Health management data processing method and system based on dynamic feedback of patients

Abstract

The present invention relates to a health management data processing method and system based on patient dynamic feedback, belonging to the field of health management. By downloading from the chain multiple patient physical sign change curves whose recorded disease states are consistent with the patient's disease states after blockchain certification and whose recorded health management plans are consistent with the health management plans, traversing the multiple curves to statistically calculate the ratio of the number of neighboring curves to the mean and setting it as multiple confidence densities, and extracting the physical sign change curve with the maximum confidence density; during health management intervention, statistically calculating the deviation vector between the change monitoring curve of the patient feedback information and the curve in the previous section of the change to correct the curve in the subsequent section of the change to obtain the corrected curve in the subsequent section of the change; when the deviation between the corrected curve in the subsequent section of the change and the physical sign change curve is greater than or equal to the expected deviation, feeding back the inadaptability identification of the health management plan execution to the health management terminal, which solves the technical problem of being difficult to accurately process health management data to generate a health management plan that conforms to individual characteristics.

Description

Technical Field

[0001] The present invention relates to the field of health management, and particularly to a health management data processing method and system based on dynamic feedback of patients. Background Art

[0002] In the traditional field of health data management, the analysis of patients' health data often relies on the statistical results of group samples. Although this method can reveal the general trend of a certain disease or health condition in the group, due to the neglect of individual differences between people, it is difficult to accurately characterize an individual's health status or abnormal conditions in practical applications. Specifically, traditional health data analysis usually collects the health data of a large number of patients, and then conducts statistical analysis based on these data to obtain the general laws of a certain disease or health condition in the group. However, a significant drawback of this method is that it cannot fully consider the uniqueness of each patient. Each person's physiological condition, genetic factors, lifestyle habits, etc. may all affect their health status, and these factors are often ignored in traditional group sample analysis.

[0003] Therefore, there are often technical problems in the prior art that it is difficult to accurately process health management data to generate a health management plan that conforms to individual characteristics. Summary of the Invention

[0004] In view of the technical problem in the prior art that it is difficult to accurately process health management data to generate a health management plan that conforms to individual characteristics, the present invention provides a health management data processing method and system based on dynamic feedback of patients to solve this problem.

[0005] The technical solutions of the present invention to solve the above technical problems are as follows:

[0006] In the first aspect, the present invention provides a health management data processing method based on dynamic feedback of patients, including:

[0007] Downloading from the chain multiple patient body sign change curves with the recorded disease state stored on the blockchain being consistent with the patient's disease state and the recorded health management plan being consistent with the health management plan; traversing the multiple patient body sign change curves to statistically calculate the ratio of the number of neighborhood curves to the mean number of neighborhood curves, which is set as multiple confidence densities, and extracting the body sign change curve with the maximum value of the multiple confidence densities; when performing health management intervention, statistically calculating the deviation vector between the body sign change monitoring curve of the patient feedback information and the front section curve of the body sign change, and correcting the rear section curve of the body sign change to obtain a corrected rear section curve of the body sign change, where the front section curve of the body sign change is a partial curve obtained by time-sequence alignment of the body sign change curve and the body sign change monitoring curve; when the deviation between the corrected rear section curve of the body sign change and the body sign change curve is greater than or equal to the expected deviation, feedback an inadaptability mark to the health management terminal for the health management plan.

[0008] In a second aspect, the present invention provides a health management data processing system based on patient dynamic feedback, including:

[0009] A physical sign monitoring module, configured to download from the chain multiple patient physical sign change curves whose recorded disease states are consistent with the actual disease states of the patients after being stored on the blockchain and whose recorded health management plans are consistent with the health management plans; a variable extraction module, configured to traverse the multiple patient physical sign change curves to statistically calculate the ratio of the number of neighborhood curves to the average value of the number of neighborhood curves, set it as multiple confidence densities, and extract the physical sign change curve with the maximum value of the multiple confidence densities; a deviation correction module, configured to, when performing health management intervention, statistically calculate the deviation vector between the physical sign change monitoring curve of the patient feedback information and the front-section curve of the physical sign change, and correct the rear-section curve of the physical sign change to obtain a corrected rear-section curve of the physical sign change, where the front-section curve of the physical sign change is a partial curve obtained by time-sequentially aligning the physical sign change curve and the physical sign change monitoring curve; a feedback management module, configured to, when the deviation between the corrected rear-section curve of the physical sign change and the physical sign change curve is greater than or equal to the expected deviation, feed back an inadaptability identification of the health management plan to the health management terminal.

[0010] The beneficial effects of the present invention are: Through the patient data stored on the blockchain and the dynamic feedback mechanism, high-confidence physical sign change curves are accurately screened, and the physical sign change curves are corrected in real time according to the patient feedback, effectively identifying and feeding back the inadaptability of the health management plan, thereby improving the personalization and accuracy of health management and enhancing individual adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is a schematic flow chart of a health management data processing method based on patient dynamic feedback provided by the present invention.

[0012] Figure 2 It is a schematic structural diagram of a health management data processing system based on patient dynamic feedback provided by the present invention.

[0013] Description of the reference numerals: Physical sign monitoring module 11, variable extraction module 12, deviation correction module 13, feedback management module 14. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0014] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present invention.

[0015] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality of" means two or more unless otherwise specifically defined.

[0016] In the description of the present invention, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present invention is not necessarily construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to implement and use the present invention. In the following description, details are set forth for purposes of explanation. It should be understood that those of ordinary skill in the art can recognize that the present invention can be implemented without the use of these specific details. In other instances, well-known structures and processes are not elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

[0017] Embodiment 1:

[0018] As Figure 1 shown, the embodiment of the present invention provides a method for processing health management data based on dynamic feedback of patients, including:

[0019] S10: Download from the chain multiple patient physical sign change curves whose recorded disease states are consistent with the blockchain-certified disease states of the patients and whose recorded health management plans are consistent with the health management plans.

