Medical platform user follow-up visit management system and method based on Internet of Things

Abstract

The invention relates to the technical field of remote patient monitoring, in particular to a medical platform user follow-up visit management system and method based on the Internet of Things, and the system comprises a data collection module which extracts timestamp information and behavior data from Internet of Things equipment, combines the behavior data of each patient with a unique identifier, and sends the behavior data to a server; and recording the collection time and quantity of each data point, and generating a data integration table. According to the method, the subtle change of the health condition is effectively captured by monitoring the behavior change rate and trend of the patient in real time, the influence of the behavior mode is accurately evaluated through the correlation analysis of the characteristic value and the health result, in addition, the future behavior trend can be predicted through the prediction model, and the prediction accuracy is improved. According to the method, more personalized and timely treatment suggestions are provided for the patient, the adaptability of the treatment scheme is improved, the monitoring task is optimized, the speed and the accuracy of medical response are improved, and the treatment efficiency and the patient satisfaction degree are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of remote patient monitoring, and in particular to a medical platform user follow-up management system and method based on the Internet of Things. Background Art

[0002] The field of remote patient monitoring technology involves the use of information technology and communication devices to track the health status of patients, who may be anywhere outside of a medical institution. The core of this technology is to collect health data such as vital signs such as heart rate, blood pressure, blood sugar, etc. in real time through Internet of Things (IoT) devices such as sensors, wearable devices, and other endpoint devices, and send them to medical providers via wireless networks. This enables doctors and caregivers to monitor the patient's condition in real time and provide remote diagnosis or intervention when necessary. This technology also includes the use of advanced data analytics and machine learning technologies to analyze health data to predict the progression of the disease and optimize treatment plans.

[0003] Among them, the user follow-up management system of the medical platform based on the Internet of Things refers to a medical health management system implemented using the Internet of Things technology. Its main purpose is to conduct continuous health monitoring and management of patients with chronic diseases or patients in recovery. The system automatically collects the patient's health data and transmits it to the medical service provider by connecting to the smart device worn by the patient or the medical monitoring equipment at home. This system enables medical service providers to remotely monitor the patient's health status and adjust the treatment plan in time. It also helps patients to self-manage their diseases, reduce the need for hospital follow-up visits, and improve treatment efficiency and patient satisfaction.

[0004] Existing technologies usually rely on intermittent data collection and lack real-time monitoring and prediction capabilities, resulting in untimely responses to changes in patient behavior and health, increasing the risk of follow-up visits and hospitalizations. For example, patients with intermittent monitoring of arrhythmias may miss critical intervention opportunities. In addition, the shortcomings of existing technologies in data analysis and future behavior prediction limit the adjustment and foresight of treatment plans, which may lead to reduced treatment effectiveness and patient satisfaction. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a medical platform user follow-up management system and method based on the Internet of Things.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solution: A medical platform user follow-up management system based on the Internet of Things includes: The data collection module extracts timestamp information and behavioral data from IoT devices, combines each patient's behavioral data with a unique identifier, records the collection time and quantity of each data point, and generates a data integration table; The behavior analysis module extracts the behavior data of consecutive time points from the data integration table, calculates the change rate of each patient's behavior, and obtains the change trend of each behavior pattern by comparing the differences between adjacent data points, thereby obtaining the behavior change characteristic value; The impact assessment module uses the behavior change characteristic value to assess the impact of each behavior change on the patient's health status, calculates the sensitivity of the health change through correlation analysis between each behavior pattern and health results, and generates a health impact assessment index; The behavior prediction module uses the health impact assessment indicators and the patient's behavior data to analyze the behavior trends that will appear in the future, establishes a regression analysis model to predict the patient's behavior changes at multiple time points in the future, and generates future behavior prediction results; The multi-task processing module adjusts the priority of the monitoring tasks in real time according to the future behavior prediction results, matches the response plan for each patient behavior, evaluates the execution efficiency of multiple tasks, optimizes the task execution process, and establishes task optimization indicators.

[0007] The data integration table specifically includes timestamp information, patient unique identifier, behavioral data record, data point collection time, and number of data points. The behavioral change characteristic value specifically includes change rate, difference between adjacent data points, and behavior pattern change trend. The health impact assessment indicators include behavior pattern and health outcome correlation analysis, and health change sensitivity. The future behavior prediction results specifically include regression analysis model and future multi-time point behavior change prediction. The task optimization indicators include monitoring task priority adjustment, matching response plan, execution efficiency evaluation, and task execution process optimization.

