Patient health monitoring method and system based on hemodialysis data analysis
By constructing a correlation data set of hemodialysis data and using clustering algorithms to analyze patient characteristics and predict health scores, the problem of traditional systems failing to effectively utilize multidimensional data is solved, personalized health monitoring and treatment suggestions are realized, and the Data usage efficiency.
Patent Information
- Application Number
- CN202411224477.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-03
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-09-03
AI Technical Summary
Traditional hemodialysis data monitoring systems fail to effectively use modern data analysis technology to deeply explore and correlate complex multidimensional data, and fail to fully consider the individual differences of patients and the impact of daily routines on treatment effects, resulting in inefficient data use. .
By collecting patient physical data and hemodialysis data, a correlation data set is constructed, patient characteristics are analyzed using clustering algorithms, improvement values are calculated, health scores are predicted, and data encryption transmission and report sending is carried out based on score adjustment suggestions.
Personalized health monitoring and treatment recommendations for patients are achieved, the efficiency of data usage is improved, targeted treatment and analysis are ensured, and improvement suggestions are dynamically adjusted to adapt to changes in patient conditions.
Smart Images

Figure CN118737473B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of hemodialysis data analysis, and in particular to a patient health monitoring method and system based on hemodialysis data analysis. Background Art
[0002] In the field of modern medicine, hemodialysis has become one of the main treatment methods for patients with chronic renal failure to maintain life. With the development of medical technology and the popularization of smart devices, data collection and analysis during the dialysis process have become more efficient and accurate. By monitoring multiple physiological indicators of patients during the dialysis process, valuable reference information can be provided.
[0003] However, traditional dialysis data monitoring systems mostly use linear analysis methods, and fail to effectively use modern data analysis technology to conduct in-depth mining and correlation analysis of complex multidimensional data. They often fail to fully consider the impact of individual differences in patients and their daily routines on treatment outcomes, lack comprehensive consideration of factors such as living habits, amount of exercise, and rest time, and do not fully utilize the data during the dialysis process to give recommendations, thus reducing the efficiency of data use. Summary of the invention
[0004] In view of the problems existing in the above-mentioned existing patient health monitoring methods and systems based on hemodialysis data analysis, the present invention is proposed.
[0005] Therefore, the problem to be solved by the present invention is that traditional dialysis data monitoring systems mostly use linear analysis methods, fail to effectively use modern data analysis technology to conduct in-depth mining and correlation analysis of complex multidimensional data, and often fail to fully consider the individual differences of patients and the impact of their daily routines on the treatment effect. There is a lack of comprehensive consideration of factors such as living habits, exercise volume and rest time, and the data during the dialysis process are not fully utilized to give suggestions, which reduces the efficiency of data use.
[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a patient health monitoring method based on hemodialysis data analysis, comprising:
[0007] Collect patient physical data and hemodialysis data, analyze the correlation values between the data and construct a correlation data set;
[0008] Perform clustering calculations based on patient-related data sets using clustering algorithms, and analyze patient feature set data of different clusters;
[0009] Based on the patient feature set data of different clusters, analyze the patient's physical data and calculate the improvement value, predict the patient's health score, analyze the improvement value to judge the patient's physical condition, and adjust the improvement suggestions according to the patient's physical condition;
[0010] Data is encrypted and transmitted based on different clustered patients, and reports are generated and sent.
[0011] As a preferred solution of the patient health monitoring method based on hemodialysis data analysis of the present invention, wherein: the collecting of patient physical data and hemodialysis data, analyzing the correlation values between the data to construct a correlation data set includes:
[0012] The patient's physical data is collected through health monitoring equipment, including exercise data and rest data. The exercise data includes the total number of steps, exercise time, and calorie consumption value, and the rest data includes sleep duration and sleep quality index;
[0013] Collect patients’ hemodialysis data through medical devices, including weight, blood pressure, heart rate before and after dialysis, as well as dialysis duration and dialysis efficiency;
[0014] Synchronize the collected physical data and hemodialysis data with the same time tag;
[0015] Standardize physical and hemodialysis data;
[0016] The Pearson correlation coefficient was used to calculate the correlation between each physical data and hemodialysis data, expressed as: ;in, represents the correlation value of the xth item of physical data and the yth item of hemodialysis data, represents the ith collected value of the xth item of physical data, represents the i-th collected value of the y-th hemodialysis data, n represents the total number of collected values; i represents the collected value index, represents the mean of the x-th item of body data, represents the mean of the y-th item of hemodialysis data;
[0017] The analysis was performed by testing based on the calculated correlation values, expressed as: ;in, Indicates the test value of the x-th physical data and the y-th hemodialysis data; the correlation value greater than the test value Perform statistics to form a correlation set R.