[0020] S20: Traverse the multiple patient physical sign change curves to statistically calculate the ratio of the number of neighborhood curves to the mean value of the number of neighborhood curves, set it as multiple confidence densities, and extract the physical sign change curve with the maximum value of the multiple confidence densities.

[0021] S30: When performing health management intervention, statistically calculate the deviation vector between the physical sign change monitoring curve of the patient feedback information and the front-section curve of the physical sign change, and correct the rear-section curve of the physical sign change to obtain the corrected rear-section curve of the physical sign change. The front-section curve of the physical sign change is a partial curve obtained by time-sequence alignment of the physical sign change curve and the physical sign change monitoring curve.

[0022] S40: When the deviation between the corrected rear-section curve of the physical sign change and the physical sign change curve is greater than or equal to the expected deviation, feedback an inadaptability identifier for the health management plan to the health management terminal.

[0023] Exemplarily, blockchain is a distributed database technology that collectively maintains a reliable database in a decentralized and trustless manner. In this database, information is packaged into individual blocks, and cryptographic techniques are used to ensure the authenticity and integrity of each block. These blocks are linked together in chronological order to form an immutable data chain. Blockchain technology has two core features: difficult data tampering and decentralization, which make the information recorded by the blockchain more authentic and reliable and can solve the trust problem. In this solution, blockchain is used as a platform to store the patient's disease record status, health management plan, and multiple patient vital sign change curves to ensure that the subsequent data is true, complete, and untampered, providing strong support for the formulation and implementation of subsequent health management plans.

[0024] Specifically, each distributed platform first uses the immutability and distributed storage characteristics of blockchain technology to deposit data such as the patient's disease record status, health management plan, and vital sign change curves on the blockchain. The disease record status refers to a set of detailed information recording the patient's current or historical disease conditions, including but not limited to disease diagnosis, disease severity, medical history records, symptom manifestations, examination results, treatment responses, etc. The health management plan refers to a series of strategies and measures formulated for a specific patient to improve or maintain their health status. This step ensures the originality and authenticity of the data. When these data need to be obtained for health management decisions, the system will interact with the blockchain network through specific interfaces or protocols. This interaction process follows the data access rules of the blockchain to ensure that only authorized users or systems can access sensitive health data. In the data acquisition stage, the system sends a request to the blockchain network, which contains specific identifiers or indexes of the data to be downloaded. After receiving the request, the blockchain network searches for matching data on the chain based on these identifiers or indexes. Once the matching data is found, the blockchain network packages these data into a specific format and securely transmits them to the requester through an encrypted channel. During the transmission process, the data is encrypted to ensure its confidentiality until it is safely received and decrypted. After receiving the data, the system performs a series of verification steps to ensure that the downloaded data is consistent with the patient's actual disease status and health management plan. This step ensures the accuracy of subsequent health management decisions. If the verification passes, the system stores these data in the local database for subsequent analysis and processing. These data include multiple patient vital sign change curves, which reflect the changes in the patient's health status at different time points. The above steps provide reliable data support for health management decisions, helping to improve the efficiency and accuracy of health management.

[0025] Furthermore, analyze multiple patient vital sign change curves. Identify the vital sign change curves with specific characteristics by calculating the ratio of the number of neighborhood curves to the mean number of neighborhood curves (i.e., confidence density). First, the data of multiple patient vital sign change curves collected may represent the changes of different physiological parameters (such as heart rate, blood pressure, body temperature, etc.) over time. Each curve can be regarded as a time series data, containing a series of time points and corresponding vital sign values. For each vital sign change curve, define a neighborhood range. This neighborhood can be temporal (e.g., select a time period before and after a certain time point), or spatial (e.g., select a set of curves similar to the current curve in a multi-dimensional space). The specific definition of the neighborhood depends on the purpose of the analysis and the characteristics of the data. For each curve, calculate the number of curves within its neighborhood, which can be achieved by comparing the similarity between curves (such as using a distance metric). Curves with high similarity are considered to be in the same neighborhood. Then, statistically calculate the mean number of neighborhood curves for all curves. This mean reflects how many similar neighborhood curves each curve has on average in the entire dataset. For each curve, compare the number of its neighborhood curves with the mean calculated above to obtain a ratio, i.e., the confidence density. The confidence density reflects the uniqueness or representativeness of the curve in the dataset. If a curve has a high confidence density, it means that it is relatively isolated or has unique characteristics in the dataset; if the confidence density is low, it means that it is similar to many other curves. Among all the curves, find the curve with the highest confidence density, which can be achieved by comparing the confidence density values of all curves; or multiple curves with relatively high confidence densities can be extracted as needed for further analysis or verification. Finally, based on the maximum value (or relatively high values) of the confidence density, identify the vital sign change curves with specific characteristics. These curves may represent certain special physiological states, disease progressions, or treatment responses, etc. In summary, by statistically analyzing the ratio of the number of neighborhood curves to the mean for multiple patient vital sign change curves, curves with unique or representative characteristics are extracted, providing valuable information for further adaptive clinical applications.

