A chronic disease patient health management system based on big data

Abstract

The present application relates to the technical field of medical health management, in particular to a chronic disease patient health management system based on big data, comprising: constructing a full-dimension personal health record containing static and dynamic information. The data analysis module performs abnormal pattern mining on continuous sign data, and generates a trend graph by associating periodic test data with historical records. The data correlation module constructs a risk correlation matrix based on environmental exposure data, which quantifies the contribution of environmental behavior factors to health indicators. The risk modeling module integrates the above information. The prediction and intervention generation module uses a chronic disease progression prediction model to calculate and synchronously output individualized disease risk warning signals and the corresponding phased health intervention programs. The present application realizes the quantitative attribution of health risk factors and dynamic, phased precision health management.

Description

Technical Field

[0001] This invention relates to the field of medical and health management technology, and in particular to a health management system for chronic disease patients based on big data. Background Technology

[0002] Most existing health management systems for chronic disease patients are based on the integration of electronic health records and wearable device data access, providing data visualization and simple early warning functions based on fixed thresholds. These solutions mainly achieve centralized storage and trend display of data such as vital signs and test results, triggering alerts when monitored data exceeds preset general safety ranges. The core flaw of such solutions lies in the fact that their analysis and early warning mechanisms remain at the level of comparing superficial data, failing to deeply explain the potential causes of fluctuations in vital signs or changes in indicators. The alerts generated by the system often only indicate "an abnormal indicator" but cannot answer "why the abnormality occurred at this time." In addition, the health recommendations provided by the system are usually static and general knowledge base entries, failing to be closely integrated with the patient's current specific and dynamically changing risk status, resulting in insufficient personalization and timeliness of interventions.

[0003] Current technologies lack the ability to systematically quantify and attribute the diverse external factors influencing chronic diseases. A patient's health is affected by a complex interplay of genetic, behavioral, and environmental factors. While conventional systems may record environmental or behavioral data, they lack quantitative models establishing the correlation between these external factors and changes in internal physiological indicators. This prevents health management from accurately identifying and focusing on modifiable key risk drivers. Furthermore, due to the lack of predictive judgment regarding the dynamic trajectory of disease progression, systematic intervention recommendations are often reactive and one-off, failing to develop a progressive, dynamic management plan that matches the evolving stages of disease risk, thus hindering truly effective long-term health outcome improvements. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a health management system for chronic disease patients based on big data.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a health management system for chronic disease patients based on big data, comprising:

[0006] The file construction module constructs a comprehensive personal health file for patients with chronic diseases. The comprehensive personal health file includes a static basic information set and a dynamic monitoring information set. The static basic information set includes records of genetic disease history, previous diagnosis records, and long-term medication records. The dynamic monitoring information set includes continuous vital sign monitoring data, periodic test data, and environmental exposure data.

[0007] The data analysis module performs time-series anomaly pattern mining on the continuous vital sign monitoring data, identifies segments of vital sign fluctuations that deviate from the individual's baseline pattern, and performs correlation mapping analysis between the periodic test data and the previous diagnostic records to generate trend evolution maps of test indicators.

[0008] The data association module constructs a risk association matrix between patient behavior and the external environment based on the environmental exposure data. The risk association matrix quantifies the contribution of different environmental factors and behavioral patterns to fluctuations in various vital signs and changes in laboratory indicators.

[0009] The risk modeling module integrates the vital sign fluctuation segments, the trend evolution map of the test indicators, and the risk correlation matrix;

[0010] The prediction and intervention generation module uses a pre-set chronic disease progression prediction model to calculate and generate individualized disease risk warning signals and phased health intervention plans.

[0011] As a further aspect of the present invention, the step of performing time-series abnormal pattern mining on the continuous vital sign monitoring data to identify vital sign fluctuation segments that deviate from the individual baseline pattern includes:

[0012] The segments of vital signs fluctuation include abnormal heart rate variability intervals, periods of blood pressure diurnal rhythm disorder, and periodic decreases in blood oxygen saturation.

[0013] A historical data time window is set for calculating the individual baseline pattern, and historical monitoring sequences within the historical data time window are extracted from the continuous vital signs monitoring data;

[0014] The mean, standard deviation, and periodic component of different vital signs in the historical monitoring sequence are calculated respectively to determine the normal fluctuation range and rhythm characteristics of personalized vital signs.

[0015] The newly acquired real-time monitoring sequence is compared point by point with the normal fluctuation range and rhythm characteristics of the vital signs, and data points that exceed the fluctuation range or violate the rhythm characteristics are marked.

[0016] Cluster analysis is performed on the marked abnormal data points to aggregate the marked points that are continuous or close in time and have the same abnormal type into the vital sign fluctuation segments.

[0017] Each of the aforementioned vital sign fluctuation segments is assigned a type label and a severity score, the type label including abnormal heart rate variability, circadian rhythm disorder of blood pressure, or periodic decrease in blood oxygen saturation.