[0008] As a further solution of the present invention, the steps of obtaining the data integration table are specifically as follows: Extract behavioral data and timestamps of multiple patients from IoT devices, combine with unique identifiers, parse the identification structure of data points, verify the uniqueness of identifiers and timestamp format, and obtain primary data tables; According to the information in the primary data table, record the number and time of each data point, parse the periodic characteristics of the timestamp and perform frequency analysis on the behavior data, screen abnormal data and perform time series normalization processing to generate a complete data record table; Extract key data from the complete data record table, analyze the time correlation characteristics between data points, and calibrate the time behavior weight based on the time offset factor, using the formula: ; Calculate the weighted average to obtain the time behavior weighted table; in, represents the total number of data points, represents the number of behaviors of the ith data point, Indicates the corresponding timestamp, represents the time bias factor, Represents the time-weighted average of behaviors, which is used to evaluate the time-weighted contribution of the behaviors of each data point in the overall data set; The data in the time behavior weighted table is integrated and unified, the distribution pattern of the weighted data is analyzed and classified and standardized, the unified time and behavior data format is adjusted to conform to the integration rules, and a data integration table is created.

[0009] As a further solution of the present invention, the step of obtaining the behavior change characteristic value is specifically as follows: Extracting the behavioral data of each patient at consecutive time points from the data integration table, ensuring data continuity and excluding missing or abnormal data, to obtain a purified continuous behavioral data table; Based on the purified continuous behavior data table, calculating the change rate of each patient's behavior data, and obtaining a behavior change rate table by calculating the difference between adjacent data points point by point; Analyze the data in the behavior change rate table using the formula: ; Add the impact of abnormal behavior to obtain the weighted behavior change feature table, Represents the rate of change of behavior, which is an indicator used to measure how fast a behavior changes over time. Indicated in Behavioral data at each time point is used to Compare to calculate the change, Indicates Time point to The absolute value of the behavior change at a time point is used to obtain the magnitude of the change, whether positive or negative. is the normalization coefficient used to average the calculation of the behavior change rate to each time interval; in, represents the total number of data points, Indicated in Behavioral data at each time point, Represents the weight of each data point adjusted based on the time interval, which is used to strengthen the influence of data points with short time intervals; Extract key statistical indicators from the weighted behavior change feature table, calculate the average change rate and standard deviation, combine the data to evaluate the change trend of each behavior pattern, and obtain the behavior change feature value.

[0010] As a further solution of the present invention, the steps for obtaining the health impact assessment index are specifically as follows: Based on the behavior change characteristic value, the data of each behavior pattern is integrated, and atypical behavior records are excluded through data cleaning to obtain cleaned behavior characteristic data; Applying a statistical model to the cleaned behavioral feature data to analyze the correlation between multiple behavioral patterns and patient health outcomes, evaluating the impact intensity of the behavioral patterns through correlation coefficients, and obtaining behavioral health correlation analysis results; Based on the behavioral health association analysis results, the formula is used: ; Calculate the sensitivity of each behavior to health status and obtain a health sensitivity analysis table; in, is the number of behavior patterns, Indicates behavior pattern Impact weights, weighted health impacts of behaviors, It is behavior The direct impact coefficient on health, is a constant used to adjust the stability of the calculation, stands for health sensitivity, which is used to quantify the overall impact of multiple behaviors on health status; Based on the health sensitivity analysis table, the average value and coefficient of variation are used to summarize the impact of each behavior pattern on health and generate a health impact assessment index.

[0011] As a further solution of the present invention, the step of obtaining the future behavior prediction result is specifically as follows: Integrate the health impact assessment indicators with the patient's behavioral data, determine the statistical correlation between behavioral patterns and health outcomes through data cross-validation, and obtain a behavior-health correlation model; Based on the behavior and health association model, a multivariate linear regression analysis is designed and implemented to predict the patient's behavior pattern in the future time period, and the regression coefficient is adjusted through data to obtain an adjusted prediction model; Apply the adjusted prediction model, combined with real-time behavioral data, using the formula: ; Computationally generate behavioral predictions at multiple time points in the future; in, Indicates a future time point The behavioral prediction value of is the baseline forecast value, Represents the influence weight of multiple behavioral variables, is the exponential form of the behavior variable, which increases the sensitivity of the model to changes in behavior. It represents the number of behavioral variables, that is, the total number of multiple behavioral factors referenced in the model; Based on the behavior predictions at multiple future time points, key behavior trends and change predictions are extracted from the model output, the health impact potential of the prediction results is analyzed, and future behavior prediction results are output.

[0012] As a further solution of the present invention, the step of obtaining the task optimization index is specifically as follows: Analyze the future behavior prediction results, determine the high-frequency behavior pattern, and set the priority of the monitoring task according to the pattern to obtain a priority allocation plan; According to the priority allocation scheme, a targeted response scheme is designed, and the potential effects of multiple schemes are verified through actual data simulation to obtain the efficiency evaluation results of multiple response schemes; Using the data from the efficiency evaluation results, the formula is used: ; Calculate and generate comprehensive indicators of execution efficiency and response speed for each task; in, Indicates the task The optimization index of is the efficiency score of task execution, Indicates the response speed score, Represents the adjustment factor of the efficiency score, adjustment The influence of It is a constant offset, which ensures that the denominator is not zero and enhances the stability of the calculation; The comprehensive indicators of execution efficiency and response speed of each task are analyzed, the task flow of the monitoring system is adjusted according to actual operation requirements, and task optimization indicators are integrated and generated.