[0026] Next, when implementing health management intervention measures for patients, collect the feedback information of patients after the intervention. Among them, the management intervention measures may include drug treatment, lifestyle adjustment, nutritional guidance, etc., aiming to improve the health status of patients. Continuously monitor the physical signs changes of patients during the implementation of the plan, and accordingly draw a physical signs change curve, which reflects the dynamic changes of patients' physical signs after receiving the health management plan. The collected feedback information may include patients' subjective feelings (such as symptom improvement, comfort, etc.) and objective physical signs data (such as blood pressure, heart rate, weight, etc.). Subsequently, based on the collected feedback information, especially the objective physical signs data, generate a physical signs change monitoring curve. This curve reflects the changes of patients' physical signs over time after the intervention. In the generated physical signs change monitoring curve, identify the part that is temporally aligned with the physical signs change curve as the front-section curve of physical signs change. This part of the curve represents the physical signs state of patients before or at the initial stage of the intervention. And, compare the physical signs change monitoring curve with the front-section curve of physical signs change, and calculate the deviation vector between them. The deviation vector quantifies the difference between the two, reflecting the degree and direction of the change in patients' physical signs after the intervention. After that, use the calculated deviation vector to correct the latter part of the physical signs change monitoring curve (i.e., the longer time period after the intervention). The purpose of the correction is to adjust the curve according to the previous change trend and patients' feedback information to more accurately reflect the actual change of patients' physical signs. After correction, obtain the corrected curve of the latter part of physical signs change. This curve more accurately reflects the change trend of patients' physical signs after the implementation of the health management intervention, providing a basis for subsequent evaluation and adjustment of intervention measures. To sum up, this process aims to obtain more accurate physical signs change information by collecting patients' feedback information, generating a physical signs change monitoring curve, calculating the deviation vector, and correcting the latter part of the curve, providing strong support for the evaluation and adjustment of health management interventions. This process helps to improve the pertinence and effectiveness of health management and promote the continuous improvement of patients' health status.

[0027] Finally, according to the preset expected deviation threshold, determine whether the deviation between the corrected curve in the latter segment of the physical sign change and the original curve is greater than or equal to this threshold. This threshold is obtained based on clinical experience and statistical analysis and is used to evaluate the suitability of the health management plan. If the deviation between the corrected curve in the latter segment of the physical sign change and the original curve is greater than or equal to the expected deviation threshold, it is considered that the current health management plan may not be applicable to this patient. In this case, the system will generate a non-suitability identifier and feedback it to the health management terminal. This identifier contains information such as the degree of deviation, possible reasons, and recommended adjustment directions. After receiving the non-suitability identifier, the health management terminal will adjust the health management plan according to the specific situation of the patient, the reasons for the deviation, and the adjustment directions recommended by the system. The adjusted plan will be applied to the patient again, and the physical sign changes will continue to be monitored to evaluate the suitability and effectiveness of the new plan. To sum up, by continuously monitoring physical sign changes, correcting curves, calculating deviations and judging suitability, and providing feedback and adjustment suggestions to the health management terminal, it aims to provide personalized health management services for patients and ensure the pertinence and effectiveness of the health management plan.

[0028] To sum up, by using blockchain technology, the integrity and authenticity of the disease records and health management plans are ensured, and multiple physical sign change curves consistent with the actual state of the patient are downloaded from the chain. By intelligently analyzing these curves, the ratio of the number of neighborhood curves to the mean, that is, the confidence density, is calculated, and the physical sign change curve with the highest confidence is selected as the benchmark. During the health management intervention process, this method can monitor the physical sign changes of patients in real time and compare them with the previous physical sign change curves, and accurately correct the latter segment curve by calculating the deviation vector. When the deviation between the corrected curve and the original curve exceeds the expected range, the system will immediately feedback the non-suitability information of the health management plan to the health management terminal. This method not only improves the personalization and accuracy of health management, but also effectively reduces the health risks caused by the non-suitability of the plan, providing strong technical support for the health management and treatment of patients.

[0029] In a preferred embodiment, the downloaded disease records stored on the blockchain have the same status as the patient's disease status, and multiple patient physical sign change curves with the recorded health management plan being consistent with the health management plan are included: according to the patient's disease status, a status comparison binary tree is constructed, and according to the health management plan, a plan comparison binary tree is constructed. Any layer of the comparison binary tree represents a comparison index attribute and a consistency rule. Only when the upper-level comparison is consistent can the lower-level comparison be entered. When all levels of comparison are consistent, it is considered to meet the comparison binary tree. When any level of comparison is inconsistent, it is considered not to meet the comparison binary tree; download the first disease record status, the first recorded health management plan, and the first patient physical sign change curve of the patient to be analyzed that are stored on the blockchain; input the first disease record status into the status comparison binary tree, and input the first recorded health management plan into the plan comparison binary tree. When the output results are both consistent, it is considered that the disease record status is consistent with the patient's disease status, and the recorded health management plan is consistent with the health management plan, and add the first patient physical sign change curve to the multiple patient physical sign change curves.

[0030] Specifically, in the blockchain-based health management data processing flow, to ensure the consistency of the disease records status, health management plan downloaded from the chain with the actual status of the patient, the solution adopts the method of constructing a comparison binary tree for hierarchical comparison. First, a status comparison binary tree is constructed based on the patient's disease status. Each layer of this binary tree represents an index attribute to be compared and its consistency rule. Similarly, a plan comparison binary tree is constructed based on the health management plan. The structures of these two comparison binary trees are similar and are both used to verify the consistency of information hierarchically. Next, the first disease record status, the first recorded health management plan, and the first patient physical sign change curve of the patient to be analyzed are downloaded from the blockchain. Then the comparison process begins, that is, the first disease record status is first input into the status comparison binary tree and compared layer by layer from the root node downwards. At each layer of the comparison, it is checked whether the index attribute to be compared and its consistency rule are satisfied. Only when the comparison results of the previous layer are consistent can the comparison enter the next layer. If the comparison results of any layer are inconsistent, then the entire comparison process will stop immediately and output the result that does not meet the comparison binary tree. Such a design can avoid invalid comparisons and save computing power. Similarly, the first recorded health management plan is also input into the plan comparison binary tree for hierarchical comparison. Only when the output results of both the status comparison binary tree and the plan comparison binary tree are consistent is it considered that the disease record status is consistent with the patient's disease status, and the recorded health management plan is consistent with the health management plan. Finally, when both of these conditions are met, the first patient physical sign change curve is added to the set of multiple patient physical sign change curves for subsequent analysis and processing. By constructing a comparison binary tree and performing hierarchical comparison, not only can the computing power dependence of a single comparison be reduced, but also an invalid comparison process can be avoided, thereby improving the efficiency and accuracy of the entire health management data processing flow.