[0018] As a preferred solution of the patient health monitoring method based on hemodialysis data analysis of the present invention, wherein: the clustering calculation is performed according to the patient correlation data set by a clustering algorithm, and the patient feature set data of different clusters are analyzed, including:
[0019] Based on the correlation set R of different patients, the physical data and dialysis parameters corresponding to the statistically significant correlation coefficients are selected to construct the feature set J;
[0020] Use the K-means clustering algorithm to randomly select patient records from the patient record set as the initial cluster centers , define clustering represents the i-th cluster, cluster Including all the cluster centers assigned to the initial cluster Patient characteristics set J;
[0021] Based on each patient’s feature set J, calculate the initial cluster center The Euclidean distance of is used to perform initial clustering, which is expressed as: ;in, represents the jth feature of patient feature set J, Represents the cluster center The jth feature of the patient feature set J, m represents the total number of features of the patient feature set J, Represents the patient feature set J and cluster center The Euclidean distance of
[0022] Based on the initial clustering, update the cluster center to the cluster to which it belongs The average position of all patient feature sets J in is expressed as: ;in, Representation clustering The number of patient records in represents the updated cluster center;
[0023] Calculate the total squared error, expressed as: ; where SSE represents the total square error, Represents the patient feature set J to its cluster center The square of the distance, k represents the number of clusters;
[0024] Repeated calculation and update of cluster centers Calculate SSE, and stop the iteration when the update transformation of the cluster center is not obvious;
[0025] Apply the elbow rule to determine the optimal number of clusters, repeat the K-means clustering algorithm based on different k values, count the SSE of each k value, and draw a curve. According to the elbow rule of the curve, the k value with the SSE value is determined as the optimal k value;
[0026] The patient feature set data of different clusters were analyzed based on the calculation of k value.
[0027] As a preferred solution of the patient health monitoring method based on hemodialysis data analysis of the present invention, wherein: the patient feature set data based on different clusters, analyzing the patient's physical data and calculating the improvement value, includes:
[0028] Physical data collected from patients based on different clusters included calorie consumption data and rest duration data;
[0029] Calculate the mean of calorie consumption data for patients in the same cluster , and calculate the standard deviation of the calorie consumption data of patients in the same cluster ;
[0030] The activity improvement value was calculated for each patient and expressed as: ;in, represents the activity improvement value of the i-th patient, represents the calorie consumption data of the i-th patient;
[0031] Calculate the mean of rest time data of patients in the same cluster , and calculate the standard deviation of the rest time data of patients in the same cluster ;
[0032] The improvement in rest volume was calculated for each patient and expressed as: ;in, represents the improvement value of rest volume of the i-th patient, Represents the rest time data of the i-th patient.
[0033] As a preferred embodiment of the patient health monitoring method based on hemodialysis data analysis of the present invention, the predicting of the patient health score and analyzing the improvement value to judge the patient's physical condition include:
[0034] According to the patients in different clusters, a feedforward neural network model based on a fully connected layer was constructed, including an input layer, a hidden layer, and an output layer;
[0035] The loss is calculated using the mean square error loss function, and the model parameters are iteratively optimized using the Adam optimizer and the gradient descent method. When the loss of the feedforward neural network model no longer decreases during the continuous iteration process, the iteration is stopped to output the model parameters and update the feedforward neural network model.
