Health monitoring system and method for hemodialysis patient

Abstract

The invention relates to the technical field of health monitoring, and discloses a hemodialysis patient health monitoring system and method.Based on basic information, previous dialysis data and state parameters of a target patient, on the premise of comprehensively considering dialysis safety and effectiveness reflected by the state parameters, the hemodialysis patient health monitoring system and method are provided. According to the method, the relation between the dialysis parameters and the dialysis monitoring indexes of the target patient can be learned in the whole dialysis treatment process of the target patient, so that the dialysis parameters which can be adopted in the future dialysis process of the target patient are predicted through the personalized dialysis parameter prediction model, effective dialysis parameter selection suggestions are provided for medical staff, and the dialysis treatment efficiency is improved. Therefore, the dialysis efficiency and the dialysis quality are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of health monitoring, and in particular to a health monitoring system and method for hemodialysis patients. Background Art

[0002] Hemodialysis is the main means of treating patients with end-stage renal disease. It maintains electrolyte balance and metabolic stability in patients by removing excess fluid and metabolic waste from the body.

[0003] However, existing dialysis systems usually use fixed dialysis parameters, which are difficult to dynamically adjust according to the individual differences of different patients, and lack feedback on real-time data during the dialysis process. Once the dialysis parameters are set, they usually remain unchanged throughout the dialysis process, which can easily lead to restrictions on dialysis monitoring indicators, especially when the patient's physiological state fluctuates, and adjustments may not be made in time. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a health monitoring system and method for hemodialysis patients to solve the current problems of being unable to achieve personalized prediction and suggestions for dialysis parameters during dialysis treatment for patients, and being unable to dynamically adjust dialysis parameters based on evidence during the dialysis process.

[0005] The first aspect of the present invention discloses a health monitoring system for hemodialysis patients, the system comprising a data acquisition and recording module, a personalized prediction module, an adjustment module and a real-time calculation module; Among them, the data acquisition and recording module is used to collect and record the patient's status parameters, dialysis data and basic information during the dialysis process; The personalized prediction module is used to establish a personalized dialysis parameter prediction model based on the patient's historical dialysis data, status parameters and basic information of the patient using a machine learning algorithm, and predict the patient's dialysis parameters based on the model to obtain a first dialysis parameter; The adjustment module is used for the medical staff to adjust the first dialysis parameter and determine the adjusted second dialysis parameter; The real-time calculation module is used to calculate the dialysis monitoring index based on the second dialysis parameter and the state parameter collected in real time when the dialysis process is performed based on the second dialysis parameter, and feed the calculated dialysis monitoring index back to the personalized prediction module; The personalized prediction module adjusts the model based on the dialysis monitoring indicators fed back by the real-time calculation module.

[0006] Further, the state parameters include the total amount of water in the patient's body, electrolyte concentration, urea concentration, blood pressure, heart rate, biomarker data, hemoglobin level, creatinine level, electrocardiogram data, blood oxygen saturation, dialysate temperature, dialysate pH and dialysate electrolyte concentration; wherein the biomarker data include β2 microglobulin, potassium, calcium, and phosphorus; The dialysis parameters include blood flow rate, dialysate flow rate, dialysis time, and electrolyte composition and concentration of the dialysate.

[0007] Furthermore, the patient's historical dialysis data includes dialysis parameters of the patient's previous dialysis sessions, dialysis monitoring indicators, and the patient's dialysis frequency; The patient's basic information includes age, gender, weight, basal metabolic rate, and comorbidities.

[0008] Furthermore, the personalized dialysis parameter prediction model is a neural network and a deep learning model.

[0009] Furthermore, the personalized dialysis parameter prediction model includes setting a physiological feedback mechanism during the training process to dynamically adjust the weights and output parameters of the model according to the patient's physiological parameters in the state parameters collected in real time.

[0010] Furthermore, the physiological feedback mechanism is determined based on basic information of the patient.

[0011] Furthermore, the system also includes a biomarker monitoring module, which is used to comprehensively evaluate the patient's metabolic state and dialysis clearance effect based on the biomarker detection results.