[0018] As a further aspect of the present invention, the step of comparing the newly acquired real-time monitoring sequence with the normal fluctuation range and rhythmic characteristics of the vital signs point by point, and marking data points that exceed the fluctuation range or violate the rhythmic characteristics, includes:

[0019] Extract the time series of individual vital signs from the real-time monitoring sequence and compare them with the normal fluctuation range of the corresponding vital signs. If multiple consecutive data points exceed the upper or lower limit of the normal fluctuation range, they are marked as continuous abnormalities.

[0020] The changes in the time series of the individual characteristic index are analyzed over a 24-hour period. The similarity between the index and the typical diurnal variation curve defined in the rhythm characteristics is calculated. If the similarity is lower than a preset threshold, it is marked as a rhythm abnormality.

[0021] The system detects whether there are segments in the time series of the individual characteristic index that rise or fall sharply in a short period of time. It calculates the difference between adjacent data points and compares it with the historical average rate of change. If the difference exceeds a certain multiple of the historical average rate of change, it is marked as a sudden change anomaly.

[0022] Data points that meet any of the following conditions—persistent abnormality, rhythm abnormality, or mutation abnormality—are recorded as abnormal data points, and the specific abnormality type is recorded.

[0023] As a further aspect of the present invention, the step of performing correlation mapping analysis between the periodic test data and the previous diagnostic records to generate a trend evolution map of the test indicators includes:

[0024] The trend evolution graph of the test indicators reflects the trajectory of the value of a specific test item over time and its correspondence with historical diagnostic events.

[0025] Extract multiple test results of the same test item from the periodic test data, and sort them by test time to form a numerical sequence of the specific item;

[0026] Extract the time point of the diagnostic event and the diagnostic conclusion from the previous diagnostic records, wherein the diagnostic conclusion includes the disease stage or severity level;

[0027] Align the numerical sequence of the specific item with the time point of the diagnostic event on the time axis to establish the time correspondence between the test value and the diagnostic event.

[0028] The analysis examines the trends, rates of change, and peak or trough values ​​of the numerical sequence of the specific item before and after each diagnostic event.

[0029] Based on the time correspondence and change characteristics, a graphical atlas is generated with time as the horizontal axis, test item values ​​as the vertical axis, and diagnostic event points marked, namely the trend evolution atlas of the test indicators.

[0030] As a further aspect of the present invention, the analysis of the changing trend, rate of change, and peak or trough values ​​of the numerical sequence of the specific item before and after each diagnostic event includes:

[0031] Using the time of the diagnostic event as the dividing point, time windows of predetermined lengths are selected before and after the event, and the average value of the numerical sequence of the specific items within the window is calculated as the baseline value before diagnosis and the baseline value after diagnosis.

[0032] The difference between the baseline value after diagnosis and the baseline value before diagnosis is calculated as the net change in the impact of the diagnostic event on the test indicators.

[0033] Within the pre-diagnosis time window, the linear regression slope of the numerical series is calculated as the pre-event trend slope; within the post-diagnosis time window, the linear regression slope of the numerical series is calculated as the post-event trend slope.

[0034] Identify the global maximum and global minimum values ​​of the entire numerical sequence over the entire time range, and record them as peak values ​​and valley values, respectively;

[0035] The average daily change from the pre-diagnosis baseline value to the post-diagnosis baseline value was recorded as the average daily rate of change.

[0036] As a further aspect of the present invention, the step of constructing a risk correlation matrix between patient behavior and the external environment based on the environmental exposure data includes:

[0037] Multiple environmental exposure factors and patient behavioral activity records were extracted from the environmental exposure data. The environmental exposure factors included air quality index, temperature and humidity levels, and noise decibel levels. The patient behavioral activity records included physical activity intensity, sleep duration, and dietary content.

[0038] For each health outcome associated with a chronic disease, a set of relevant physical and laboratory indicators are defined as target variables for risk association.

[0039] Multivariate correlation analysis and regression analysis were used to calculate the statistical association strength coefficient between each environmental exposure factor and each patient behavioral activity record on each target variable.

[0040] All the calculated statistical association strength coefficients are organized into a multi-dimensional numerical table, namely the risk association matrix, according to the environmental exposure factor, the behavioral activity record rows, and the target variable as the column.

[0041] The coefficients in the risk correlation matrix are standardized to make the coefficients of different target variables comparable.

[0042] As a further aspect of the present invention, the method of employing multivariate correlation analysis and regression analysis to calculate the statistical association strength coefficient between each environmental exposure factor and each patient behavioral activity record on each target variable includes:

[0043] For each specific pair of environmental exposure factors and target variables, historical data pairs of environmental exposure factors and target variables at the same time point are collected.

[0044] Calculate the correlation coefficient between the historical data series of the environmental exposure factors and the historical data series of the target variable;

[0045] A linear regression model was established with the environmental exposure factors as independent variables and the target variable as the dependent variable, and the regression coefficients were obtained by fitting the model.