[0031] In a preferred embodiment, constructing a state comparison binary tree according to the patient's disease state includes: analyzing the correlation between the patient's disease state indicators and the amplitude of the fluctuation of the physical sign change curve, sorting the patient's disease state indicators according to the Pearson correlation coefficient from large to small, and obtaining the sorting result of the patient's disease state indicators, wherein for the indicators with the same Pearson correlation coefficient, random continuous sorting is performed; according to the sorting result of the patient's disease state indicators, extracting the first serial number of the patient's disease state from the patient's disease state to construct the first-level parameters to be compared; wherein when the first serial number of the patient's disease state indicator is a type indicator, the first-level consistency rule is: consistent if the types are the same, otherwise inconsistent, and when the first serial number of the patient's disease state indicator is a quantitative indicator, the first-level consistency rule is: consistent if the input data deviation from the first serial number of the patient's disease state is less than or equal to the threshold, otherwise inconsistent; constructing the first level of the binary tree according to the first-level consistency rule and the first-level parameters to be compared, until the Nth level of the binary tree is constructed, and fusing the first N levels to generate the state comparison binary tree.

[0032] Further, in the process of constructing the state comparison binary tree, attention should be paid to the correlation between the patient's disease state indicators and the amplitude of the fluctuation curve of physical signs. The core of this step lies in understanding which disease state indicators have a greater influence on the prediction or explanation of physical sign changes. This correlation is quantified by using the Pearson correlation coefficient. The Pearson correlation coefficient is a value between -1 and 1, which is used to measure the linear correlation degree between two variables. Calculate the Pearson correlation coefficient between each disease state indicator and the amplitude of the fluctuation curve of physical signs, and sort them according to the magnitude of the coefficient. It should be noted that for indicators with the same Pearson correlation coefficient, a random continuous sorting method is adopted to ensure the uniqueness of the sorting. After sorting, the disease state indicator with the first serial number is extracted from the patient's disease state as the parameter to be compared at the first level of the binary tree construction. The intention of this step is to start the comparison from the indicators most relevant to the physical sign changes, because these indicators may play a more crucial role in judging the changes in the patient's state. Next, consistency rules are set for the first level, and these rules vary according to the type of disease state indicators (type indicators or quantitative indicators). For type indicators, the rules are relatively simple. As long as the input type is the same as the type of the disease state of the patient with the first serial number, it is considered consistent; otherwise, it is considered inconsistent. For quantitative indicators, a threshold is set. Only when the deviation between the input data and the disease state of the patient with the first serial number is less than or equal to this threshold, is it considered consistent; otherwise, it is considered inconsistent. With the parameters to be compared and the consistency rules at the first level, the first level of the binary tree can be constructed. Subsequently, the subsequent levels are continued to be constructed according to the same logic until the required Nth level is reached. In this process, the parameters to be compared at each level are determined based on the comparison results and sorting of the previous level, ensuring the hierarchy and progression of the comparison. Finally, the first N levels are integrated to generate a complete state comparison binary tree. The structure and comparison rules of this binary tree are designed to ensure the objectivity and accuracy of the comparison, especially by comparing the indicators with a large correlation with the physical sign state first. When these more important indicators are abnormal, it can be directly regarded as inconsistent, thus avoiding unnecessary subsequent comparisons and improving the efficiency of the comparison. Generally speaking, this method of comparing layer by layer according to the magnitude of the correlation not only ensures the objectivity of the comparison, but also improves the efficiency and accuracy of the entire comparison process by preferentially comparing more important indicators.

[0033] In a preferred embodiment, analyzing the correlation between the patient's disease state indicators and the amplitude fluctuations of the physical sign change curve, according to the Pearson correlation coefficient, includes: obtaining the patient's disease state indicators and extracting the first patient's disease state indicators; using the amplitude fluctuation of the first patient's disease state indicators as the only independent variable and the amplitude fluctuation of the physical sign change curve as the dependent variable, collecting the amplitude fluctuation sequence of the first patient's disease state indicators and the amplitude fluctuation sequence of the physical sign change curve to perform Pearson correlation coefficient analysis, obtaining the Pearson correlation coefficient of the first patient's disease state indicators; adding the Pearson correlation coefficient of the first patient's disease state indicators into the Pearson correlation coefficient.

[0034] Optionally, when analyzing the correlation between the patient's disease state indicators and the amplitude fluctuations of the physical sign change curve, the Pearson correlation coefficient, a statistical measure, is used to quantify the linear relationship between the two. First, a series of disease state indicators are collected from the patient's medical records or health monitoring devices. These indicators may include heart rate, blood pressure, blood glucose levels, etc., which can reflect the patient's health status. Select the first (or any one) of the collected disease state indicators as the object of the current analysis, i.e., the first patient's disease state indicators. For the first patient's disease state indicators, collect the sequence of their amplitude fluctuations. At the same time, also collect the sequence of the amplitude fluctuations of the patient's physical sign change curve as the dependent variable. These two sequences will be used for Pearson correlation coefficient analysis. Then, use the Pearson correlation coefficient formula to calculate the correlation between the amplitude fluctuations of the first patient's disease state indicators and the amplitude fluctuations of the physical sign change curve. The formula for the Pearson correlation coefficient (r) is as follows: ; where x i and y i are the i-th observations of the amplitude fluctuations of the first patient's disease state indicators and the amplitude fluctuations of the physical sign change curve respectively, and are the means of these two sequences respectively, and n is the number of observations. Finally, record the calculated Pearson correlation coefficient of the first patient's disease state indicators, which will be used for subsequent analysis and comparison.