[0036] Calculate health score based on input patient physical data and hemodialysis data;
[0037] For each patient, the feedforward neural network model is used to predict the health score under the original data and the improved data, and the difference between the health score under the improved data and the original data predicted by the model is calculated as the health score difference ;
[0038] According to whether the actual improvement values of activity and rest in the patient's physical data meet the standards, the patient is judged to be overdosed:
[0039] If the actual improvement value of the activity is greater than the calculated improvement value of the activity, and , then it is judged as excessive exercise;
[0040] If the actual activity improvement is less than the calculated activity improvement, and , then the exercise is judged to be unsatisfactory;
[0041] If the actual improvement in rest volume is greater than the calculated improvement in rest volume, and , then it is judged that the rest is excessive;
[0042] If the actual improvement in rest volume is less than the calculated improvement in rest volume, and , it is judged that the rest does not meet the standard.
[0043] As a preferred solution of the patient health monitoring method based on hemodialysis data analysis of the present invention, wherein: the improvement suggestions adjusted according to the patient's physical condition include:
[0044] Based on the judgment of whether the actual improvement value of activity and the actual improvement value of rest in the patient's physical data meet the standard, the adjustment factor is set, which is expressed as: ;in, represents the adjustment factor, represents the actual health score of the ith patient, represents the score before improvement of the i-th patient, represents the mean health score of patients;
[0045] The activity adjustment value and rest adjustment value are calculated based on the adjustment factor applied to the activity improvement value and the rest improvement value, respectively, expressed as: ; ;in, represents the rest volume adjustment value of the i-th patient, represents the activity regulation value of the i-th patient;
[0046] Adjust rest and activity adjustments based on patient dynamics.
[0047] As a preferred solution of the patient health monitoring method based on hemodialysis data analysis of the present invention, wherein: the encrypted data transmission based on different clustered patients, the report generation and report sending include:
[0048] Adopt data encryption technology, use SSL or TLS protocol to encrypt the physical data, hemodialysis data and health score data of different cluster patients, use OAuth2 protocol for identity authentication and authorization management, and store data in the cloud and local servers;
[0049] The report is generated using Matplotlib and Seaborn of Python libraries and filled with data based on the patient’s physical data, hemodialysis data, and health score. The report is sent as an attachment to the specified email address using the SMTP library.
[0050] Another object of the present invention is to provide a system for patient health monitoring method based on hemodialysis data analysis, which comprises:
[0051] A data collection module collects patient physical data and hemodialysis data;
[0052] Data clustering module, which analyzes the correlation values between data, performs clustering calculations through clustering algorithms, and analyzes patient feature set data of different clusters;
[0053] Improvement suggestion module, based on patients in different clusters, analyzes physical data and calculates improvement values, predicts health scores, makes excessive judgments, and adjusts improvement suggestions;
[0054] The data transmission module encrypts and transmits the data of patients in different clusters, generates reports and sends them.
[0055] The beneficial effects of the present invention are as follows: by comprehensively collecting the patient's physical data and hemodialysis data to ensure the comprehensiveness and accuracy of the data, patients with similar characteristics are grouped and clustered according to the patient's related data sets, so that the patients in each cluster have similar health conditions and living habits, which is helpful for targeted treatment and analysis of different types of patients, and the patient's physical data and hemodialysis data are used to predict health scores, and improvement suggestions for exercise and rest are given according to the patient grouping and clustering, and the improvement suggestions are dynamically adjusted based on changes in the health score, making full use of various data in the dialysis process and giving patients health suggestions. BRIEF DESCRIPTION OF THE DRAWINGS
[0056] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.
[0057] Figure 1 Schematic diagram of the process of patient health monitoring method based on hemodialysis data analysis.
[0058] Figure 2 Schematic diagram of the structure of the patient health monitoring system based on hemodialysis data analysis. DETAILED DESCRIPTION
[0059] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.
[0060] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.
[0061] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.
