performed on the processed data. By calculating the mean μ and standard deviation σ of the data, the calculation formula for the mean: The calculation formula for the standard deviation: Analyze the data. According to the patient's historical data and clinical standards, set the upper threshold as: T upper = μ + 2σ and the lower threshold as: T lonver = μ - 2σ. For each newly arrived real - time data point, calculate its relationship with the mean μ and standard deviation σ and compare it with the threshold. If x > T upper and x < T lowerWhen any one of the above occurs, the data point is considered an abnormal data point. When an abnormal data point is detected, it is immediately marked and recorded, and an alarm mechanism is triggered. The alarm information is sent to the patient, family members, and medical team in real time via text message. Continuously monitor the patient's status, dynamically adjust the warning rules and intervention strategies based on new data points and feedback information, monitor the operating status of the device to ensure its stability, regularly introduce new technologies and methods, and continuously improve the intelligence level and user experience of the system.

[0034] Example Two:

[0035] A home management system for hemodialysis patients based on information technology in this example particularly emphasizes the application of the personalized dialysis plan generation module:

[0036] Select features that have an important impact on the generation of dialysis plans from the patient's health record, such as the patient's age, weight, renal function indicators, historical dialysis effect, and current vital signs. Standardize and normalize the extracted features to ensure that data with different dimensions and ranges can participate in model training fairly among numerous features.

[0037] Use a decision tree to predict the patient's dialysis needs and complication risks, screen out feature data highly correlated with dialysis needs and complication risks, and based on the Gini index indicator, select the best feature as the root node. The Gini index calculation formula is: According to the selected features, divide the data set into multiple sub-data sets, repeat the above steps for each sub-data set to construct a decision tree until the number of samples in the node is too small. Use the decision tree model to predict the dialysis needs and complication risks of new patient data to provide support for medical decision-making.

[0038] Use a linear regression equation to predict the expected effect after dialysis and changes in physiological indicators of the patient. Extract features associated with the expected effect after dialysis and changes in physiological indicators from the collected data, such as renal function indicators before dialysis, dialysis duration, and dialysis fluid flow rate. The general form of the linear regression equation is: y = β0 + β1x1 + β2x2 + … + β n x n + ∈. Estimate the regression coefficients by the least squares method to minimize the sum of squared errors between the predicted values and the actual values. Input the relevant features of the patient before dialysis into the linear regression model to predict the expected effect after dialysis and changes in physiological indicators of the patient, providing a reference for medical decision-making.

[0039] Dynamically adjust the dialysis plan through the Q-learning algorithm, clarify the environment of dialysis treatment, represent the patient's current state (such as comfort level, various symptom indicators) and relevant parameters of the dialysis plan with the mathematical vector S, and represent the adjustment operations of the dialysis plan with the mathematical vector A, such as increasing dialysis time, decreasing dialysis time, and adjusting the dialysis fluid flow rate. Design the reward function R(S,A) according to the patient's comfort level and symptom improvement. Give a positive reward when the comfort level improves and the symptoms are significantly improved, and give a negative reward when the situation deteriorates and there is no obvious change. Initialize the Q-value table Q(S,A) to 0. For each dialysis treatment, select an action according to the patient's current state, and observe the immediate reward and the new state obtained, and use the update formula of Q-learning to update the value table: Through continuous iteration, the value table gradually converges to find the optimal dialysis plan adjustment strategy.

[0040] According to the prediction results of the decision tree and linear regression model, combined with the optimal strategy found by the Q-learning algorithm, generate a personalized dialysis plan.

[0041] Example 3:

[0042] Based on Example 1 and Example 2, the specific process steps of the system are as follows:

[0043] 1. Collect and preprocess the patient's multi-dimensional health data in real time to form a complete data set;

[0044] 2. Deeply mine the health data to identify the correlation of health indicators and abnormal points;

[0045] 3. Extract key features from the patient's health record, after standardization processing, use the decision tree and linear regression model to predict the dialysis demand and effect

[0046] 4. According to the patient's comfort level and symptom improvement, dynamically adjust the dialysis plan through the Q-learning algorithm to optimize the long-term treatment effect;

[0047] 5. Combine the predictions and strategies of the decision tree, linear regression, and Q-learning algorithms to generate a personalized dialysis plan that suits the actual situation of the patient;

[0048] 6. Collect the patient's dialysis feelings and real-time monitoring data through the user interface and mobile application, combined with the professional advice of the doctor, to make personalized adjustments and optimizations to the dialysis plan to ensure that the plan adapts to the actual situation of the patient;

[0049] 7. Collect the patient's vital signs and dialysis parameters in real time, through cloud processing and real-time anomaly detection, provide early warning notifications and preliminary treatment suggestions, continuously optimize the early warning rules and intervention strategies, and improve the system intelligence and user experience.