[0013] A method for follow-up management of medical platform users based on the Internet of Things, the method for follow-up management of medical platform users based on the Internet of Things is executed based on the above-mentioned medical platform user follow-up management system based on the Internet of Things, and comprises the following steps: S1: Extract patient behavior data with timestamps from IoT devices, perform data merging, combine behavior data with the patient’s unique identifier, record the data collection time and quantity, and generate a data consolidation table; S2: extracting the behavioral data of consecutive time points according to the data integration table, calculating the change rate of each patient's behavior, determining the change trend of the behavior pattern by comparing the differences between adjacent data points, and generating a behavior characteristic value result; S3: using the behavior feature value results, evaluating the impact of the feature value on the patient's health status, performing correlation analysis between the behavior pattern and the health outcome, thereby calculating the sensitivity of health changes and generating a health impact assessment indicator; S4: Based on the health impact assessment indicators and combined with the patient's behavioral data, a regression analysis model is established to predict patient behavior changes at multiple time points in the future, generate future behavior prediction results, and adjust the priority of monitoring tasks in real time according to the future behavior prediction results, match response plans for each patient behavior, and optimize the task execution process.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by real-time monitoring of the rate and trend of patient behavior changes, subtle changes in health status can be effectively captured, and the impact of behavioral patterns can be accurately evaluated through correlation analysis of characteristic values ​​and health outcomes. In addition, through the prediction model, the solution can predict future behavioral trends and provide patients with more personalized and timely treatment recommendations. This prediction not only improves the adaptability of the treatment plan, but also optimizes the monitoring task, improves the speed and accuracy of medical response, and significantly improves treatment efficiency and patient satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a system flow chart of the present invention; Figure 2 A flowchart of the steps for obtaining the data integration table of the present invention; Figure 3 A flowchart of the steps for obtaining the behavior change characteristic value of the present invention; Figure 4 A flowchart of the steps for obtaining the health impact assessment index of the present invention; Figure 5 A flow chart of the steps for obtaining the future behavior prediction results of the present invention; Figure 6 The present invention is a flowchart of the steps for obtaining the task optimization index. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0017] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0018] Example 1: Please refer to Figure 1 The present invention provides a technical solution: a medical platform user follow-up management system based on the Internet of Things includes: The data collection module extracts timestamp information and behavioral data from IoT devices, combines each patient's behavioral data with a unique identifier, records the collection time and quantity of each data point, and generates a data integration table; The behavior analysis module extracts the behavior data of consecutive time points from the data integration table, calculates the change rate of each patient's behavior, and obtains the change trend of each behavior pattern by comparing the differences between adjacent data points, thereby obtaining the behavior change characteristic value; The impact assessment module uses the behavioral change characteristic values ​​to assess the impact of each behavioral change on the patient's health status. By analyzing the association between each behavioral pattern and health outcomes, it calculates the sensitivity of health changes and generates health impact assessment indicators. The behavior prediction module uses health impact assessment indicators and patient behavior data to analyze future behavior trends, establishes a regression analysis model to predict patient behavior changes at multiple time points in the future, and generates future behavior prediction results; The multi-task processing module adjusts the priority of monitoring tasks in real time according to the future behavior prediction results, matches response plans for each patient behavior, evaluates the execution efficiency of multiple tasks, optimizes the task execution process, and establishes task optimization indicators.

[0019] The data integration table specifically includes timestamp information, patient unique identifier, behavioral data records, data point collection time, and number of data points. The behavioral change characteristic values ​​specifically include change rate, difference between adjacent data points, and behavioral pattern change trend. Health impact assessment indicators include correlation analysis between behavioral patterns and health outcomes, and sensitivity to health changes. Future behavior prediction results specifically include regression analysis models and predictions of behavioral changes at multiple time points in the future. Task optimization indicators include monitoring task priority adjustment, matching response plans, execution efficiency evaluation, and task execution process optimization.

[0020] See also Figure 2 , the specific steps for obtaining the data integration table are: Extract behavioral data and timestamps of multiple patients from IoT devices, combine with unique identifiers, parse the identification structure of data points, verify the uniqueness of identifiers and timestamp format, and obtain primary data tables; Parse the identification structure of the data point and verify the uniqueness of the identifier and the timestamp format, obtain device data through the network communication protocol, standardize the data points to ensure the consistency and traceability of the data, and analyze the integrity and accuracy of the data through the correspondence between the timestamp and identifier of the data point. These data are then used to form the primary data table.

[0021] Based on the information in the primary data table, record the number and time of each data point, parse the periodic characteristics of the timestamp and perform frequency analysis on the behavior data, screen abnormal data and perform time series normalization to generate a complete data record table; By analyzing the periodic characteristics of timestamps, using time series analysis methods to perform frequency analysis on behavioral data, using statistical methods to screen and process abnormal data points, and normalizing the data to improve data consistency and comparability, this series of processing can enhance data accuracy and availability and generate a complete data record table.