[0062] Example 1, reference Figure 1 , which is the first embodiment of the present invention, and provides a patient health monitoring method based on hemodialysis data analysis. The patient health monitoring method based on hemodialysis data analysis includes:
[0063] S1: Collect patient physical data and hemodialysis data, analyze the correlation values between the data and construct a correlation data set;
[0064] Preferably, collecting patient physical data and hemodialysis data, analyzing the correlation values between the data to construct a correlation data set includes:
[0065] The patient's physical data is collected through health monitoring equipment, including exercise data and rest data. The exercise data includes the total number of steps, exercise time, and calorie consumption value, and the rest data includes sleep duration and sleep quality index;
[0066] Collect patients’ hemodialysis data through medical devices, including weight, blood pressure, heart rate before and after dialysis, as well as dialysis duration and dialysis efficiency;
[0067] Synchronize the collected physical data and hemodialysis data with the same time tag;
[0068] Standardize physical and hemodialysis data;
[0069] The Pearson correlation coefficient was used to calculate the correlation between each physical data and hemodialysis data, expressed as: ;in, represents the correlation value of the xth item of physical data and the yth item of hemodialysis data, represents the ith collected value of the xth item of physical data, represents the i-th collected value of the y-th hemodialysis data, n represents the total number of collected values; i represents the collected value index, represents the mean of the x-th item of body data, represents the mean of the y-th item of hemodialysis data;
[0070] The analysis was performed by testing based on the calculated correlation values, expressed as: ;in, represents the test value of the xth item of physical data and the yth item of hemodialysis data;
[0071] Because the calculation of the correlation value involves the loss of two degrees of freedom, which is used to calculate the average of the x-th physical data and the y-th hemodialysis data, The test value has n-2 degrees of freedom;
[0072] The significance level is determined based on historical experience (usually 0.05), and the correlation value greater than the test value Perform statistics to form a correlation set R.
[0073] Comprehensively collect patients' physical data (such as exercise data and rest data) and hemodialysis data through health monitoring equipment and medical devices to ensure the comprehensiveness and accuracy of the data. Synchronize the collected physical data and hemodialysis data with the same time label and perform standardized processing to ensure the comparability and consistency of data from different sources. Calculate the correlation value between each physical data and hemodialysis data through the Pearson correlation coefficient, deeply explore the intrinsic connection between the data, and provide a scientific basis for subsequent health analysis and intervention. Use t-test to analyze the calculated correlation values to ensure the statistical significance and reliability of the correlation results, eliminate the interference of random factors, screen out physical data and hemodialysis data with significant correlation based on the significance level, and construct the correlation set R to provide accurate data support for personalized health management and improvement suggestions.
[0074] S2: Perform clustering calculations based on the patient-related data set using a clustering algorithm, and analyze the patient feature set data of different clusters;
[0075] Preferably, clustering calculation is performed based on the patient correlation data set by using a clustering algorithm to analyze the patient feature set data of different clusters, including:
[0076] Based on the correlation set R of different patients, the physical data and dialysis parameters corresponding to the statistically significant correlation coefficients are selected to construct the feature set J. Each record of each patient forms a row, each row includes all its significant correlation values, and each column represents a specific variable pair (such as the correlation coefficient between the number of steps and the change in weight);
[0077] Use the K-means clustering algorithm to randomly select patient records from the patient record set as the initial cluster centers , define clustering represents the i-th cluster, cluster Including all the cluster centers assigned to the initial cluster Patient characteristics set J;
[0078] Based on each patient’s feature set J, calculate the initial cluster center The Euclidean distance of is used to perform initial clustering, which is expressed as: ;in, represents the jth feature of patient feature set J, Represents the cluster center The jth feature of the patient feature set J, m represents the total number of features of the patient feature set J, Represents the patient feature set J and cluster center The Euclidean distance of
[0079] Update cluster centers based on initial clustering Belong to the cluster The average position of all patient feature sets J in is expressed as: ;in, Representation clustering The number of patient records in represents the updated cluster center;
[0080] Calculate the total squared error, expressed as: ; where SSE represents the total square error, Represents the patient feature set J to its cluster center The square of the distance, k represents the number of clusters;
[0081] Repeated calculation and update of cluster centers Calculate SSE, and stop the iteration when the update transformation of the cluster center is not obvious;
[0082] Apply the elbow rule to determine the optimal number of clusters, repeat the K-means clustering algorithm based on different k values, count the SSE of each k value, and draw a curve. According to the elbow rule of the curve, the k value with the SSE value is determined as the optimal k value;
[0083] Analyze the patient feature set data of different clusters based on the calculation of k value;
[0084] For example, if the optimal k value is 3 clusters, the patient feature set data of different clusters are summarized as follows:
[0085] Cluster 1:
[0086] center : The correlation between total steps and weight change is 0.6, and the correlation between total steps and blood pressure change is -0.4;
[0087] Characteristic description: high activity level, high sleep quality, good dialysis effect, stable changes in weight and blood pressure;
[0088] Cluster 2:
[0089] center : The correlation between total steps and weight change is 0.4, and the correlation between total steps and blood pressure change is -0.2;
[0090] Characteristic description: moderate activity level, moderate sleep quality, and average dialysis effect;
[0091] Cluster 3:
[0092] center : The correlation between total steps and weight change is 0.1, and the correlation between total steps and blood pressure change is 0;
[0093] Characteristic description: low activity, poor sleep quality, poor dialysis effect, unstable weight and blood pressure changes.