[0050] It will be apparent to those skilled in the art that the present invention is not limited to the details of the above-described exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, in all respects, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the appended claims rather than the above description. Accordingly, all changes that fall within the meaning and scope of the equivalent elements of the claims are intended to be embraced within the present invention. Any reference signs in the claims should not be construed as limiting the claims concerned.

Claims

1. A home management system for hemodialysis patients based on information technology, characterized in that: The system consists of the following components: Data collection and analysis module: collects patients’ vital signs, dialysis data, dietary intake and living habits in real time through wearable devices and smart home sensors, and uses big data analysis technology and historical data of patients to form personal health records of patients; Personalized dialysis plan generation module: conducts in-depth analysis of the patient's personal health records, identifies key factors that affect the patient's dialysis effect, and customizes dialysis plans for the patient accordingly, to maximize the patient's individual needs, improve dialysis effects, and reduce the risk of complications. The continuous optimization and learning capabilities of machine learning algorithms enable the dialysis plan to automatically adjust as the patient's health status changes; Patient feedback and plan adjustment module: Supports patients to provide feedback on their feelings and effects during dialysis through the user interface and mobile applications. Based on patient feedback and real-time monitoring data, the dialysis plan is automatically optimized and adjusted in combination with the doctor's opinions. Remote monitoring and early warning module: Use IoT technology to monitor the patient's health status and dialysis process in real time, and immediately notify the patient and medical team if any abnormality is found, so as to achieve early intervention and early warning.

2. The home management system for hemodialysis patients based on information technology according to claim 1, characterized in that: The data collection and analysis module collects the patient's vital signs, dialysis data, dietary intake and living habits in real time through wearable devices and smart home sensors, removes duplication, denoises and fills missing values ​​in the received raw data, converts data from different sources into a unified format to form a complete health data set, applies K-means clustering analysis to deeply mine the health data set, discovers the correlation between health indicators, compares the clustering results with the patient's historical data, identifies the trend and abnormal points of health indicator changes, provides a basis for formulating personalized dialysis plans, and integrates the real-time collected data with the patient's historical data to form a personal health record.

3. The home management system for hemodialysis patients based on information technology according to claim 2 is characterized in that: The data collection and analysis module uses K-means clustering analysis to analyze the health data set. Perform deep mining to find the correlation between health indicators, select health indicators that are meaningful for cluster analysis from the data set as features, use the elbow rule to determine the number of K, and randomly select K data points as the initial centroids μ1,μ2,…,μ K , for each data point x, use Euclidean distance to calculate its distance to each centroid. The Euclidean distance formula is: Where: n is the number of features, x j is the jth eigenvalue of data point x, μ ij is the center of mass μ i The j-th eigenvalue of is used to assign the data point to the cluster to which the nearest centroid belongs, and the centroid of each cluster is recalculated. The centroid is the mean of all data points in the cluster. The mean calculation formula is: Where: |C i | is cluster C i Repeat the above steps until the centroid no longer changes significantly, analyze the final clustering results, and observe the distribution and characteristics of health indicators in different clusters.

4. The home management system for hemodialysis patients based on information technology according to claim 1, characterized in that: The personalized dialysis plan generation module selects features that have an important impact on the generation of the dialysis plan from the patient's health record, standardizes and normalizes the extracted features, and uses a decision tree to predict the patient's dialysis needs and complication risks based on historical data and current health conditions. The linear regression equation is used to predict the patient's expected effects and changes in physiological indicators after dialysis to guide the formulation of the dialysis plan. According to the patient's comfort and symptom improvement, the dialysis plan is dynamically adjusted through the Q-learning algorithm to optimize the long-term treatment effect. According to the prediction results of the decision tree and linear regression model, combined with the optimal strategy found by the Q-learning algorithm, a personalized dialysis plan is generated.

5. The home management system for hemodialysis patients based on information technology according to claim 4 is characterized in that: The personalized dialysis plan generation module uses a decision tree to predict the patient's dialysis needs and complication risks, selects feature data that are highly correlated with dialysis needs and complication risks from a large number of features, and uses the Gini index indicator to select the best feature as the root node. The calculation formula of the Gini index is: Where: y is the category set, p k is the probability of category k appearing in the data set. According to the selected features, the data set is divided into multiple sub-datasets. The above steps are repeated for each sub-dataset until the number of samples in the node is too small. Then the decision tree model is constructed. The decision tree model is applied to new patient data to predict their dialysis needs and complication risks, providing support for medical decision-making.