[0046] The correlation coefficient and the regression coefficient are weighted together to obtain the statistical association strength coefficient of the environmental exposure factor on the target variable;

[0047] For each combination of patient behavior activity records and target variables, the steps from data collection to comprehensive weighting are repeated.

[0048] As a further aspect of the present invention, the integration of the vital sign fluctuation segments, the trend evolution map of the test indicators, and the risk correlation matrix, and the calculation using a preset chronic disease progression prediction model, generates a personalized disease risk early warning signal, including:

[0049] All the aforementioned vital sign fluctuation segments identified during the current monitoring period are scored according to their type and severity, and converted into input feature vectors for the prediction model.

[0050] From the latest trend evolution map of the test indicators, the recent trend slope and the offset of the current value relative to the historical baseline are extracted as another set of input features;

[0051] Based on the patient's recently recorded environmental exposure data and behavioral activity records, combined with the risk correlation matrix, a comprehensive risk exposure score under the current combination of environment and behavior is calculated as the third set of input features;

[0052] The input feature vector, recent trend slope and offset, and comprehensive risk exposure score are input together into the chronic disease progression prediction model;

[0053] The chronic disease progression prediction model outputs the probability value of a specific chronic disease acute event or disease stage progression occurring within a specified future time period. When the probability value exceeds a preset warning threshold, a corresponding disease risk warning signal is generated. The signal includes the risk event type and the expected occurrence time window.

[0054] As a further aspect of the present invention, generating the phased health intervention plan includes:

[0055] The health intervention program includes medication adjustment recommendations, a list of behavioral pattern optimizations, and a follow-up visit plan;

[0056] Based on the risk event type indicated in the disease risk warning signal, a set of corresponding intervention measures is matched from a pre-set intervention knowledge base;

[0057] Based on the patient's current long-term medication records, a compatibility check is performed on the medication recommendations in the intervention set to screen out medication adjustment recommendations without conflicts;

[0058] Based on the patient's personal daily routine preferences and the recorded behavioral activities, specific execution times, frequencies, and intensities are set for the behavioral intervention recommendations in the intervention set, forming a list of behavioral pattern optimizations;

[0059] Based on the monitoring needs of key indicators indicated in the trend evolution graph of the test indicators, and the urgency of the disease risk warning signals, a follow-up plan including the next test items and the recommended follow-up time is formulated.

[0060] The medication adjustment recommendations, behavioral pattern optimization list, and follow-up visit plan are integrated into a phased health intervention plan document with a clear implementation cycle.

[0061] As a further aspect of the present invention, the construction steps of the chronic disease progression prediction model include:

[0062] Collect a complete set of health monitoring data for a historical chronic disease patient population. The set of health monitoring data includes continuous vital sign monitoring data, periodic test data, environmental exposure data, and corresponding disease progression event records.

[0063] The continuous vital signs monitoring data are processed by time series segmentation to extract the feature vector of vital signs fluctuation patterns for each patient at different time periods;

[0064] The periodic test data are processed to extract trend features and generate a descriptor for the trajectory of test index changes for each patient.

[0065] The environmental exposure data and patient behavioral activity records are subjected to feature encoding processing to obtain an environmental behavior feature matrix;

[0066] The feature vector of the vital sign fluctuation pattern, the descriptor of the trajectory of the test index change, and the environmental behavior feature matrix are subjected to multi-source feature fusion processing to generate a comprehensive feature representation.

[0067] Each patient sample is labeled with a disease progression status tag based on the disease progression event records;

[0068] The time-series deep learning network is trained using the comprehensive feature representation and disease progression status labels. The network parameters are then optimized using the backpropagation algorithm to obtain the trained chronic disease progression prediction model.

[0069] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0070] By using environmental exposure data to structurally link patients' behaviors with their external environment, a risk correlation matrix is ​​constructed. Statistical analysis or machine learning algorithms are then used to calculate the specific contribution of different environment-behavior combinations to various health indicators. This approach transforms the previously vague environmental and behavioral risks into quantifiable, rankable, and comparable numerical indicators. This clearly reveals the key external drivers influencing a specific patient's health fluctuations and their relative importance, enabling health management to move beyond simply focusing on "whether indicators are normal" to understanding "why indicators change." This provides direct quantitative decision-making support for targeted and precise environmental adjustments and behavioral interventions, surpassing traditional models based on experience or qualitative recommendations.