[0012] Furthermore, the system also includes a dialysis device performance monitoring module for real-time monitoring of the water permeability, membrane resistance and toxin removal capacity of the dialysis device, and predicting the replacement time of each component unit of the dialysis device.

[0013] Furthermore, the system also includes a remote monitoring module for transmitting the patient's real-time and historical dialysis data, status parameters and basic patient information to the cloud platform through the Internet of Things technology, and analyzing and remotely monitoring the data through the cloud platform.

[0014] The second aspect of the present invention discloses a health monitoring method for hemodialysis patients, which is implemented based on the system disclosed in the first aspect, and comprises: Collect and record the patient's status parameters, dialysis data and basic information during dialysis; Based on the patient's historical dialysis data, status parameters and basic information of the patient, a personalized dialysis parameter prediction model is established using a machine learning algorithm, and the patient's dialysis parameters are predicted based on the model to obtain a first dialysis parameter; The medical staff adjusts the first dialysis parameter and determines the adjusted second dialysis parameter; Calculating a dialysis monitoring index based on the second dialysis parameter and a state parameter collected in real time when the dialysis process is performed based on the second dialysis parameter, and feeding the calculated dialysis monitoring index back to the personalized dialysis parameter prediction model; The model was adjusted based on the dialysis monitoring indicators fed back.

[0015] Further, the state parameters include the total amount of water in the patient's body, electrolyte concentration, urea concentration, blood pressure, heart rate, biomarker data, hemoglobin level, creatinine level, electrocardiogram data, blood oxygen saturation, dialysate temperature, dialysate pH and dialysate electrolyte concentration; wherein the biomarker data include β2 microglobulin, potassium, calcium, and phosphorus; The dialysis parameters include blood flow rate, dialysate flow rate, dialysis time, and electrolyte composition and concentration of the dialysate.

[0016] Furthermore, the patient's historical dialysis data includes dialysis parameters of the patient's previous dialysis sessions, dialysis monitoring indicators, and the patient's dialysis frequency; The patient's basic information includes age, gender, weight, basal metabolic rate, and comorbidities.

[0017] Furthermore, the personalized dialysis parameter prediction model is a neural network and a deep learning model.

[0018] Furthermore, the personalized dialysis parameter prediction model includes setting a physiological feedback mechanism during the training process to dynamically adjust the weights and output parameters of the model according to the patient's physiological parameters in the state parameters collected in real time.

[0019] Furthermore, the physiological feedback mechanism is determined based on basic information of the patient.

[0020] Furthermore, the method also includes comprehensively evaluating the patient's metabolic status and dialysis clearance effect based on the biomarker detection results.

[0021] Furthermore, the method also includes real-time monitoring of the water permeability, membrane resistance and toxin removal capacity of the dialysis device, and predicting the replacement time of each component unit of the dialysis device.

[0022] Furthermore, the method also includes transmitting the patient's real-time and historical dialysis data, status parameters and basic patient information to a cloud platform through Internet of Things technology, and analyzing and remotely monitoring the data through the cloud platform.

[0023] Compared with the prior art, the present invention has the following beneficial effects: Based on the basic information of the target patient, the previous dialysis data and the status parameters, the present invention can learn the relationship between the dialysis parameters and the dialysis monitoring indicators of the target patient during the entire dialysis treatment process of the target patient, while ensuring that the dialysis safety and effectiveness reflected by the status parameters are comprehensively considered. Therefore, the dialysis parameters that can be adopted in the future dialysis process of the target patient can be predicted through the personalized dialysis parameter prediction model, and the medical staff can be provided with effective dialysis parameter selection suggestions, thereby improving the dialysis efficiency and dialysis quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of the economic application, and do not constitute a limitation of the embodiments of the present invention. In the drawings: Figure 1 This is a schematic diagram of the structure of a health monitoring system for hemodialysis patients disclosed in Embodiment 1 of the present invention; Figure 2 A schematic flow chart of a health monitoring method for hemodialysis patients disclosed in yet another embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments.