6. The home management system for hemodialysis patients based on information technology according to claim 4, characterized in that: The personalized dialysis plan generation module uses a linear regression equation to predict the expected effect and physiological index changes of the patient after dialysis, and extracts features associated with the expected effect and physiological index changes after dialysis from the collected data. The general form of the linear regression equation is: y=β0+β1x1+β2x2+…+β n x n +∈, where: y is the predicted expected effect and changes in physiological indicators after dialysis, x1, x2,…, x n is the selected feature, β0 is the intercept, β1,β2,…,β n is the regression coefficient, ∈ is the error term, and the regression coefficient is estimated by the least squares method. The linear regression model is applied to new patient data to minimize the sum of square errors between the predicted value and the actual value. The relevant characteristics before dialysis are input into the linear regression model to predict the expected effect and changes in physiological indicators after dialysis, providing a reference for medical decision-making.

7. The home management system for hemodialysis patients based on information technology according to claim 4 is characterized in that: The personalized dialysis plan generation module dynamically adjusts the dialysis plan to clarify the environment of dialysis treatment through the Q-learning algorithm, uses a mathematical vector S to represent the patient's current comfort level, various symptom indicators, and a combination of relevant parameters of the dialysis plan, and uses a mathematical vector A to represent the adjustment operation of the dialysis plan, such as increasing the dialysis time, reducing the dialysis time, and adjusting the dialysate flow rate. The reward function R(S, A) is designed according to the patient's comfort and symptom improvement. If the patient's comfort improves and the symptoms are significantly improved, a positive reward is given, and if the situation deteriorates and there is no obvious change, a negative reward is given. The Q value table Q(S, A) is initialized to 0, and the update formula of Q-learning is: Where: S t and A t They represent the state and action taken at time t, α is the learning rate, γ is used to weigh the importance of future rewards, and R t+1 In state S t Take action A t After receiving the instant reward, In the subsequent state S t+1 The maximum Q value that can be obtained by taking all possible actions. After each dialysis treatment, the Q value table is updated using the update formula according to the patient's new state and reward. Through continuous iteration, the Q value table gradually converges to find the optimal strategy.

8. The information technology-based home management system for hemodialysis patients according to claim 1, characterized in that: The patient feedback and regimen adjustment module uses a user interface and mobile application to provide feedback on the effects felt during dialysis, obtains patient monitoring data in real time through the connected hemodialysis equipment, analyzes and quantifies the feedback information, evaluates the patient's status and treatment effects in combination with the monitoring data, and quantitatively evaluates the patient's status and treatment effects, providing a viewing interface for doctors. Based on this information, combined with their own professional knowledge and experience, the doctor gives personalized adjustment suggestions, based on the patient's feedback, monitoring data and the doctor's adjustment suggestions, evaluates and verifies the initial plan generated by the personalized dialysis regimen generation module, organizes the medical team to discuss and determine the final plan, communicates the plan to the treatment team, and continuously tracks the feedback and data after the implementation of the new plan.

9. The home management system for hemodialysis patients based on information technology according to claim 1, characterized in that: The remote monitoring and early warning module of the remote monitoring device collects the patient's vital signs and dialysis parameters in real time, encrypts and transmits them to the cloud, applies the threshold detection algorithm to identify abnormal data, combines clinical early warning rules to achieve instant anomaly detection, and notifies the patient, family members and medical team of abnormal data points in real time, and provides preliminary treatment suggestions, continuously monitors the patient's status, adjusts the early warning rules and intervention strategies based on feedback to optimize the treatment effect, and regularly introduces new technologies to improve system intelligence and user experience.

10. The information technology-based home management system for hemodialysis patients according to claim 9, characterized in that: The remote monitoring and early warning module uses a threshold detection algorithm to identify abnormal data, and analyzes the data by calculating the mean μ and standard deviation σ of the processed data. The calculation formula of the mean is: Where: x i is the data point, N is the total number of data points, and the standard deviation is calculated as: Based on the distribution of data, the upper threshold is set to: T upper =μ+2σ, the lower limit of the threshold is: T lonver =μ-2σ, for each newly arrived real-time data point x i , compare it with the set threshold, and x>T upper and x <T lower When one of the items is found, the data point is considered to be an abnormal data point, and the data point identified as abnormal is marked and recorded, and an alarm is issued.

Images