[0071] By integrating multi-source fusion data through a pre-defined chronic disease progression prediction model, a dynamic health intervention plan is generated that is directly linked to individualized disease risk warning signals and implemented in stages. This plan outputs not a static list of recommendations, but a dynamic action plan sequence containing different implementation time points and intervention focuses. This allows health interventions to proactively match the predicted pace of risk evolution with the patient's current specific health stage. Intervention measures can be scientifically arranged in sequence based on risk priority and feasibility, forming a complete management path combining immediate short-term measures to ongoing long-term plans, thus transforming a one-off reminder into an executable and adjustable continuous health management process. Attached Figure Description

[0072] Figure 1 This is a sequence diagram of the big data-based chronic disease patient health management system described in this invention;

[0073] Figure 2 A flowchart for generating the trend evolution map of the indicator. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0075] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0076] See Figure 1 The patient profile construction module is responsible for building a comprehensive personal health profile for patients with chronic diseases. This profile includes a static basic information set and a dynamic monitoring information set. The static basic information set includes records of genetic disease history, previous diagnosis records, and long-term medication records. The dynamic monitoring information set includes continuous vital sign monitoring data, periodic test data, and environmental exposure data. The data analysis module performs time-series anomaly pattern mining on the continuous vital sign monitoring data to identify segments of vital sign fluctuations that deviate from the individual's baseline pattern. The data analysis module also performs correlation mapping analysis between periodic test data and previous diagnosis records to generate trend evolution maps of test indicators. The data correlation module constructs a risk correlation matrix between patient behavior and the external environment based on environmental exposure data. This matrix quantifies the contribution of different environmental factors and behavioral patterns to fluctuations in various vital signs and changes in test indicators. The risk modeling module integrates vital sign fluctuation segments, trend evolution maps of test indicators, and the risk correlation matrix. The prediction and intervention generation module uses a pre-set chronic disease progression prediction model to calculate and generate individualized disease risk warning signals and phased health intervention plans.

[0077] See Figure 2 In one embodiment of the present invention, time-series anomaly pattern mining is performed on continuous vital sign monitoring data to identify vital sign fluctuation segments that deviate from the individual's baseline pattern. These fluctuation segments include abnormal heart rate variability intervals, periods of disrupted diurnal blood pressure rhythms, and periodic decreases in blood oxygen saturation. A historical data time window is set for calculating the individual's baseline pattern, for example, the past thirty days. Historical monitoring sequences within this time window are extracted from the continuous vital sign monitoring data. These sequences contain hourly records of heart rate, systolic blood pressure, diastolic blood pressure, and blood oxygen saturation. The mean, standard deviation, and periodic component of different vital sign indicators in the historical monitoring sequences are calculated to determine the personalized normal fluctuation range and rhythmic characteristics of the vital sign indicators. The personalized normal fluctuation range is defined by the mean plus or minus twice the standard deviation, and the rhythmic characteristics are extracted using Fourier analysis to extract typical waveforms within a 24-hour period. In some embodiments, the length of the historical data time window can be dynamically adjusted according to the data acquisition frequency.

[0078] In practice, newly acquired real-time monitoring sequences are compared point-by-point with the normal fluctuation range and rhythm characteristics of vital signs, marking data points that exceed the fluctuation range or violate rhythm characteristics. Individual vital sign time series are extracted from the real-time monitoring sequences, such as a daily heart rate time series, and compared numerically with the corresponding normal fluctuation range of the heart rate indicator. If three consecutive data points exceed the upper or lower limit of the normal fluctuation range, they are marked as persistent abnormalities. The morphological changes of individual vital sign time series over 24 hours are analyzed, and their morphological similarity is calculated with the typical diurnal variation curve defined in the rhythm characteristics. The morphological similarity is calculated using a dynamic time warping algorithm. If the calculated dynamic time warping distance is higher than a preset threshold, it is marked as a rhythm abnormality. The presence of segments with sharp increases or decreases in the short term in the individual vital sign time series is detected by calculating the difference between adjacent data points and comparing it with the historical average rate of change. If the difference exceeds three times the historical average rate of change, it is marked as a sudden change abnormality. It is understood that the marking rules are configurable. Data points that meet any of the conditions of persistent abnormality, rhythm abnormality, or sudden change abnormality are recorded as abnormal data points, and the specific abnormality type is recorded.

[0079] In practice, cluster analysis is performed on the marked abnormal data points, aggregating those that are temporally continuous or close and have the same abnormality type into vital sign fluctuation segments. This aggregation process is based on the timestamps and type labels of the abnormal data points, using a density-based clustering algorithm. Each vital sign fluctuation segment is assigned a type label and a severity score. Type labels include abnormal heart rate variability, circadian rhythm disturbances in blood pressure, or periodic decreases in blood oxygen saturation. The severity score is calculated using a formula:

[0080]

[0081] in: Severity rating This represents the number of abnormal data points contained within a segment of vital sign fluctuations. Representing the The actual monitored value of each abnormal data point. This represents the expected value of that vital sign in the individual's baseline pattern at the corresponding time. This represents the standard deviation of that vital sign indicator in the individual's baseline pattern at the corresponding time. This represents the weighting coefficient preset based on the anomaly type.