[0026] Embodiment 1 The first aspect of the present invention discloses a health monitoring system for hemodialysis patients. Figure 1 , Figure 1 It is a structural schematic diagram of a health monitoring system for hemodialysis patients disclosed in an embodiment of the present invention, the system comprising a data acquisition and recording module, a personalized prediction module, an adjustment module and a real-time calculation module; Among them, the data acquisition and recording module is used to collect and record the patient's status parameters, dialysis data and basic information during the dialysis process.

[0027] The personalized prediction module is used to establish a personalized dialysis parameter prediction model based on the patient's historical dialysis data, status parameters and basic patient information using a machine learning algorithm, and predict the patient's dialysis parameters based on the model to obtain a first dialysis parameter.

[0028] Furthermore, the status parameters include, but are not limited to, the total amount of water in the patient's body, electrolyte concentration, urea concentration, blood pressure, heart rate, biomarker data, hemoglobin level, creatinine level, electrocardiogram data, blood oxygen saturation, dialysate temperature, dialysate pH, and dialysate electrolyte concentration. The biomarker data include, but are not limited to, β2 microglobulin, potassium, calcium, and phosphorus.

[0029] Specifically, in the embodiment of the present invention, the state parameter is a parameter directly or indirectly related to the patient's physiological condition. For example, the dialysate temperature can improve the patient's comfort and reduce the discomfort that may occur during the dialysis process. The pH value of the dialysate is used to maintain the patient's acid-base balance. During the dialysis process, the correct pH value can help adjust the patient's blood acid-base state.

[0030] It is understandable that during the dialysis process, the patient's physiological parameters are comprehensively considered because these parameters can reflect the patient's overall health status and physiological response, which helps to optimize dialysis treatment. Specifically, physiological parameters such as body water, urea concentration, blood pressure, heart rate and biomarker data provide key information about the patient's metabolism, fluid balance and organ function. By real-time monitoring and analysis of these parameters, the dialysis plan can be adjusted in time to ensure the safety and effectiveness of the patient's dialysis process while reducing the risk of complications.

[0031] Furthermore, dialysis parameters include, but are not limited to, blood flow rate, dialysate flow rate, dialysis duration, and electrolyte composition and concentration of the dialysate. The dialysis parameters here refer to key parameters used to adjust and control the dialysis process, and the adjustment of dialysis parameters directly affects the efficiency and safety of dialysis.

[0032] The patient's historical dialysis data includes but is not limited to the dialysis parameters of the patient's previous dialysis, dialysis monitoring indicators, and the patient's dialysis frequency. The patient's basic information includes but is not limited to age, gender, weight, basal metabolic rate, and comorbidities.

[0033] Specifically, the personalized dialysis parameter prediction model in the present invention is established based on the basic information of the target patient, the previous dialysis data and the state parameters. Under the premise of ensuring the comprehensive consideration of the dialysis safety and effectiveness reflected by the state parameters, the relationship between the dialysis parameters of the target patient and the dialysis monitoring indicators can be learned for the entire dialysis treatment process of the target patient, so as to predict the dialysis parameters that can be adopted in the future dialysis process of the target patient through the personalized dialysis parameter prediction model. It can be understood that the basic information corresponding to the patient may change during the previous dialysis processes. Therefore, there is a corresponding relationship between the dialysis data of the previous dialysis in the historical dialysis process, the basic information of the patient and the state parameters.

[0034] Further, the adjustment module is used for the medical staff to adjust the first dialysis parameter and determine the adjusted second dialysis parameter. It is understandable that the patient's dialysis course usually includes several specific dialysis processes, and the specific dialysis parameters used in each dialysis process are different. The output of the personalized dialysis parameter prediction model is the dialysis parameter recommended for a future dialysis process. However, in view of the fact that the adoption of the dialysis parameter is related to the patient's physiological parameter changes and physical condition, therefore, in order to ensure safety, the first dialysis parameter is only used as a recommended parameter, and it needs to be further determined by the medical staff before it can be used as the dialysis parameter used in actual dialysis. However, it should be noted that the second dialysis parameter can be the same as the first dialysis parameter. With the confirmation of the medical staff, the first dialysis parameter output by the model can be directly used as the parameter used during the dialysis operation.