[0022] Extract key data from the complete data record table, analyze the time correlation characteristics between data points, and calibrate the time behavior weight based on the time offset factor, using the formula: ; Calculate the weighted average to get the time behavior weighted table, Represents the time-weighted average of behaviors, which is used to evaluate the time-weighted contribution of the behaviors of each data point in the overall data set; in, represents the total number of data points, represents the number of behaviors of the ith data point, Indicates the corresponding timestamp, represents the time bias factor; formula: ; The benefit of the formula is that by introducing the time bias factor And exponential smoothing, adjusts the weight of the timestamp of the behavior data point to the number of behaviors, increases the impact of timestamp accuracy on behavior analysis, and improves the time sensitivity and response speed of the data.

[0023] Detailed explanation of the formula and the process of formula calculation and derivation: There are 3 data points, , They are 10, 15, 20, 1, 2, 3 hours respectively. They are 0.5, 0.3, and 0.1 respectively. The contribution of the first data point is: ; Contribution of the second data point: ; Contribution of the third data point: ; The results show that by referring to the weighting of time bias and number of behaviors, the distribution of behavioral data in the time dimension can be more accurately measured. For the step results, it means that a time-weighted behavioral data table can be obtained. The data table can better reflect the time distribution characteristics of the behavior and provide a basis for subsequent data analysis.

[0024] Integrate and unify the data in the time behavior weighted table, analyze the distribution pattern of the weighted data and classify and standardize it, adjust the unified time and behavior data format to comply with the integration rules, and create a data integration table.

[0025] The data in the time-behavior weighted table is integrated and unified by analyzing the distribution pattern of the weighted data and classifying and standardizing it, and adjusting the unified time and behavior data format to comply with the integration rules. These steps are automatically completed through program scripts, ensuring the accuracy and efficiency of the data integration process. The created data integration table meets the predetermined data standards and format requirements.

[0026] See also Figure 3 , the specific steps for obtaining the behavior change characteristic value are: Extract the behavioral data of each patient at consecutive time points from the data integration table to ensure data continuity and exclude missing or abnormal data, and obtain a purified continuous behavioral data table; In the process of extracting the behavioral data of each patient at consecutive time points from the data integration table to ensure data continuity and exclude missing or abnormal data in time, the data integrity of the data table is first verified, and any missing data or abnormal values ​​are excluded through the data verification algorithm. This involves checking the continuity of the data points in the time series, and using threshold analysis to identify and exclude data points that exceed normal behavioral patterns. Through these steps, it can be ensured that the extracted behavioral data has a high degree of reliability and integrity in the time series, and a purified continuous behavioral data table is obtained.

[0027] Based on the purified continuous behavior data table, the change rate of each patient's behavior data is calculated, and the behavior change rate table is obtained by calculating the difference between adjacent data points point by point; First, the difference algorithm is applied to the time series data to solve the difference between consecutive data points. This calculation involves subtracting the behavioral data of each data point from the previous data point to obtain the behavior change rate. These differential values ​​reflect the instantaneous rate of behavior change. In this way, the change rate of the patient's behavior at each time point can be obtained, and a behavior change rate table can be obtained.

[0028] Analyze the data in the behavior change rate table using the formula: ; Add the impact of abnormal behavior to obtain a weighted behavior change feature table; in, represents the total number of data points, Indicated in Behavioral data at each time point, Represents the weight of each data point adjusted based on the time interval, which is used to strengthen the influence of data points with short time intervals. Represents the rate of change of behavior, which is an indicator used to measure how fast a behavior changes over time. Indicated in Behavioral data at each time point is used to Compare to calculate the change, Indicates Time point to The absolute value of the behavior change at a time point is used to obtain the magnitude of the change, whether positive or negative. is the normalization coefficient used to average the calculation of the behavior change rate to each time interval; formula: ; The benefit of the formula is that, through weighted processing, it emphasizes the criticality of behavioral changes, especially the more prominent change points in time series data, and enhances the model's sensitivity to behavioral mutations, which is particularly critical for analyzing changing trends in patient behavior patterns.

[0029] Detailed explanation of the formula and the process of formula calculation and derivation: First determine the total number of data points =100, indicating the total number of behavioral data points of the patient during the observation period. The behavioral value of each data point and Can be obtained from the continuous behavior data table, set in the and The behavioral data at time points are 5 and 7. is adjusted based on the time interval of the data points and is set to 1.2. Substituting these values ​​into the formula: ; The results showed that the average rate of behavioral change at each time point was 2.4. This result helps to iteratively analyze the changing trends and patterns of patients' behaviors, thereby providing data support for subsequent behavioral interventions.

[0030] Extract key statistical indicators from the weighted behavior change feature table, calculate the average change rate and standard deviation, combine the data to evaluate the change trend of each behavior pattern, and obtain the behavior change feature value.

[0031] Through statistical analysis methods, we first calculate the mean and standard deviation of all change rates in the weighted behavior change characteristic table, including summing up each change rate value and then dividing it by the total number of data points to get the average change rate. Then, we calculate the square of the difference between each rate value and the average value, sum it up and divide it by the number of data points minus one to get the variance, and finally take the square root to get the standard deviation. This series of statistical calculations helps to quantitatively describe the normality and volatility of behavior changes, thereby deriving the changing trend of each behavior pattern and obtaining the behavior change characteristic value.