[0094] Through the K-means clustering algorithm, patients with similar characteristics are grouped according to the patients' related data sets. This method can accurately divide patients into different clusters, so that patients in each cluster have similar health conditions and living habits, which is convenient for subsequent personalized health management and intervention. By calculating the Euclidean distance between each patient feature set and the initial cluster center, and dynamically updating the cluster center to the average position of all patient feature sets, the cluster center can accurately reflect the feature distribution of patients in the cluster, thereby improving the accuracy of clustering. By calculating and minimizing the total square error (SSE), the distance between the patient feature set and the cluster center is ensured to be as small as possible, ensuring the compactness and clustering effect of the clustering results. By applying the elbow rule, the optimal number of clusters is determined by drawing SSE curves for different k values, so that the selection of the number of clusters has a scientific basis, thereby improving the rationality of the clustering results.
[0095] S3: Based on the patient feature set data of different clusters, analyze the patient's physical data and calculate the improvement value, predict the patient's health score, analyze the improvement value to judge the patient's physical condition, and adjust the improvement suggestions according to the patient's physical condition;
[0096] Preferably, based on the patient feature set data of different clusters, analyzing the patient's physical data and calculating the improvement value comprises:
[0097] Physical data were collected from patients based on different clusters, including calorie consumption data and rest duration data;
[0098] Calculate the mean of calorie consumption data for patients in the same cluster , and calculate the standard deviation of the calorie consumption data of patients in the same cluster ;
[0099] The activity improvement value was calculated for each patient and expressed as: ;in, represents the activity improvement value of the i-th patient, represents the calorie consumption data of the i-th patient;
[0100] Calculate the mean of rest time data of patients in the same cluster , and calculate the standard deviation of the rest time data of patients in the same cluster ;
[0101] The improvement in rest volume was calculated for each patient and expressed as: ;in, represents the improvement value of rest volume of the i-th patient, Represents the rest time data of the i-th patient.
[0102] By calculating the mean and standard deviation of the calorie consumption data of patients in the same cluster, personalized activity improvement suggestions are provided to each patient. Personalized improvement suggestions based on cluster analysis can more effectively improve the health management effect of patients. Cluster analysis enables the system to identify patient groups with similar characteristics, thereby providing more accurate health suggestions and preventing the suggested improvement values from varying too much. It is applicable to patients in the same cluster. By using the mean and standard deviation to calculate the activity and rest improvement values, the scientific rationality of health intervention measures can be ensured, which not only avoids excessive intervention, but also ensures that the improvement suggestions are operational. This method allows dynamic adjustment of improvement strategies based on the patient's actual calorie consumption and rest duration data, ensuring that health suggestions can be updated in a timely manner as the patient's condition changes.
[0103] Furthermore, the patient's health score is predicted and the improvement value is analyzed to determine the patient's physical condition, including:
[0104] According to the patients in different clusters, a feedforward neural network model based on a fully connected layer was constructed, including an input layer, a hidden layer, and an output layer;
[0105] The input layer inputs the patient's physical data and hemodialysis data. The hidden layer includes three fully connected layers. Each neuron in each fully connected layer is connected to all neurons in the previous layer. After each fully connected layer, the ReLU activation function is applied, and the output layer outputs the predicted value.