[0035] Exemplarily, assume there are observed data at 5 time points, and the amplitude fluctuations of the patient's heart rate (bpm, beats per minute) and blood pressure (mmHg, millimeters of mercury) are recorded at each time point. The amplitude fluctuation can be understood as the maximum change in the measured value within a certain time window. The following are the specific observed values:

[0036]

[0037] Based on the above data, calculate the average value. The average value of the heart rate amplitude fluctuations = = 10 bpm; the average value of the blood pressure amplitude fluctuations = = 14.8 mmHg. Calculate the sum of products of differences: = 14.4. Calculate the sum of squares of differences: = 10; = 19.72. Substitute into the Pearson correlation coefficient formula for calculation: . That is, the calculated Pearson correlation coefficient r is approximately 0.91, which indicates a strong positive correlation between the amplitude of heart rate fluctuations and the amplitude of blood pressure fluctuations. That is, when the amplitude of heart rate fluctuations increases, the amplitude of blood pressure fluctuations also tends to increase. This coefficient can be used for subsequent analysis and comparison to determine whether heart rate is one of the important factors affecting blood pressure fluctuations.

[0038] In a preferred embodiment, when performing health management intervention, the deviation vector between the physical sign change monitoring curve and the front-segment curve of physical sign change of the patient feedback information is statistically analyzed, and the rear-segment curve of physical sign change is corrected to obtain the corrected rear-segment curve of physical sign change, including: obtaining a physical sign change prediction unit, wherein the physical sign change prediction unit is generated by training a long short-term memory neural network with multiple groups of data, and any one of the multiple groups of data includes a physical sign change monitoring record curve, a front-segment record curve of physical sign change, and a label identifying the rear-segment record curve of physical sign change; the deviation vector includes the first physical sign attribute deviation vector time series information from the physical sign change monitoring curve to the Mth physical sign attribute deviation vector time series information of the front-segment curve of physical sign change; statistically analyze the proportion of the first moment greater than or equal to the deviation modulus value in the first physical sign attribute deviation vector time series information until the proportion of the Mth moment greater than or equal to the deviation modulus value in the Mth physical sign attribute deviation vector time series information is statistically analyzed; according to the number of the first moment proportion to the Mth moment proportion greater than or equal to the moment proportion, integrate at least twice the number of physical sign change prediction units for output mean fusion to obtain a physical sign change prediction model, and process the physical sign change monitoring curve and the front-segment curve of physical sign change of the patient feedback information to obtain the rear-segment curve of physical sign change.

[0039] Specifically, a long short-term memory neural network (LSTM) can be used to predict and correct the health management intervention process of the latter segment curve of the patient's physical sign changes. First, multiple sets of data are collected for training the LSTM neural network. Each set of data contains three key parts: the monitoring record curve of physical sign changes, the record curve of the former segment of physical sign changes, and a label identifying the record curve of the latter segment of physical sign changes. These curves and labels together constitute the input and output required for training the neural network. The monitoring record curve of physical sign changes reflects the physical sign changes of the patient over a period of time, while the record curve of the former segment of physical sign changes provides the baseline information before the changes occur. The label is used to indicate the expected record curve of the latter segment of physical sign changes, that is, the target that the neural network needs to learn. The multiple sets of collected data are used to train the LSTM neural network to generate one or more physical sign change prediction units, which can predict the possible record curve of the latter segment of physical sign changes based on the input monitoring curve of physical sign changes and the former segment curve. For the given patient feedback information, calculate the deviation vector between the monitoring curve of physical sign changes and the former segment curve of physical sign changes. This deviation vector contains the deviation vector time series information from the first physical sign attribute to the Mth physical sign attribute, reflecting the differences between the two at each time point. For the deviation vector time series information of each physical sign attribute, count the proportion of the moments greater than or equal to a certain deviation modulus value, which helps to identify which physical sign attributes have more significant changes and when these significant changes occur. According to the number of the proportions of the moments greater than or equal to the deviation modulus value in the deviation vector time series information of each physical sign attribute, select at least twice the number of physical sign change prediction units for integration. This integration method improves the accuracy and stability of the prediction through output mean fusion. Finally, use the integrated physical sign change prediction model to process the monitoring curve of physical sign changes and the former segment curve of the patient feedback information. This model can predict and output the corrected record curve of the latter segment of physical sign changes based on the input information. In summary, this process realizes the accurate prediction and correction of the record curve of the latter segment of the patient's physical sign changes by using the LSTM neural network and deviation vector analysis. This method not only improves the accuracy of health management but also provides a more personalized health intervention plan for patients.

[0040] In a preferred embodiment, any one of the multiple sets of data includes the monitoring record curve of physical sign changes, the record curve of the former segment of physical sign changes, and the label identifying the record curve of the latter segment of physical sign changes, including: collecting the record curve of physical sign changes and the initial monitoring record curve of physical sign changes, wherein the initial monitoring record curve of physical sign changes is aligned with the two time series ends of the record curve of physical sign changes; bisecting the initial monitoring record curve of physical sign changes to obtain the monitoring record curve of physical sign changes and the label identifying the record curve of the latter segment of physical sign changes.