[0032] See also Figure 4 , the specific steps for obtaining health impact assessment indicators are as follows: Based on the behavior change characteristic value, the data of each behavior pattern is integrated, and the atypical behavior records are excluded through data cleaning to obtain the cleaned behavior characteristic data; Collect behavioral change feature values. In order to maintain data consistency and integrity, remove abnormal data points caused by equipment failure or external interference during data cleaning. Verify the consistency of timestamps of behavioral records to ensure that the analyzed data is not affected by time errors. Standardize the original data to remove potential biases and errors, thereby ensuring the accuracy and reliability of behavioral data. Data processing at this stage provides a solid foundation for subsequent analysis.

[0033] Apply statistical models to the cleaned behavioral feature data to analyze the correlation between multiple behavioral patterns and patient health outcomes, evaluate the impact of behavioral patterns through correlation coefficients, and obtain behavioral health correlation analysis results; Statistical models are used for analysis, and the correlation coefficients between differentiated behavioral patterns and health outcomes are calculated to quantitatively assess the impact of behavior. In this process, not only conventional correlation analysis is relied upon, but multivariate statistical analysis techniques are introduced to handle a variety of potential confounding variables to ensure that the obtained correlation coefficients reflect the true correlation between behavioral changes and health outcomes. The calculation of these correlation coefficients is based on data processing and mathematical models, which makes the analysis results highly reliable and of reference value.

[0034] Based on the results of behavioral health association analysis, the formula is used: ; Calculate the sensitivity of each behavior to health status and obtain a health sensitivity analysis table; in, is the number of behavior patterns, Indicates behavior pattern Impact weights, weighted health impacts of behaviors, It is behavior The direct impact coefficient on health, is a constant used to adjust the stability of the calculation, stands for health sensitivity, which is used to quantify the overall impact of multiple behaviors on health status; formula: ; The benefit of the formula is that by introducing the weight and health impact factor And the constant term , not only refers to the impact strength of each behavior, but also increases the sensitivity to small values ​​through square and root processing, improving the model's responsiveness to small changes in health effects.

[0035] Detailed explanation of the formula and the process of formula calculation and derivation: There are three behavior modes, corresponding to the weights , health impact coefficient ,constant , first calculate the sum of the squares of each term: The value after adding the constant is: , take the square root of each term and we get: ; ; ; Apply the weights to calculate the result: ; The results show that after integrating all behavioral patterns, the overall sensitivity of behavior to health is 1.0933, which indicates that behavioral changes have a slight but significant impact on health status, and the numerical value can be used for iterative analysis and health risk assessment.

[0036] The comprehensive health sensitivity analysis table uses mean values ​​and coefficients of variation to summarize the impact of each behavioral pattern on health and generate health impact assessment indicators.

[0037] Based on the health sensitivity analysis table, in order to extract key statistical indicators, it is necessary to group and sort the sensitivity values ​​of each behavior pattern, analyze the distribution characteristics of the sensitivity by calculating the mean value and coefficient of variation within the group, and combine the time series trend chart of the behavior pattern to verify the sensitivity fluctuation of the behavior pattern in differentiated time periods. Finally, the statistical data are summarized and organized, and a health risk assessment model is established to generate health impact assessment indicators.

[0038] See also Figure 5 , the specific steps for obtaining the future behavior prediction results are: Integrate health impact assessment indicators with patient behavior data, determine the statistical correlation between behavior patterns and health outcomes through data cross-validation, and obtain a behavior-health correlation model; Integrate health impact assessment indicators with patients' behavioral data. In order to ensure the reliability of the data, first perform data cleaning to remove outliers and missing data. Then use statistics to calculate the correlation coefficient between behavioral data and health impacts to evaluate the intensity of the impact of differentiated behavioral patterns on health. Based on this analysis result, establish a preliminary health impact model. This model refers to the potential impact of behavioral frequency and duration on health. Through regression analysis of behavioral data, determine which behavioral patterns are significantly correlated with changes in health status. This analysis process not only assists medical providers in understanding patients' behavioral health patterns, but also provides data support for subsequent health interventions.

[0039] Based on the behavior and health association model, a multivariate linear regression analysis was designed and implemented to predict the patient's behavior pattern in the future time period, and the regression coefficient was adjusted through data to obtain an adjusted prediction model; On the basis of the behavior and health association model, the currently collected behavioral data is used to predict the patient's behavioral trends in future time periods through a multiple linear regression model. This regression model includes time series analysis, referring to the dynamic changes of behavioral patterns over time. The construction of the prediction model first includes determining the model variables, including behavior frequency and duration, and then encoding and normalizing the variables to match the model requirements. The model parameters are optimized through cross-validation methods to obtain a model that can reflect the patient's future behavior, provide decision support for medical intervention, and assist in predicting health risks.