[0035] Furthermore, the real-time calculation module is used to calculate the dialysis monitoring index based on the second dialysis parameter and the state parameter collected in real time when the dialysis process is performed based on the second dialysis parameter, and feed the calculated dialysis monitoring index back to the personalized prediction module.

[0036] Specifically, in an embodiment of the present invention, the dialysis monitoring index includes but is not limited to dialysis efficiency, electrolyte clearance rate, and liquid state. Among them, the dialysis efficiency can be calculated based on the urea concentration before and after dialysis, the dialysis time, and the blood flow rate. The electrolyte clearance rate can be calculated based on the electrolyte concentration in the dialysate, the electrolyte concentration in the patient's blood, the dialysate flow rate, and the dialysis time. The liquid state can be evaluated by monitoring the changes in the patient's heart rate and blood pressure, and combining the liquid load formula. The present invention does not specifically limit the specific dialysis monitoring index calculated by the real-time calculation module and the calculation method corresponding to the index.

[0037] Furthermore, the personalized prediction module adjusts the model based on the dialysis monitoring indicators fed back by the real-time calculation module.

[0038] Furthermore, the personalized dialysis parameter prediction model is a neural network and a deep learning model.

[0039] Furthermore, the personalized dialysis parameter prediction model includes setting a physiological feedback mechanism during the training process to dynamically adjust the weights and output parameters of the model according to the patient's physiological parameters in the state parameters collected in real time.

[0040] Specifically, as a preferred implementation of Example 1 of the present invention, the personalized dialysis parameter prediction model adopts an LSTM (Long Short-Term Memory) network structure to predict personalized dialysis parameters of hemodialysis patients by learning the long-term dependencies in the time series data of the patient's previous dialysis processes. The input of the model includes the patient's historical dialysis data, state parameters, and basic information.

[0041] Specifically, the process of establishing a personalized dialysis parameter prediction model includes: Data preprocessing: Collect and organize the patient's historical dialysis data, status parameters and basic information to form a training data set. After obtaining the training data set, perform normalization on the data. The purpose of this process is to eliminate the impact of dimensions and ensure that different parameters are within the same range to facilitate model learning.

[0042] Model structure design: The dialysis parameter prediction model of the embodiment of the present invention is composed of multiple LSTM layers and fully connected layers. Among them, the input layer receives the preprocessed data, the LSTM layer is used to learn the time series features, and the fully connected layer is used to output the predicted value of the personalized dialysis parameter.

[0043] Choice of loss function: This process uses mean square error (MSE) as the loss function to calculate the error between the predicted value and the actual value, with the aim of guiding the optimization process of the model.

[0044] Furthermore, the training model also includes randomly initializing the weights and bias parameters of the LSTM layer. After the model is initialized, the input data is passed into the model, information is transmitted through the LSTM layer, the hidden state and cell state of the current time step are calculated, and finally, the output of the LSTM layer is passed through the fully connected layer to generate the predicted value of the personalized dialysis parameter. In addition, the loss function is used to calculate the error between the current predicted value and the actual value. The gradient of the loss function to the weights of each layer is calculated by the chain rule, and the model parameters are updated using an optimization algorithm (such as Adam or SGD) to reduce the loss.

[0045] In the model optimization in the embodiment of the present invention, in addition to adjusting the model based on the dialysis monitoring indicators fed back by the real-time computing module, it also includes setting a physiological feedback mechanism during the training process to dynamically adjust the weights and output parameters of the model according to the patient's physiological parameters in the real-time collected status parameters.

[0046] It can be understood that the specific relationship between the physiological parameter range determined in the physiological feedback mechanism and the model parameter adjustment is determined based on the condition of the target patient. The patient's basic physical conditions such as age, weight, complications, and basic medical conditions are different, and the normal physiological parameter ranges that can be tolerated are different. Therefore, the physiological feedback mechanism set for different patients should be different.