[0082] In one embodiment of the invention, periodic test data is correlated and mapped with previous diagnostic records to generate a trend evolution map of test indicators. This map reflects the trajectory of specific test item values ​​over time and their correspondence with historical diagnostic events. Multiple test results for the same test item are extracted from the periodic test data, such as the glycated hemoglobin (HbA1c) value of a diabetic patient. These results are sorted by test time to form a numerical sequence for the specific item, containing multiple timestamps and corresponding test values. The time points of diagnostic events and diagnostic conclusions are extracted from previous diagnostic records. Diagnostic conclusions include disease stage or severity level, such as a diagnosis of "diabetic nephropathy stage III". The numerical sequence of the specific item is aligned with the time points of the diagnostic events on the timeline to establish a temporal correspondence between test values ​​and diagnostic events, ensuring that each HbA1c test value can be associated with its most recent preceding or following diagnostic event on the timeline.

[0083] In some embodiments, the analysis examines the trend, rate of change, and peak or trough values ​​of the numerical sequences of specific items before and after each diagnostic event. Using the diagnostic event time as a dividing point, predetermined time windows of varying lengths are selected before and after the event, for example, six months before and six months after the diagnosis. The average value of the numerical sequence of the specific item within each window is calculated as the pre-diagnosis baseline value and the post-diagnosis baseline value. The difference between the post-diagnosis baseline value and the pre-diagnosis baseline value is calculated as the net change in the impact of the diagnostic event on the test indicator. Within the pre-diagnosis time window, the linear regression slope of the numerical sequence is calculated as the pre-event trend slope; within the post-diagnosis time window, the linear regression slope of the numerical sequence is calculated as the post-event trend slope. It is understood that the length of the time window can be adjusted according to the disease type and the clinical significance of the test item. The global maximum and global minimum values ​​of the entire numerical sequence over the entire time range are identified and recorded as peak and trough values, respectively. The average daily change from the pre-diagnosis baseline value to the post-diagnosis baseline value is recorded as the average daily rate of change, calculated using the following formula:

[0084]

[0085] in: Represents the average daily rate of change. Represents the baseline value before diagnosis. Represents the baseline value after diagnosis. This represents the actual number of days elapsed from the time point corresponding to the previous baseline value to the time point corresponding to the subsequent baseline value. Optionally, a moving average of all values ​​within the time window can be used when calculating the baseline values ​​before and after diagnosis.

[0086] In practical implementation, based on the time correspondence and change characteristics, a graphical atlas is generated with time as the horizontal axis, test item values ​​as the vertical axis, and diagnostic event points labeled, i.e., a test indicator trend evolution atlas. It can be understood that the graphical atlas will plot all historical test values ​​in scatter plot form, display the trend of the value sequence with connecting lines, and clearly mark the occurrence time and conclusion of each diagnostic event using vertical lines and labels at corresponding positions on the time axis. In some embodiments, the atlas can overlay the calculated pre-event trend lines and post-event trend lines. Optionally, the atlas can also label the calculated peak values, trough values, and the positions of the pre-diagnosis baseline value and the post-diagnosis baseline value.

[0087] In one embodiment of the invention, a risk association matrix between patient behavior and the external environment is constructed based on environmental exposure data. Various environmental exposure factors and patient behavior records are extracted from the environmental exposure data. Environmental exposure factors include air quality index, temperature and humidity levels, and noise levels (decibels). Patient behavior records include physical activity intensity, sleep duration, and dietary content. For each chronic disease-related health outcome, such as acute cardiovascular events in hypertension, a set of relevant vital signs and laboratory indicators are defined as target variables for risk association. Target variables may include systolic blood pressure, diastolic blood pressure, and C-reactive protein concentration in the blood. In some embodiments, the selection of target variables follows clinical guidelines.

[0088] In practice, multivariate correlation and regression analyses are used to calculate the statistical association strength coefficient between each environmental exposure factor and each patient behavioral activity record for each target variable. For each specific pair of environmental exposure factors and target variables, such as outdoor PM2.5 concentration and systolic blood pressure, historical data pairs of the environmental exposure factor and target variable at the same time point are collected, accurate to the day. The correlation coefficient between the historical data series of the environmental exposure factor and the historical data series of the target variable is calculated using the Pearson correlation coefficient. A linear regression model is established with the environmental exposure factor as the independent variable and the target variable as the dependent variable, and the regression coefficients are obtained through fitting. The regression coefficients reflect the impact of a unit change in the environmental exposure factor on the expected change in the target variable. The correlation coefficients and regression coefficients are then weighted to obtain the statistical association strength coefficient between the environmental exposure factor and the target variable.

[0089] In some embodiments, all calculated statistical association strength coefficients are organized into a multi-dimensional numerical table, namely a risk association matrix, with rows for environmental exposure factors and rows for behavioral activity records, and columns for target variables. The row headings of the risk association matrix are the specific names of the environmental exposure factors and patient behavioral activity records, the column headings are the names of the target variables, and each cell in the matrix contains the corresponding statistical association strength coefficient. The coefficients in the risk correlation matrix are standardized to make them comparable across different target variables. The standardization method used is Z-score standardization, which involves subtracting the mean of all coefficients in each column (corresponding to a target variable) and then dividing by its standard deviation. It can be understood that the standardized coefficients directly reflect the relative strength of the influence of different environmental or behavioral factors under the same target variable dimension.