[0040] Apply the adjusted prediction model, combined with real-time behavioral data, using the formula: ; Computationally generate behavioral predictions at multiple time points in the future; in, Indicates a future time point The behavioral prediction value of is the baseline forecast value, Represents the influence weight of multiple behavioral variables, is the exponential form of the behavior variable, which increases the sensitivity of the model to changes in behavior. It represents the number of behavioral variables, that is, the total number of multiple behavioral factors referenced in the model; formula: ; The benefit of the formula is that by introducing the square term of behavioral data, the model's sensitivity to changes in behavioral frequency and intensity is enhanced, and it can more accurately predict how behavioral patterns change over time and their impact on health status.

[0041] Detailed explanation of the formula and the process of formula calculation and derivation: In the setting scenario, there are three behavior modes, and the coefficient of the multi-mode is , at the time point , behavioral index The calculation steps are as follows: 1. Calculate multiple behavior pattern pairs Contributions: ; ; ; 2. Add up all contributions and add the intercept : ; The results showed that: at the time point , the predicted behavior pattern index is 297.5, which indicates that as the behavior index increases, the predicted health risk also increases accordingly. This value can assist medical experts in assessing the patient's future health status and adjust the treatment plan accordingly.

[0042] Based on behavioral predictions at multiple time points in the future, key behavioral trends and change predictions are extracted from the model output, the health impact potential of the prediction results is analyzed, and future behavior prediction results are output.

[0043] By preprocessing the patient behavior data collected in real time, including data normalization and outlier screening, we ensure that the data matches the model. Then, we extract the behavior pattern variables from the processed data and input them into the adjusted prediction model. The model calculates the patient's behavior patterns at multiple time points in the future through regression coefficient weights and time series distribution to generate prediction values. These prediction values ​​are then statistically analyzed in time segments to evaluate the changing trends of behavior patterns in differentiated time periods and derive future behavior prediction results.

[0044] See also Figure 6 , the specific steps for obtaining task optimization indicators are: Analyze the future behavior prediction results, determine the high-frequency behavior patterns, and set the priority of monitoring tasks according to the patterns to obtain the priority allocation plan; Based on patient behavior data and monitoring records, a clustering algorithm is used to classify behavior patterns and calculate the frequency of occurrence of each type of behavior. Behaviors with high frequency are classified as high priority, while behaviors with low frequency are classified as low priority. The clustering algorithm used in this process refers to the fit of the time series to ensure the recognition of behavior patterns. Through statistical analysis methods, the frequency of occurrence of differentiated behavior patterns is calculated, which provides a basis for setting the priority of monitoring tasks and thus obtaining a priority allocation plan.

[0045] Design targeted response plans based on the priority allocation plan, verify the potential effects of multiple plans through actual data simulation, and obtain the efficiency evaluation results of multiple response plans; Differentiated interventions are designed, including more frequent monitoring and immediate interventions for high-priority behavioral patterns and routine monitoring for low-priority behavioral patterns. Each response plan is simulated and tested based on previous patient response data and intervention effects to ensure the feasibility and effectiveness of each plan. Data simulation is used to predict the potential effects of multiple plans, including the patient's response time to interventions and the duration of the effects. These data are obtained through analysis of monitoring data, providing a scientific basis for selecting the most matching response plan.

[0046] Using the data from the efficiency evaluation results, the formula is: ; Calculate and generate comprehensive indicators of execution efficiency and response speed for each task; in, Indicates the task The optimization index of is the efficiency score of task execution, Indicates the response speed score, Represents the adjustment factor of the efficiency score, adjustment The influence of It is a constant offset, which ensures that the denominator is not zero and enhances the stability of the calculation; formula: ; The formula is useful in that it introduces the efficiency score and responsiveness score , and the adjustment factor and , providing a comprehensive reference method for task execution efficiency and response speed. By adjusting these parameters, the speed and efficiency requirements of differentiated tasks can be flexibly matched. The square root and absolute value operations in the formula increase the sensitivity to data anomalies, making the evaluation results more stable and reliable.

[0047] Detailed explanation of the formula and the process of formula calculation and derivation: Settings , , Substituting the formula into the calculation steps is as follows: 1. Calculation ; 2. Calculate the denominator ; 3. Calculate the entire fraction ; The results show that: under given efficiency and response speed scores, the task The comprehensive optimization index is about 0.698, which indicates that the task has achieved a good balance between response speed and execution efficiency. The larger the value of the optimization index, the more efficient the task execution and the faster the response.

[0048] Analyze the comprehensive indicators of execution efficiency and response speed of each task, adjust the task flow of the monitoring system according to actual operational needs, and integrate and generate task optimization indicators.

[0049] The value of the comprehensive index directly affects the adjustment of tasks and resource allocation. Tasks with high index values ​​will obtain more resources and be executed first, while tasks with low index values ​​will be delayed or rescheduled. This analysis process involves data normalization and priority sorting. By comparing the comprehensive index values ​​of differentiated tasks, administrators can identify tasks that need to be prioritized and optimize the task flow of the entire monitoring system. This not only improves the operating efficiency of the system, but also ensures that key tasks receive sufficient attention and resources. It integrates and generates task optimization indicators to provide a basis for real-time adjustment and resource allocation.