[0047] Specifically, as a preferred implementation, the process of dynamically adjusting the model weights is as follows, in combination with the state parameters collected in real time: ; in, is the model weight vector after dynamic adjustment, is the current model weight vector, To adjust the learning rate of the factor, it is used to control the amplitude of weight update. is the current value of the real-time state parameter collected at time step t, specifically representing the patient’s current physiological state, is the value of the historical state parameter collected at time step t, representing the physiological state used by the model during training, is the hidden state at time step t, corresponding to the hidden state in LSTM, is the minimum value of the state parameter, indicating the lower limit of the state parameter in the entire data set. is the maximum value of the state parameter, indicating the upper limit of the state parameter in the entire data set. It means that by normalizing the current state parameters to the range of (0,1), the purpose is to ensure that the impact of feedback adjustment is not too large, preventing the model from overfitting or producing unstable output.

[0048] ; is the feedback function, where is the i-th state parameter related to the patient's physiological state collected at time step t, For The relevant weight coefficient indicates the contribution of each parameter to the feedback. For the i-th The threshold value is used to determine the impact of this parameter on the feedback. The function is used to ensure that the feedback value is between (0,1) and limit the amplitude of the feedback.

[0049] In an embodiment of the present invention, by introducing a physiological feedback mechanism, the model can be adjusted in real time according to the patient's physiological state, that is, the output of the model is adjusted based on the unique physiological condition of each patient, so that the prediction results are more personalized and meet the actual needs of the patient.

[0050] Embodiment 2 Furthermore, based on the content of Example 1, as another embodiment of the present invention, the system also includes a biomarker monitoring module, and the biomarker monitoring module is used to comprehensively evaluate the patient's metabolic state and dialysis clearance effect based on the biomarker detection results.

[0051] Specifically, the biomarker monitoring module performs data preprocessing operations on the collected biomarker data and analyzes the processed data. Based on the analysis results, the changes in multiple biomarkers are comprehensively considered to evaluate the patient's overall metabolic state. For example, changes in electrolyte levels in serum can indicate the state of water balance. The dialysis clearance effect is evaluated by comparing the changes in biomarker concentrations before and after dialysis, such as the removal efficiency of waste products such as urea and creatinine.

[0052] Furthermore, as a preferred embodiment, the evaluation results are fed back to the personalized prediction module for further adjusting the dialysis parameters to ensure the effectiveness and safety of the treatment plan. In addition, real-time suggestions can be provided to medical staff to help them make quick decisions based on the evaluation results.

[0053] Embodiment 3 Furthermore, based on the contents of Example 1 or Example 2, as another embodiment of the present invention, the system also includes a dialysis device performance monitoring module for real-time monitoring of the water permeability, membrane resistance and toxin removal capacity of the dialysis device, and predicting the replacement time of each component unit of the dialysis device.

[0054] In an embodiment of the present invention, a collection operation is performed on the performance parameters of each component unit of the dialysis device, specifically, the water permeability of the dialysis device membrane is monitored in real time by means of sensors, such as using a combination of a flow sensor and a pressure sensor to record the flow parameters in the dialysate and the patient's blood, and calculate the water permeability of the membrane; the resistance of the membrane is calculated by measuring the inflow and outflow pressures of the membrane, and an increase in the membrane resistance indicates an increase in the degree of blockage or contamination of the membrane, which is not conducive to the dialysis process; real-time monitoring of biomarker concentrations is used to compare concentration changes before and after dialysis, and to evaluate the toxin removal efficiency.

[0055] The real-time collected data is compared with the set threshold to evaluate the performance of the dialysis device. For example, when the water permeability of the membrane is lower than a certain standard, the system will issue an alarm to prompt replacement. By comprehensively evaluating various performance indicators, the overall health status of the dialysis device is determined.

[0056] Based on the monitoring data and performance evaluation results, a prediction model is used to calculate the expected replacement time of each component of the dialysis device. Furthermore, this prediction can be dynamically adjusted based on factors such as the frequency of use of the device and the rate of performance degradation.