[0090] In one embodiment of the present invention, vital sign fluctuation segments, laboratory indicator trend evolution maps, and risk correlation matrices are integrated, and a pre-defined chronic disease progression prediction model is used to calculate and generate a personalized disease risk warning signal. All vital sign fluctuation segments identified within the current monitoring period are scored according to their type and severity, and converted into input feature vectors for the prediction model. The input feature vector is a multi-dimensional numerical array, where each dimension corresponds to a type of vital sign fluctuation segment, and its value is aggregated from the severity scores of that type of segment. The aggregation method can be summation or taking the maximum value. From the latest laboratory indicator trend evolution map, the recent trend slope and the offset of the current value relative to the historical baseline are extracted as another set of input features. The recent trend slope is usually taken from the post-event trend slope, and the historical baseline value can be the post-diagnosis baseline value or the average value of a longer historical window. Based on the patient's recently recorded environmental exposure data and behavioral activity records, combined with the risk correlation matrix, a comprehensive risk exposure score under the current environmental and behavioral combination is calculated as a third set of input features. It is understandable that calculating the comprehensive risk exposure score requires traversing the rows in the risk association matrix corresponding to the patient's current environment and behavioral records, and then weighted summing their coefficients with respect to a specific target variable. The input feature vector, recent trend slope and offset, along with the comprehensive risk exposure score, are input into the chronic disease progression prediction model. The chronic disease progression prediction model outputs the probability value of a specific acute event or disease stage progression occurring within a specified future time period. When the probability value exceeds a preset warning threshold, a corresponding disease risk warning signal is generated, which includes the risk event type and the expected occurrence time window. See Table 1 for a specific example of converting a segment of vital sign fluctuations into an input feature vector.

[0091] Table 1: Examples of converting vital sign fluctuation segments into input feature vectors

[0092] Vital Sign Fluctuation Fragment Type Label Severity rating Corresponding input feature vector dimension Aggregated eigenvalues Heart rate variability 8.5 Dimension 1 8.5 Heart rate variability 6.2 Dimension 1 8.5 (maximum value) Blood pressure circadian rhythm disorder 12.1 Dimension 2 12.1 Blood oxygen saturation decreases periodically 3.0 Dimension 3 3.0

[0093] In some embodiments, the comprehensive risk exposure score can be calculated using the following formula:

[0094]

[0095] in: Representative regarding the first The comprehensive risk exposure score of the target variables, This represents the total number of entries in the risk association matrix that are related to environmental exposure factors and behavioral activity records. It is an indicator variable, when the first A value of 1 is assigned when an environmental or behavioral factor is observed in recent records, and 0 otherwise. Represents the risk correlation matrix of the first The factor affects the first The standardized statistical association strength coefficients of the target variables. Optional, indicator variables. You can replace it with the specific value or level of that factor from recent records.

[0096] The steps for constructing a chronic disease progression prediction model include: collecting a complete health monitoring dataset of a historical chronic disease patient population, including continuous vital sign monitoring data, periodic laboratory test data, environmental exposure data, and corresponding disease progression event records. The continuous vital sign monitoring data is segmented over time to extract feature vectors of vital sign fluctuation patterns for each patient at different time periods. These feature vectors may include statistics such as fluctuation frequency, average duration, and average severity. The periodic laboratory test data undergoes trend feature extraction to generate descriptors of the trajectory changes in laboratory test indicators for each patient. These descriptors may include long-term slope, periodic amplitude, and coefficient of variation. Environmental exposure data and patient behavioral activity records are combined using feature encoding to obtain an environmental behavior feature matrix. Feature encoding can employ one-hot encoding or discretization binning based on clinical knowledge. Finally, the vital sign fluctuation pattern feature vectors, laboratory test indicator trajectory descriptors, and environmental behavior feature matrix are fused using multi-source feature fusion to generate a comprehensive feature representation. This multi-source feature fusion can be a simple vector concatenation or a weighted fusion using an attention mechanism. Each patient sample is labeled with a disease progression status tag based on disease progression event records. The tag can be a binary "progression status" or a multi-class progression type. A temporal deep learning network is trained using the comprehensive feature representation and disease progression status tags. The network parameters are optimized using a backpropagation algorithm to obtain a trained chronic disease progression prediction model. The temporal deep learning network can be a long short-term memory network or a temporal convolutional network. In some embodiments, the training process divides the model into training, validation, and test sets. Optionally, the network structure and parameters are saved after model training for use by the prediction and intervention generation module.