[0050] A method for follow-up management of medical platform users based on the Internet of Things, the method for follow-up management of medical platform users based on the Internet of Things is executed based on the above-mentioned medical platform user follow-up management system based on the Internet of Things, and comprises the following steps: S1: Extract patient behavior data with timestamps from IoT devices, perform data merging, combine behavior data with the patient’s unique identifier, record the data collection time and quantity, and generate a data consolidation table; S2: According to the data integration table, extract the behavioral data of consecutive time points, calculate the change rate of each patient's behavior, determine the change trend of the behavior pattern by comparing the differences between adjacent data points, and generate the behavioral characteristic value results; S3: Use the behavioral feature value results to evaluate the impact of the feature value on the patient's health status, perform correlation analysis between behavioral patterns and health outcomes, thereby calculating the sensitivity of health changes and generating health impact assessment indicators; S4: Based on the health impact assessment indicators and combined with the patient's behavioral data, a regression analysis model is established to predict the patient's behavioral changes at multiple time points in the future, generate future behavior prediction results, and adjust the priority of the monitoring task in real time according to the future behavior prediction results, match the response plan for each patient behavior, and optimize the task execution process.

[0051] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A medical platform user follow-up management system based on the Internet of Things, characterized in that: The system comprises: The data collection module extracts timestamp information and behavioral data from IoT devices, combines each patient's behavioral data with a unique identifier, records the collection time and quantity of each data point, and generates a data integration table; The behavior analysis module extracts the behavior data of consecutive time points from the data integration table, calculates the change rate of each patient's behavior, and obtains the change trend of each behavior pattern by comparing the differences between adjacent data points, thereby obtaining the behavior change characteristic value; The impact assessment module uses the behavior change characteristic value to assess the impact of each behavior change on the patient's health status, calculates the sensitivity of the health change through correlation analysis between each behavior pattern and health results, and generates a health impact assessment index; The behavior prediction module uses the health impact assessment indicators and the patient's behavior data to analyze the future behavior trends, establishes a regression analysis model to predict the patient's behavior changes at multiple time points in the future, and generates future behavior prediction results; The multi-task processing module adjusts the priority of the monitoring tasks in real time according to the future behavior prediction results, matches the response plan for each patient behavior, evaluates the execution efficiency of multiple tasks, optimizes the task execution process, and establishes task optimization indicators.

2. The user follow-up management system for a medical platform based on the Internet of Things according to claim 1, characterized in that: The data integration table specifically includes timestamp information, patient unique identifier, behavioral data record, data point collection time, and number of data points. The behavioral change characteristic value specifically includes change rate, difference between adjacent data points, and behavior pattern change trend. The health impact assessment indicators include behavior pattern and health outcome correlation analysis, and health change sensitivity. The future behavior prediction results specifically include regression analysis model and future multi-time point behavior change prediction. The task optimization indicators include monitoring task priority adjustment, matching response plan, execution efficiency evaluation, and task execution process optimization.

3. The user follow-up management system for a medical platform based on the Internet of Things according to claim 2, characterized in that: The steps for obtaining the data integration table are specifically as follows: Extract behavioral data and timestamps of multiple patients from IoT devices, combine with unique identifiers, parse the identification structure of data points, verify the uniqueness of identifiers and timestamp format, and obtain primary data tables; According to the information in the primary data table, record the number and time of each data point, analyze the periodic characteristics of the timestamp and perform frequency analysis on the behavior data, screen abnormal data and perform time series normalization processing to generate a complete data record table; Extract key data from the complete data record table, analyze the time correlation characteristics between data points, and calibrate the time behavior weight based on the time offset factor, using the formula: ； Calculate the weighted average to obtain the time behavior weighted table; in, represents the total number of data points, represents the number of behaviors of the ith data point, Indicates the corresponding timestamp, represents the time bias factor, Represents the time-weighted average of behaviors, which is used to evaluate the time-weighted contribution of the behaviors of each data point in the overall data set; The data in the time behavior weighted table is integrated and unified, the distribution pattern of the weighted data is analyzed and classified and standardized, the unified time and behavior data format is adjusted to conform to the integration rules, and a data integration table is created.

4. The user follow-up management system for a medical platform based on the Internet of Things according to claim 3 is characterized in that: The steps for obtaining the behavior change characteristic value are specifically as follows: Extracting the behavioral data of each patient at consecutive time points from the data integration table, ensuring data continuity and excluding missing or abnormal data, to obtain a purified continuous behavioral data table; Based on the purified continuous behavior data table, calculating the change rate of each patient's behavior data, and obtaining a behavior change rate table by calculating the difference between adjacent data points point by point; Analyze the data in the behavior change rate table using the formula: ； Add the impact of abnormal behavior to obtain a weighted behavior change feature table; in, represents the total number of data points, Indicated in Behavioral data at each time point, Represents the weight of each data point adjusted based on the time interval, which is used to strengthen the influence of data points with short time intervals. Represents the rate of change of behavior, which is an indicator used to measure how fast a behavior changes over time. Indicated in Behavioral data at each time point is used to Compare to calculate the change, Indicates Time point to The absolute value of the behavior change at a time point is used to obtain the magnitude of the change, whether positive or negative. is the normalization coefficient used to average the calculation of the behavior change rate to each time interval; Extract key statistical indicators from the weighted behavior change feature table, calculate the average change rate and standard deviation, combine the data to evaluate the change trend of each behavior pattern, and obtain the behavior change feature value.