[0057] Real-time monitoring of dialysis device performance can detect problems in a timely manner, reduce treatment risks caused by equipment failure, and ensure that toxin removal capacity remains within an effective range to optimize dialysis treatment effects. In addition, predicting replacement time also helps with equipment maintenance and management, reducing treatment interruptions caused by equipment failure and improving equipment utilization efficiency.

[0058] Embodiment 4 Furthermore, based on the contents of Example 1, Example 2 or Example 3, as another embodiment of the present invention, the system also includes a remote monitoring module, which is used to transmit the patient's real-time and historical dialysis data, status parameters and patient basic information to the cloud platform through the Internet of Things technology, and analyze and remotely monitor the data through the cloud platform.

[0059] Specifically, the cloud platform receives and stores real-time and historical data, analyzes the data, and extracts valuable information through machine learning and data mining technology, including generating health trend analysis reports, identifying potential abnormalities, and evaluating dialysis effects and the patient's overall health status. Furthermore, medical staff and patients can access this data through a dedicated interface and receive real-time alerts and health advice when necessary, ensuring that patients can be effectively monitored and managed in their daily lives.

[0060] Embodiment 5 The second aspect of the present invention discloses a health monitoring method for hemodialysis patients, see Figure 2 , Figure 2 FIG. 1 is a flow chart of a health monitoring method for a hemodialysis patient disclosed in another embodiment of the present invention, the method comprising: Collect and record the patient's status parameters, dialysis data and basic information during dialysis; Based on the patient's historical dialysis data, status parameters and basic information of the patient, a personalized dialysis parameter prediction model is established using a machine learning algorithm, and the patient's dialysis parameters are predicted based on the model to obtain a first dialysis parameter; The medical staff adjusts the first dialysis parameter and determines the adjusted second dialysis parameter; Calculating a dialysis monitoring index based on the second dialysis parameter and a state parameter collected in real time when the dialysis process is performed based on the second dialysis parameter, and feeding the calculated dialysis monitoring index back to the personalized dialysis parameter prediction model; The model was adjusted based on the dialysis monitoring indicators fed back.

[0061] Further, the state parameters include the total amount of water in the patient's body, electrolyte concentration, urea concentration, blood pressure, heart rate, biomarker data, hemoglobin level, creatinine level, electrocardiogram data, blood oxygen saturation, dialysate temperature, dialysate pH and dialysate electrolyte concentration; wherein the biomarker data include β2 microglobulin, potassium, calcium, and phosphorus; The dialysis parameters include blood flow rate, dialysate flow rate, dialysis time, and electrolyte composition and concentration of the dialysate.

[0062] Furthermore, the patient's historical dialysis data includes dialysis parameters of the patient's previous dialysis sessions, dialysis monitoring indicators, and the patient's dialysis frequency; The patient's basic information includes age, gender, weight, basal metabolic rate, and comorbidities.

[0063] Furthermore, the personalized dialysis parameter prediction model is a neural network and a deep learning model.

[0064] Furthermore, the personalized dialysis parameter prediction model includes setting a physiological feedback mechanism during the training process to dynamically adjust the weights and output parameters of the model according to the patient's physiological parameters in the state parameters collected in real time.

[0065] Furthermore, the physiological feedback mechanism is determined based on basic information of the patient.

[0066] Furthermore, the method also includes comprehensively evaluating the patient's metabolic status and dialysis clearance effect based on the biomarker detection results.

[0067] Furthermore, the method also includes real-time monitoring of the water permeability, membrane resistance and toxin removal capacity of the dialysis device, and predicting the replacement time of each component unit of the dialysis device.

[0068] Furthermore, the method also includes transmitting the patient's real-time and historical dialysis data, status parameters and basic patient information to a cloud platform through Internet of Things technology, and analyzing and remotely monitoring the data through the cloud platform.

[0069] It should be noted that the specific implementation process of the fifth embodiment is similar to that of the first to fourth embodiments, and will not be repeated in this embodiment.