[0106] The loss is calculated using the mean square error loss function, expressed as: ;in, MSE represents the calculation loss, M represents the total number of samples, represents the true health score of the i-th sample, represents the model prediction value;
[0107] The model parameters are iteratively optimized using the Adam optimizer and the gradient descent method. When the loss of the feedforward neural network model no longer decreases significantly during the continuous iteration, the iteration output model parameters are stopped and the feedforward neural network model is updated.
[0108] Calculate health score based on input patient physical data and hemodialysis data;
[0109] For each patient, the feedforward neural network model is used to predict the health score under the original data and the improved data, and the difference between the health score under the improved data and the original data predicted by the model is calculated as the health score difference ;
[0110] According to whether the actual improvement values of activity and rest in the patient's physical data meet the standards, the patient is judged to be overdosed:
[0111] If the actual improvement value of the activity is greater than the calculated improvement value of the activity, and , then it is judged as excessive exercise;
[0112] If the actual activity improvement is less than the calculated activity improvement, and , then the exercise is judged to be unsatisfactory;
[0113] If the actual improvement in rest volume is greater than the calculated improvement in rest volume, and , then it is judged that the rest is excessive;
[0114] If the actual improvement in rest volume is less than the calculated improvement in rest volume, and , it is judged that the rest does not meet the standard.
[0115] By constructing a feedforward neural network model based on a fully connected layer and using the patient's physical data and hemodialysis data to predict health scores, an accurate health score can be provided. This method can comprehensively consider multiple data sources, making the health score prediction more scientific and accurate. The feedforward neural network model is used to predict the original data and improved data of each patient, and the difference between the health score under the improved data and the original data predicted by the model is calculated as the health score difference. Personalized health management suggestions are provided to each patient. The improvement strategy is adjusted according to the patient's specific situation to ensure that each patient can obtain the most appropriate health intervention measures. By calculating whether the actual improvement value of the activity and the actual improvement value of the rest amount meet the standards, the patient is judged to be excessive to ensure the scientificity and rationality of the improvement suggestion. Through this method, the improvement strategy can be dynamically adjusted to avoid excessive or insufficient health intervention.
[0116] Further, the improvement suggestions are adjusted according to the patient's physical condition, including:
[0117] Based on the judgment of whether the actual improvement value of activity and the actual improvement value of rest in the patient's physical data meet the standard, the adjustment factor is set, which is expressed as: ;in, represents the adjustment factor, represents the actual health score of the ith patient, represents the score before improvement of the i-th patient, represents the mean health score of patients;
[0118] The activity adjustment value and rest adjustment value are calculated based on the adjustment factor applied to the activity improvement value and the rest improvement value, respectively, expressed as: ; ;in, represents the rest volume adjustment value of the i-th patient, represents the activity regulation value of the i-th patient.
[0119] By dynamically adjusting the rest and activity adjustment values, personalized health management suggestions can be provided to each patient. According to the patient's actual health score and the score before improvement, the improvement strategy can be accurately adjusted to ensure that each patient can obtain the most suitable health intervention measures, avoiding excessive or insufficient intervention due to fixed adjustment values, ensuring the accuracy and effectiveness of health improvement suggestions, and ensuring that the improvement suggestions are carried out within a reasonable range, thereby reducing health risks and ensuring patient safety.
[0120] S4: Encrypt data transmission based on different clustered patients, generate reports and send them;
[0121] Preferably, data encryption transmission is performed based on different clustered patients, and reports are generated and sent, including:
[0122] Adopt data encryption technology, use SSL or TLS protocol to encrypt the physical data, hemodialysis data and health score data of different cluster patients, use OAuth2 protocol for identity authentication and authorization management, and store data in the cloud and local servers;
[0123] The report is generated using Matplotlib and Seaborn of Python libraries and filled with data based on the patient’s physical data, hemodialysis data, and health score. The report is sent as an attachment to the specified email address using the SMTP library.
[0124] By using SSL or TLS protocol to encrypt the physical data, hemodialysis data and health score data of different clustered patients, the confidentiality and integrity of the data during transmission are ensured, effectively preventing data leakage and tampering. The OAuth2 protocol is used for identity authentication and authorization management to ensure that only authorized users can access and operate the data, further enhancing the security and reliability of the system. The Python library Matplotlib and Seaborn are used to generate charts, which can efficiently process and visualize the patient's physical data, hemodialysis data and health score data. Through the generated charts, the medical team can more intuitively understand the patient's health status and improvement effect.