5. The user follow-up management system for a medical platform based on the Internet of Things according to claim 4, characterized in that: The steps for obtaining the health impact assessment indicators are as follows: Based on the behavior change characteristic value, the data of each behavior pattern is integrated, and atypical behavior records are excluded through data cleaning to obtain cleaned behavior characteristic data; Applying a statistical model to the cleaned behavioral feature data to analyze the correlation between multiple behavioral patterns and patient health outcomes, evaluating the impact intensity of the behavioral patterns through correlation coefficients, and obtaining behavioral health correlation analysis results; Based on the behavioral health association analysis results, the formula is used: ； Calculate the sensitivity of each behavior to health status and obtain a health sensitivity analysis table; in, is the number of behavior patterns, Indicates behavior pattern Impact weights, weighted health impacts of behaviors, It is behavior The direct impact coefficient on health, is a constant used to adjust the stability of the calculation, stands for health sensitivity, which is used to quantify the overall impact of multiple behaviors on health status; Based on the health sensitivity analysis table, the average value and coefficient of variation are used to summarize the impact of each behavior pattern on health and generate a health impact assessment index.

6. The user follow-up management system for a medical platform based on the Internet of Things according to claim 5, characterized in that: The steps for obtaining the future behavior prediction result are specifically as follows: Integrate the health impact assessment indicators with the patient's behavioral data, determine the statistical correlation between behavioral patterns and health outcomes through data cross-validation, and obtain a behavior-health correlation model; Based on the behavior and health association model, a multivariate linear regression analysis is designed and implemented to predict the patient's behavior pattern in the future time period, and the regression coefficient is adjusted through data to obtain an adjusted prediction model; Apply the adjusted prediction model, combined with real-time behavioral data, using the formula: ； Computationally generate behavioral predictions at multiple time points in the future; in, Indicates a future time point The behavioral prediction value of is the baseline forecast value, Represents the influence weight of multiple behavioral variables, is the exponential form of the behavior variable, which increases the sensitivity of the model to changes in behavior. It represents the number of behavioral variables, that is, the total number of multiple behavioral factors referenced in the model; Based on the behavior predictions at multiple future time points, key behavior trends and change predictions are extracted from the model output, the health impact potential of the prediction results is analyzed, and future behavior prediction results are output.

7. The user follow-up management system for a medical platform based on the Internet of Things according to claim 6, characterized in that: The steps for obtaining the task optimization index are specifically as follows: Analyze the future behavior prediction results, determine the high-frequency behavior pattern, and set the priority of the monitoring task according to the pattern to obtain a priority allocation plan; According to the priority allocation scheme, a targeted response scheme is designed, and the potential effects of multiple schemes are verified through actual data simulation to obtain the efficiency evaluation results of multiple response schemes; Using the data from the efficiency evaluation results, the formula is used: ； Calculate and generate comprehensive indicators of execution efficiency and response speed for each task; in, Indicates the task The optimization index of is the efficiency score of task execution, Indicates the response speed score, Represents the adjustment factor of the efficiency score, adjustment The influence of It is a constant offset, which ensures that the denominator is not zero and enhances the stability of the calculation; The comprehensive indicators of execution efficiency and response speed of each task are analyzed, the task flow of the monitoring system is adjusted according to actual operation requirements, and task optimization indicators are integrated and generated.

8. A user follow-up management method for a medical platform based on the Internet of Things, characterized in that: The user follow-up management system for the medical platform based on the Internet of Things according to any one of claims 1 to 7 comprises the following steps: Extract patient behavioral data with timestamps from IoT devices, perform data merging, combine behavioral data with the patient's unique identifier, record the time and amount of data collection, and generate a data consolidation table; According to the data integration table, the behavior data of consecutive time points are extracted, the change rate of each patient's behavior is calculated, and the change trend of the behavior pattern is determined by comparing the differences between adjacent data points to generate a behavior characteristic value result; Using the behavior feature value results, evaluating the impact of the feature value on the patient's health status, performing correlation analysis between behavior patterns and health outcomes, thereby calculating the sensitivity of health changes and generating health impact assessment indicators; Based on the health impact assessment indicators and combined with the patient's behavioral data, a regression analysis model is established to predict patient behavior changes at multiple time points in the future, generate future behavior prediction results, and adjust the priority of monitoring tasks in real time according to the future behavior prediction results, match response plans for each patient behavior, and optimize the task execution process.

Images