[0041] Furthermore, the process of collecting multiple sets of data required by the long short-term memory neural network (LSTM) is as follows: Before starting the collection, ensure that there are appropriate devices and methods to record the changes in the patient's physical signs, which may include a heart rate monitor, a blood pressure monitor, a blood glucose meter, etc., depending on the type of physical signs to be monitored. Then, continuously monitor the patient's physical signs and record the data of physical sign changes over a period of time (such as several hours or several days). These data constitute a physical sign change record curve, which reflects the overall physical sign state of the patient during this period. At the same time, extract an initial physical sign change monitoring record curve from the same period. This curve should be aligned at both ends in time series with the physical sign change record curve, that is, their starting points and ending points should match. This is to ensure that the two curves can be directly compared and analyzed. Subsequently, perform a two-segment processing on the initial physical sign change monitoring record curve. This means dividing the curve into two parts: the first half is used as the physical sign change monitoring record curve to represent the state before the physical sign change occurs, and the second half is used as the label for the record curve after the physical sign change to represent the state after the physical sign change occurs. Combine the processed physical sign change monitoring record curve, the front-segment record curve of the physical sign change (i.e., the first half after two-segmentation), and the label for the record curve after the physical sign change to form a complete data set. This process will be repeated multiple times to collect a sufficient number of data sets for training the LSTM neural network. After collecting all the data sets, perform data verification to ensure that the time series alignment of each curve is correct and the data quality meets the training requirements. Then, organize the data for subsequent use in the training of the neural network. In summary, this process generates multiple sets of data for training the LSTM neural network by collecting and analyzing the data of the patient's physical sign changes. These data sets include the physical sign change monitoring record curve, the front-segment record curve of the physical sign change, and the label for the record curve after the physical sign change, providing the necessary input and output information for the training of the neural network.

[0042] In a preferred embodiment, when the deviation between the corrected curve of the latter segment of the physical sign change and the physical sign change curve is less than the expected deviation, perform an adaptation identification on the health management plan and conduct blockchain storage and proof with the diseased record status.

[0043] Specifically, if the deviation between the corrected curve of the physical sign change in the latter stage and the physical sign change curve is less than the expected deviation, it indicates that the prediction result is relatively accurate. The health management plan may need to be adapted according to the corrected curve. The adapted health management plan may include adjusting drug doses, changing eating habits, increasing exercise plans, etc., to better meet the patient's health needs. Identify the adapted health management plan so that it can be easily identified and tracked during subsequent health management. This identification can be a code, a label, or any other mark that can uniquely identify the change in the health management plan. Subsequently, store the adapted health management plan and its identification on the blockchain together with the patient's disease record status. To sum up, this process realizes a comprehensive, accurate and secure health management process by monitoring the patient's physical sign changes, generating corrected curves, evaluating deviations, adapting the health management plan, implementing adaptation identification, and blockchain storage. This process not only improves the efficiency and quality of health management, but also enhances the patient's trust and satisfaction with the health management plan.

[0044] The health management data processing method based on patients' dynamic feedback provided by the embodiments of the present invention has at least the following technical effects:

[0045] 1. By downloading multiple physical sign change curves stored on the blockchain that are consistent with the patient's current disease state and health management plan, it provides rich historical data support for health management. By traversing these curves and statistically calculating the confidence density, the physical sign change curve that best matches the patient's current condition can be extracted, thereby improving the accuracy of the health management plan. In addition, the information of the patient's dynamic feedback is also considered, the curve in the latter stage of the physical sign change is corrected, and whether the health management plan is suitable is judged according to the deviation between the correction result and the original curve. This personalized adjustment based on the patient's real-time feedback makes the health management plan more in line with the actual needs of the patient.

[0046] 2. Using blockchain technology to store the disease record status, health management plan, and physical sign change curves ensures the immutability and traceability of the data. By constructing a state comparison binary tree and a plan comparison binary tree, strict comparison and verification are performed on the downloaded data, and only the data that meets the consistency rules will be used for subsequent analysis, thereby enhancing the consistency and credibility of the data. This data verification mechanism not only improves the accuracy of health management decisions, but also provides reliable evidence support for the resolution of medical disputes.

[0047] 3. By integrating multiple physical sign change prediction units for output mean fusion, a more accurate physical sign change prediction model is obtained. This model can quickly generate a correction curve for the latter segment of the physical sign change based on the real-time feedback information of the patient, and conduct deviation analysis with the original curve, so as to timely judge the suitability of the health management plan. This process optimization not only shortens the adjustment cycle of the health management plan, but also improves the overall efficiency of health management. At the same time, when the health management plan needs to be adjusted, the plan can automatically execute the non-conformance label or conformance label, and conduct blockchain evidence storage with the diseased record status, providing continuity and traceability for subsequent health management.

[0048] Embodiment 2:

[0049] As Figure 2 shown, based on the same inventive concept as the health management data processing method based on the dynamic feedback of patients provided in Embodiment 1, the embodiment of the present invention further provides a health management data processing system based on the dynamic feedback of patients, and the system includes:

[0050] A physical sign monitoring module 11, configured to download from the chain multiple patient physical sign change curves whose diseased record status stored in the blockchain is consistent with the patient's diseased state and whose recorded health management plan is consistent with the health management plan.

[0051] A variable extraction module 12, configured to traverse multiple patient physical sign change curves to statistically calculate the ratio of the number of neighborhood curves to the mean value of the number of neighborhood curves, set it as multiple confidence densities, and extract the physical sign change curve with the maximum value of the multiple confidence densities.

[0052] A deviation correction module 13, configured to, when there is a health management intervention, statistically calculate the deviation vector between the physical sign change monitoring curve of the patient feedback information and the front-segment curve of the physical sign change, and correct the rear-segment curve of the physical sign change to obtain a corrected rear-segment curve of the physical sign change, where the front-segment curve of the physical sign change is a partial curve obtained by time sequence alignment of the physical sign change curve and the physical sign change monitoring curve.

[0053] A feedback management module 14, configured to, when the deviation between the corrected rear-segment curve of the physical sign change and the physical sign change curve is greater than or equal to the expected deviation, execute a non-conformance label feedback on the health management plan to the health management terminal.