[0125] Example 2, reference Figure 2 , which is the second embodiment of the present invention, is different from the previous embodiment and provides a system for a patient health monitoring method based on hemodialysis data analysis, including:
[0126] A data collection module collects patient physical data and hemodialysis data;
[0127] Data clustering module, which analyzes the correlation values between data, performs clustering calculations through clustering algorithms, and analyzes patient feature set data of different clusters;
[0128] Improvement suggestion module, based on patients in different clusters, analyzes physical data and calculates improvement values, predicts health scores, makes excessive judgments, and adjusts improvement suggestions;
[0129] The data transmission module encrypts and transmits the data of patients in different clusters, generates reports and sends them.
[0130] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.
[0131] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.
[0132] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.
[0133] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit with a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit with a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
[0134] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.
Claims
1. A patient health monitoring method based on hemodialysis data analysis, characterized in that: include: Collect patient physical data and hemodialysis data, analyze the correlation values between the data and construct a correlation data set; Perform clustering calculations based on patient-related data sets using clustering algorithms, and analyze patient feature set data of different clusters; Based on the patient feature set data of different clusters, analyze the patient's physical data and calculate the improvement value, predict the patient's health score, analyze the improvement value to judge the patient's physical condition, and adjust the improvement suggestions according to the patient's physical condition; Encrypted data transmission based on different clustered patients, generating reports and sending them; The method of analyzing the patient's physical data and calculating the improvement value based on the patient feature set data of different clusters includes: Physical data were collected from patients based on different clusters, including calorie consumption data and rest duration data; Calculate the mean of calorie consumption data for patients in the same cluster , and calculate the standard deviation of the calorie consumption data of patients in the same cluster ; The activity improvement value was calculated for each patient and expressed as: ;in, represents the activity improvement value of the i-th patient, represents the calorie consumption data of the i-th patient; Calculate the mean of rest time data of patients in the same cluster , and calculate the standard deviation of the rest time data of patients in the same cluster ; The improvement in rest volume was calculated for each patient and expressed as: ;in, represents the improvement value of rest volume of the i-th patient, represents the rest time data of the i-th patient; The predicting of the patient's health score and analyzing the improvement value to determine the patient's physical condition include: According to the patients in different clusters, a feedforward neural network model based on a fully connected layer was constructed, including an input layer, a hidden layer, and an output layer; The loss is calculated using the mean square error loss function, and the model parameters are iteratively optimized using the Adam optimizer and the gradient descent method. When the loss of the feedforward neural network model no longer decreases during the continuous iteration process, the iteration is stopped to output the model parameters and update the feedforward neural network model. Calculate health score based on input patient physical data and hemodialysis data; For each patient, the feedforward neural network model is used to predict the health score under the original data and the improved data, and the difference between the health score under the improved data and the original data predicted by the model is calculated as the health score difference ; According to whether the actual improvement values of activity and rest in the patient's physical data meet the standards, the patient is judged to be overdosed: If the actual improvement value of the activity is greater than the calculated improvement value of the activity, and If the actual improvement value of the activity is greater than the calculated improvement value of the activity, and , then it is judged as excessive exercise; If the actual improvement value of the activity is less than the calculated improvement value of the activity, and , then the exercise is judged to be unsatisfactory; If the actual improvement in rest volume is greater than the calculated improvement in rest volume, and , then it is judged that the rest is excessive; If the actual improvement in rest volume is less than the calculated improvement in rest volume, and , then it is determined that the rest does not meet the standard; The improvement suggestions are adjusted according to the patient's physical condition, including: Based on the judgment of whether the actual improvement value of activity and the actual improvement value of rest in the patient's physical data meet the standard, the adjustment factor is set, which is expressed as: ;in, represents the adjustment factor, represents the actual health score of the ith patient, represents the score before improvement of the i-th patient, represents the mean health score of patients; The activity adjustment value and rest adjustment value are calculated based on the adjustment factor applied to the activity improvement value and the rest improvement value, respectively, expressed as: ; ;in, represents the rest volume adjustment value of the i-th patient, represents the activity regulation value of the i-th patient.