Claims

1. A health monitoring system for hemodialysis patients, characterized in that: The system includes a data collection and recording module, a personalized prediction module, an adjustment module and a real-time calculation module; Among them, the data acquisition and recording module is used to collect and record the patient's status parameters, dialysis data and basic information during the dialysis process; The personalized prediction module is used to establish a personalized dialysis parameter prediction model based on the patient's historical dialysis data, status parameters and basic information of the patient using a machine learning algorithm, and predict the patient's dialysis parameters based on the model to obtain a first dialysis parameter; The adjustment module is used for the medical staff to adjust the first dialysis parameter and determine the adjusted second dialysis parameter; The real-time calculation module is used to calculate the dialysis monitoring index based on the second dialysis parameter and the state parameter collected in real time when the dialysis process is performed based on the second dialysis parameter, and feed the calculated dialysis monitoring index back to the personalized prediction module; The personalized prediction module adjusts the model based on the dialysis monitoring indicators fed back by the real-time calculation module.

2. The health monitoring system for hemodialysis patients according to claim 1, characterized in that: The state parameters include the total amount of water in the patient's body, electrolyte concentration, urea concentration, blood pressure, heart rate, biomarker data, hemoglobin level, creatinine level, electrocardiogram data, blood oxygen saturation, dialysate temperature, dialysate pH and dialysate electrolyte concentration; wherein the biomarker data include β2 microglobulin, potassium, calcium and phosphorus; The dialysis parameters include blood flow rate, dialysate flow rate, dialysis time, and electrolyte composition and concentration of the dialysate.

3. The health monitoring system for hemodialysis patients according to claim 1, characterized in that: The patient's historical dialysis data includes dialysis parameters of the patient's previous dialysis sessions, dialysis monitoring indicators, and the patient's dialysis frequency; The patient's basic information includes age, gender, weight, basal metabolic rate, and comorbidities.

4. The health monitoring system for hemodialysis patients according to claim 3, characterized in that: The personalized dialysis parameter prediction model is a neural network and deep learning model.

5. The health monitoring system for hemodialysis patients according to claim 4, characterized in that: The personalized dialysis parameter prediction model includes setting a physiological feedback mechanism during the training process, and dynamically adjusting the weights and output parameters of the model according to the patient's physiological parameters in the state parameters collected in real time.

6. The health monitoring system for hemodialysis patients according to claim 5, characterized in that: The physiological feedback mechanism is determined based on the patient's basic information.

7. The health monitoring system for hemodialysis patients according to claim 1, characterized in that: The system also includes a biomarker monitoring module, which is used to comprehensively evaluate the patient's metabolic state and dialysis clearance effect based on the biomarker detection results.

8. The health monitoring system for hemodialysis patients according to claim 1, characterized in that: The system also includes a dialysis device performance monitoring module for real-time monitoring of the water permeability, membrane resistance and toxin removal capacity of the dialysis device, and predicting the replacement time of each component unit of the dialysis device.

9. The health monitoring system for hemodialysis patients according to claim 1, characterized in that: The system also includes a remote monitoring module for transmitting the patient's real-time and historical dialysis data, status parameters and basic patient information to a cloud platform through the Internet of Things technology, and analyzing and remotely monitoring the data through the cloud platform.

10. A health monitoring method for hemodialysis patients, the method being implemented based on the system of any one of claims 1 to 9, characterized in that: The method comprises: Collect and record the patient's status parameters, dialysis data and basic information during dialysis; Based on the patient's historical dialysis data, status parameters and basic information of the patient, a personalized dialysis parameter prediction model is established using a machine learning algorithm, and the patient's dialysis parameters are predicted based on the model to obtain a first dialysis parameter; The medical staff adjusts the first dialysis parameter and determines the adjusted second dialysis parameter; Calculating a dialysis monitoring index based on the second dialysis parameter and a state parameter collected in real time when the dialysis process is performed based on the second dialysis parameter, and feeding the calculated dialysis monitoring index back to the personalized dialysis parameter prediction model; The model was adjusted based on the feedback of dialysis monitoring indicators.

Images