[0054] Furthermore, the physical sign monitoring module 11 is further configured to perform the following steps:

[0055] Construct a status comparison binary tree according to the patient's disease state, construct a plan comparison binary tree according to the health management plan. Any level of the comparison binary tree represents a comparison index attribute and a consistency rule to be compared. Only when the upper-level comparison is consistent can the lower-level comparison be entered. When all levels of comparison are consistent, it is considered that the comparison binary tree is satisfied. When any level of comparison is inconsistent, it is considered that the comparison binary tree is not satisfied. Download from the chain the first disease record status, the first recorded health management plan, and the first patient physical sign change curve of the patient to be analyzed that have been stored on the blockchain. Input the first disease record status into the status comparison binary tree, input the first recorded health management plan into the plan comparison binary tree. When the output results are all consistent, it is considered that the disease record status is consistent with the patient's disease state, and the recorded health management plan is consistent with the health management plan. Add the first patient physical sign change curve to the multiple patient physical sign change curves.

[0056] Furthermore, the physical sign monitoring module 11 is further configured to perform the following steps:

[0057] Analyze the correlation between the patient's disease state indicators and the fluctuation amplitude of the physical sign change curve. Sort the patient's disease state indicators according to the Pearson correlation coefficient from large to small to obtain the sorting result of the patient's disease state indicators. Among them, for the indicators with the same Pearson correlation coefficient, they are randomly sorted continuously. According to the sorting result of the patient's disease state indicators, extract the first-order patient's disease state from the patient's disease state to construct the first-level comparison parameters to be compared. Among them, when the first-order patient's disease state indicator is a type indicator, the first-level consistency rule is: consistent if the types are the same, otherwise inconsistent. When the first-order patient's disease state indicator is a quantitative indicator, the first-level consistency rule is: consistent if the input data deviates from the first-order patient's disease state by less than or equal to the threshold, otherwise inconsistent. Construct the first level of the binary tree according to the first-level consistency rule and the first-level comparison parameters to be compared, until the Nth level of the binary tree is constructed, and fuse the first N levels to generate the status comparison binary tree.

[0058] Furthermore, the physical sign monitoring module 11 is further configured to perform the following steps:

[0059] Obtain the patient's disease state indicators and extract the first patient's disease state indicator. Use the fluctuation amplitude of the first patient's disease state indicator as the only independent variable and the fluctuation amplitude of the physical sign change curve as the dependent variable. Collect the fluctuation amplitude sequence of the first patient's disease state indicator and the fluctuation amplitude sequence of the physical sign change curve to perform Pearson correlation coefficient analysis to obtain the Pearson correlation coefficient of the first patient's disease state indicator. Add the Pearson correlation coefficient of the first patient's disease state indicator to the Pearson correlation coefficient.

[0060] Furthermore, the deviation correction module 13 is further configured to perform the following steps:

[0061] Obtain a physical sign change prediction unit, wherein the physical sign change prediction unit is generated by training a long short-term memory neural network with multiple groups of data, and any one of the multiple groups of data includes a physical sign change monitoring record curve, a physical sign change front-section record curve, and a label identifying the physical sign change back-section record curve; the deviation vector includes the first physical sign attribute deviation vector time series information of the physical sign change monitoring curve and the physical sign change front-section curve until the Mth physical sign attribute deviation vector time series information; statistically calculate the proportion of the first moment greater than or equal to the deviation modulus value in the first physical sign attribute deviation vector time series information until the proportion of the Mth moment greater than or equal to the deviation modulus value in the Mth physical sign attribute deviation vector time series information is statistically calculated; according to the number of the first moment proportion until the Mth moment proportion that is greater than or equal to the moment proportion, integrate at least twice the number of physical sign change prediction units for output mean fusion to obtain a physical sign change prediction model, and process the physical sign change monitoring curve and the physical sign change front-section curve of the patient feedback information to obtain the physical sign change back-section curve.

[0062] Furthermore, the deviation correction module 13 is further configured to perform the following steps:

[0063] Collect a physical sign change record curve and an initial physical sign change monitoring record curve, wherein the initial physical sign change monitoring record curve is aligned with the time series at both ends of the physical sign change record curve; perform binary segmentation on the initial physical sign change monitoring record curve to obtain the physical sign change monitoring record curve and the label identifying the physical sign change back-section record curve.

[0064] Furthermore, the deviation correction module 13 is further configured to perform the following steps:

[0065] When the deviation between the physical sign change back-section correction curve and the physical sign change curve is less than the expected deviation, perform an adaptation identification on the health management plan and perform blockchain evidence storage with the disease record status.

[0066] Through the foregoing detailed description of the health management data processing method based on patient dynamic feedback in this specification, those skilled in the art can clearly know the health management data processing system based on patient dynamic feedback in this embodiment. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description in the method part.