2. The patient health monitoring method based on hemodialysis data analysis according to claim 1, characterized in that: The collecting of the patient's physical data and hemodialysis data, and analyzing the correlation values between the data to construct a correlation data set include: The patient's physical data is collected through health monitoring equipment, including exercise data and rest data. The exercise data includes the total number of steps, exercise time, and calorie consumption value, and the rest data includes sleep duration and sleep quality index; Collect patients’ hemodialysis data through medical devices, including weight, blood pressure, heart rate before and after dialysis, as well as dialysis duration and dialysis efficiency; Synchronize the collected physical data and hemodialysis data with the same time tag; Standardize physical and hemodialysis data; The Pearson correlation coefficient was used to calculate the correlation between each physical data and hemodialysis data, expressed as: ;in, represents the correlation value of the xth item of physical data and the yth item of hemodialysis data, represents the ith collected value of the xth item of physical data, represents the i-th collected value of the y-th hemodialysis data, n represents the total number of collected values; i represents the collected value index, represents the mean of the x-th item of body data, represents the mean of the y-th item of hemodialysis data; The analysis was performed by testing based on the calculated correlation values, expressed as: ;in, represents the test value of the xth item of physical data and the yth item of hemodialysis data; The correlation value will be greater than the test value Perform statistics to form a correlation set R.
3. The patient health monitoring method based on hemodialysis data analysis according to claim 2, characterized in that: The clustering calculation is performed according to the patient correlation data set by using a clustering algorithm to analyze the patient feature set data of different clusters, including: Based on the correlation set R of different patients, the physical data and dialysis parameters corresponding to the statistically significant correlation coefficients are selected to construct the feature set J; Use the K-means clustering algorithm to randomly select patient records from the patient record set as the initial cluster centers Defining clusters represents the i-th cluster, cluster Including all the cluster centers assigned to the initial cluster Patient characteristics set J; Based on each patient’s feature set J, calculate the initial cluster center The Euclidean distance , and perform initial clustering, expressed as: ;in, represents the jth feature of patient feature set J, Represents the cluster center The jth feature of the patient feature set J, m represents the total number of features of the patient feature set J, Represents the patient feature set J and cluster center The Euclidean distance of Based on the initial clustering, update the cluster center to the cluster to which it belongs The average position of all patient feature sets J in is expressed as: ;in, Representation clustering The number of patient records in Represents the updated cluster center; calculate the total square error, expressed as: ; where SSE represents the total square error, Represents the patient feature set J to its cluster center The square of the distance, k represents the number of clusters; Repeated calculation and update of cluster centers Calculate SSE, and stop the iteration when the update transformation of the cluster center is not obvious; Apply the elbow rule to determine the optimal number of clusters, repeat the K-means clustering algorithm based on different k values, count the SSE of each k value, and draw a curve. According to the elbow rule of the curve, the k value with the SSE value is determined as the optimal k value; The patient feature set data of different clusters were analyzed based on the calculation of k value.
4. The patient health monitoring method based on hemodialysis data analysis according to claim 3, characterized in that: The encrypted data transmission based on different clustered patients and the report generation and report sending include: Adopt data encryption technology, use SSL or TLS protocol to encrypt the physical data, hemodialysis data and health score data of different cluster patients, use OAuth2 protocol for identity authentication and authorization management, and store data in the cloud and local servers; The report is generated using Matplotlib and Seaborn of Python libraries and filled with data based on the patient’s physical data, hemodialysis data, and health score. The report is sent as an attachment to the specified email address using the SMTP library.
5. A system for patient health monitoring based on the hemodialysis data analysis method according to any one of claims 1 to 4, characterized in that: include: A data collection module collects patient physical data and hemodialysis data; Data clustering module, which analyzes the correlation values between data, performs clustering calculations through clustering algorithms, and analyzes patient feature set data of different clusters; Improvement suggestion module, based on patients in different clusters, analyzes physical data and calculates improvement values, predicts health scores, makes excessive judgments, and adjusts improvement suggestions; The data transmission module encrypts and transmits the data of patients in different clusters, generates reports and sends them.