[0097] In one embodiment of the invention, a phased health intervention plan is generated, which includes medication adjustment recommendations, a list of optimized behavior patterns, and a follow-up visit plan. Based on the risk event type indicated in the disease risk warning signal, such as an increased risk of acute exacerbation of heart failure within the next month, a corresponding set of intervention measures is matched from a pre-set intervention knowledge base. The intervention knowledge base stores standardized intervention entries corresponding to different risk event types. The matched set of intervention measures may include "increasing the diuretic dosage," "restricting daily sodium intake," and "recommending an echocardiogram within one week." Combining the patient's current long-term medication record, a compatibility check is performed on the medication recommendations in the intervention set. Medication adjustment recommendations without conflicts are screened out. The compatibility check verifies whether there are any known serious interactions between the drugs in the long-term medication record and the drugs to be added or adjusted. For example, if a patient has been taking warfarin long-term, and the intervention recommends adding aspirin, and the check reveals an interaction that increases the risk of bleeding, then this medication recommendation will be marked as conflicting and screened out. It is understood that the compatibility check relies on a continuously updated drug interaction database.

[0098] In some embodiments, based on the patient's personal lifestyle preferences and activity records, specific execution times, frequencies, and intensities are set for behavioral intervention recommendations in the intervention set, forming a behavioral pattern optimization list. For example, for a behavioral intervention recommendation of "performing moderate-intensity aerobic exercise," the system reads the patient's personal lifestyle preferences, finds that the patient usually has free time after 7 pm on weekdays, and combines this with the activity history showing that the patient can tolerate 30 minutes of walking each time. Therefore, this recommendation is specified as "performing 30 minutes of brisk walking exercise every Monday, Wednesday, and Friday evening at 7:30 pm," and this specific item is added to the behavioral pattern optimization list. Optionally, if the activity record shows that the patient has recently experienced knee discomfort, the system may automatically adjust the exercise intensity to "low intensity." Based on the monitoring needs of key indicators indicated in the trend evolution graph of test indicators and the urgency of disease risk warning signals, a follow-up plan including the next test items and recommended follow-up appointment time is formulated. For example, if the trend chart of test indicators shows that the level of B-type natriuretic peptide (BNP) has been rising recently, and the probability value of the disease risk warning signal is high, then the follow-up visit plan will clearly list "BNP test" as the next test item, and use a formula to calculate the recommended follow-up visit time:

[0099]

[0100] in: The representative suggested a follow-up appointment time. Represents the current time. This represents the starting point of the expected occurrence time window in the disease risk warning signal. It is a safety factor determined based on the risk level. This indicates rounding up. This calculation is understood to ensure that follow-up assessments are completed before the expected occurrence window of the risk event. Medication adjustment recommendations, a list of optimized behavioral patterns, and the follow-up plan are integrated into a phased health intervention plan document with a clearly defined execution cycle. The document specifies the execution cycle for this phase of intervention. In some embodiments, the integrated document is presented to patients and healthcare professionals in a structured list and calendar view. Optionally, at the end of the execution cycle, the system automatically triggers a new round of health status assessment and plan generation.

[0101] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A health management system for chronic disease patients based on big data, characterized in that, The system includes: The file construction module constructs a comprehensive personal health file for patients with chronic diseases. The comprehensive personal health file includes a static basic information set and a dynamic monitoring information set. The static basic information set includes records of genetic disease history, previous diagnosis records, and long-term medication records. The dynamic monitoring information set includes continuous vital sign monitoring data, periodic test data, and environmental exposure data. The data analysis module performs time-series anomaly pattern mining on the continuous vital sign monitoring data, identifies segments of vital sign fluctuations that deviate from the individual's baseline pattern, and performs correlation mapping analysis between the periodic test data and the previous diagnostic records to generate trend evolution maps of test indicators. The data association module constructs a risk association matrix between patient behavior and the external environment based on the environmental exposure data. The risk association matrix quantifies the contribution of different environmental factors and behavioral patterns to fluctuations in various vital signs and changes in laboratory indicators. The risk modeling module integrates the vital sign fluctuation segments, the trend evolution map of the test indicators, and the risk correlation matrix; The prediction and intervention generation module uses a preset chronic disease progression prediction model to calculate and generate individualized disease risk warning signals and phased health intervention plans. The process integrates the vital sign fluctuation segments, the trend evolution map of the test indicators, and the risk correlation matrix, and uses a preset chronic disease progression prediction model to calculate and generate personalized disease risk warning signals, including: All the aforementioned vital sign fluctuation segments identified during the current monitoring period are scored according to their type and severity, and converted into input feature vectors for the prediction model. From the latest trend evolution map of the test indicators, the recent trend slope and the offset of the current value relative to the historical baseline are extracted as another set of input features; Based on the patient's recently recorded environmental exposure data and behavioral activity records, combined with the risk correlation matrix, a comprehensive risk exposure score under the current combination of environment and behavior is calculated as the third set of input features; The input feature vector, recent trend slope and offset, and comprehensive risk exposure score are input together into the chronic disease progression prediction model; The chronic disease progression prediction model outputs the probability value of a specific chronic disease acute event or disease stage progression occurring within a specified time period in the future. When the probability value exceeds a preset warning threshold, a corresponding disease risk warning signal is generated. The signal includes the risk event type and the expected occurrence time window.