[0067] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for processing health management data based on patients' dynamic feedback, characterized in that, Including: Downloading multiple patient physical sign change curves from the chain that are stored on the blockchain and have the same disease record status as the patient's disease status, and the recorded health management plan is the same as the health management plan; Traversing multiple patient physical sign change curves to statistically calculate the ratio of the number of neighborhood curves to the average number of neighborhood curves, which is set as multiple confidence densities, and extracting the physical sign change curve with the maximum value of the multiple confidence densities; When there is a health management intervention, statistically calculating the deviation vector between the physical sign change monitoring curve of the patient's feedback information and the front-section curve of the physical sign change, and correcting the rear-section curve of the physical sign change to obtain a corrected rear-section curve of the physical sign change. The front-section curve of the physical sign change is a partial curve obtained by time-series alignment of the physical sign change curve and the physical sign change monitoring curve; When the deviation between the corrected rear-section curve of the physical sign change and the physical sign change curve is greater than or equal to the expected deviation, feedback an inadaptability identification of the health management plan to the health management terminal; Downloading multiple patient physical sign change curves from the chain that are stored on the blockchain and have the same disease record status as the patient's disease status, and the recorded health management plan is the same as the health management plan, including: Constructing a status comparison binary tree according to the patient's disease status, and constructing a plan comparison binary tree according to the health management plan. Any layer of the comparison binary tree represents a comparison index attribute and a consistency rule to be compared. Only when the upper-level comparison is consistent can the lower-level comparison be entered. When all levels of comparison are consistent, it is considered to meet the comparison binary tree. When any level of comparison is inconsistent, it is considered not to meet the comparison binary tree; Downloading the first disease record status, the first recorded health management plan, and the first patient physical sign change curve of the patient to be analyzed that are stored on the blockchain from the chain; Inputting the first disease record status into the status comparison binary tree, and inputting the first recorded health management plan into the plan comparison binary tree. When the output results are all consistent, it is considered that the disease record status is the same as the patient's disease status, and the recorded health management plan is the same as the health management plan, and adding the first patient physical sign change curve to the multiple patient physical sign change curves; When there is a health management intervention, statistically calculating the deviation vector between the physical sign change monitoring curve of the patient's feedback information and the front-section curve of the physical sign change, and correcting the rear-section curve of the physical sign change to obtain a corrected rear-section curve of the physical sign change, including: Obtaining a physical sign change prediction unit, where the physical sign change prediction unit is generated by training a long short-term memory neural network with multiple groups of data. Any one of the multiple groups of data includes a physical sign change monitoring record curve, a front-section record curve of the physical sign change, and a label indicating the rear-section record curve of the physical sign change; The deviation vector includes the first physical sign attribute deviation vector time-series information of the physical sign change monitoring curve and the front-section curve of the physical sign change until the Mth physical sign attribute deviation vector time-series information; Statistically calculating the proportion of the first moment greater than or equal to the deviation modulus value in the first physical sign attribute deviation vector time-series information until statistically calculating the proportion of the Mth moment greater than or equal to the deviation modulus value in the Mth physical sign attribute deviation vector time-series information; According to the number of ratios of moments from the first moment ratio to the M-th moment ratio that are greater than or equal to the moment ratio, integrate at least twice the number of physical sign change prediction units for output mean fusion to obtain a physical sign change prediction model, and process the physical sign change monitoring curve and the physical sign change front segment curve of the patient feedback information to obtain the physical sign change rear segment curve.

2. The method according to claim 1, wherein, Construct a state comparison binary tree according to the patient's disease state, including: Analyze the correlation between the patient's disease state indicators and the amplitude fluctuations of the physical sign change curve. According to the Pearson correlation coefficient from large to small, sort the patient's disease state indicators to obtain the sorting result of the patient's disease state indicators. Among them, for indicators with the same Pearson correlation coefficient, perform random continuous sorting; According to the sorting result of the patient's disease state indicators, extract the disease state of the first serial number patient from the patient's disease state and construct the first-level parameters to be compared; Among them, when the disease state indicator of the first serial number patient is a type indicator, the first-level consistency rule is: consistent if the types are the same, otherwise inconsistent. When the disease state indicator of the first serial number patient is a quantitative indicator, the first-level consistency rule is: consistent if the input data deviates from the disease state of the first serial number patient by less than or equal to the threshold, otherwise inconsistent; Construct the first level of the binary tree according to the first-level consistency rule and the first-level parameters to be compared, until the N-th level of the binary tree is constructed, and fuse the first N levels to generate the state comparison binary tree.

3. The method according to claim 1, characterized in that Analyze the correlation between the patient's disease state indicators and the amplitude fluctuations of the physical sign change curve. According to the Pearson correlation coefficient, including: Obtain the patient's disease state indicators and extract the first patient's disease state indicator; Use the amplitude fluctuation of the first patient's disease state indicator as the only independent variable and the amplitude fluctuation of the physical sign change curve as the dependent variable. Collect the amplitude fluctuation sequence of the first patient's disease state indicator and the amplitude fluctuation sequence of the physical sign change curve to perform Pearson correlation coefficient analysis to obtain the Pearson correlation coefficient of the first patient's disease state indicator; Add the Pearson correlation coefficient of the first patient's disease state indicator to the Pearson correlation coefficient.

4. The method according to claim 1, characterized in that Any one of the multiple groups of data includes a physical sign change monitoring record curve, a physical sign change front segment record curve, and a label indicating the physical sign change rear segment record curve, including: Collect the physical sign change record curve and the initial physical sign change monitoring record curve, where the initial physical sign change monitoring record curve is aligned with both time series ends of the physical sign change record curve; Perform two-segment segmentation on the initial physical sign change monitoring record curve to obtain the physical sign change monitoring record curve and the label indicating the physical sign change rear segment record curve.

5. The method according to claim 1, characterized in that, Also include: When the deviation between the corrected rear segment curve of the physical sign change and the physical sign change curve is less than the expected deviation, perform an adaptation label on the health management plan and conduct blockchain evidence storage with the disease record status.

6. A health management data processing system based on dynamic feedback from patients, characterized in that, For implementing the health management data processing method based on patient dynamic feedback according to any one of claims 1-5, the system includes: The sign monitoring module is used to download multiple patient sign change curves from the chain, where the diseased record status stored on the blockchain is consistent with the patient's diseased status, and the recorded health management plan is consistent with the health management plan; The variable extraction module is used to traverse multiple patient sign change curves to statistically calculate the ratio of the number of neighborhood curves to the mean of the number of neighborhood curves, which is set as multiple confidence densities, and extract the sign change curve with the maximum value of the multiple confidence densities; The deviation correction module is used to, when there is a health management intervention, statistically calculate the deviation vector between the sign change monitoring curve of the patient feedback information and the sign change front-segment curve, and correct the sign change back-segment curve to obtain the sign change back-segment corrected curve. The sign change front-segment curve is a partial curve obtained by time-sequentially aligning the sign change curve and the sign change monitoring curve; The feedback management module is used to, when the deviation between the sign change back-segment corrected curve and the sign change curve is greater than or equal to the expected deviation, feedback an inadaptable identification of the health management plan to the health management terminal.

Images