2. The chronic disease patient health management system based on big data according to claim 1, characterized in that, The step of performing time-series anomaly pattern mining on the continuous vital sign monitoring data to identify segments of vital sign fluctuations that deviate from the individual baseline pattern includes: The segments of vital signs fluctuation include abnormal heart rate variability intervals, periods of blood pressure diurnal rhythm disorder, and periodic decreases in blood oxygen saturation. A historical data time window is set for calculating the individual baseline pattern, and historical monitoring sequences within the historical data time window are extracted from the continuous vital signs monitoring data; The mean, standard deviation, and periodic component of different vital signs in the historical monitoring sequence are calculated respectively to determine the normal fluctuation range and rhythm characteristics of personalized vital signs. The newly acquired real-time monitoring sequence is compared point by point with the normal fluctuation range and rhythm characteristics of the vital signs, and data points that exceed the fluctuation range or violate the rhythm characteristics are marked. Cluster analysis is performed on the marked abnormal data points to aggregate the marked points that are continuous or close in time and have the same abnormal type into the vital sign fluctuation segments. Each of the aforementioned vital sign fluctuation segments is assigned a type label and a severity score, the type label including abnormal heart rate variability, circadian rhythm disorder of blood pressure, or periodic decrease in blood oxygen saturation.

3. The chronic disease patient health management system based on big data according to claim 2, characterized in that, The step of comparing the newly acquired real-time monitoring sequence with the normal fluctuation range and rhythmic characteristics of the vital signs point by point, and marking data points that exceed the fluctuation range or violate the rhythmic characteristics, includes: Extract the time series of individual vital signs from the real-time monitoring sequence and compare them with the normal fluctuation range of the corresponding vital signs. If multiple consecutive data points exceed the upper or lower limit of the normal fluctuation range, they are marked as continuous abnormalities. The changes in the time series of the individual characteristic index are analyzed over a 24-hour period. The similarity between the index and the typical diurnal variation curve defined in the rhythm characteristics is calculated. If the similarity is lower than a preset threshold, it is marked as a rhythm abnormality. The system detects whether there are segments in the time series of the individual characteristic index that rise or fall sharply in a short period of time. It calculates the difference between adjacent data points and compares it with the historical average rate of change. If the difference exceeds a certain multiple of the historical average rate of change, it is marked as a sudden change anomaly. Data points that meet any of the following conditions—persistent abnormality, rhythm abnormality, or mutation abnormality—are recorded as abnormal data points, and the specific abnormality type is recorded.

4. The chronic disease patient health management system based on big data according to claim 1, characterized in that, The step of performing correlation mapping analysis between the periodic test data and the previous diagnostic records to generate a trend evolution map of the test indicators includes: The trend evolution graph of the test indicators reflects the trajectory of the value of a specific test item over time and its correspondence with historical diagnostic events. Extract multiple test results of the same test item from the periodic test data, and sort them by test time to form a numerical sequence of the specific item; Extract the time point of the diagnostic event and the diagnostic conclusion from the previous diagnostic records, wherein the diagnostic conclusion includes the disease stage or severity level; Align the numerical sequence of the specific item with the time point of the diagnostic event on the time axis to establish the time correspondence between the test value and the diagnostic event. The analysis examines the trends, rates of change, and peak or trough values ​​of the numerical sequence of the specific item before and after each diagnostic event. Based on the time correspondence and change characteristics, a graphical atlas is generated with time as the horizontal axis, test item values ​​as the vertical axis, and diagnostic event points marked, namely the trend evolution atlas of the test indicators.

5. A health management system for chronic disease patients based on big data according to claim 4, characterized in that, The analysis examines the trends, rates of change, and peak or trough values ​​of the numerical sequences of the specific item before and after each diagnostic event, including: Using the time of the diagnostic event as the dividing point, time windows of predetermined lengths are selected before and after the event, and the average value of the numerical sequence of the specific items within the window is calculated as the baseline value before diagnosis and the baseline value after diagnosis. The difference between the baseline value after diagnosis and the baseline value before diagnosis is calculated as the net change in the impact of the diagnostic event on the test indicators. Within the pre-diagnosis time window, the linear regression slope of the numerical series is calculated as the pre-event trend slope; within the post-diagnosis time window, the linear regression slope of the numerical series is calculated as the post-event trend slope. Identify the global maximum and global minimum values ​​of the entire numerical sequence over the entire time range, and record them as peak values ​​and valley values, respectively; The average daily change from the pre-diagnosis baseline value to the post-diagnosis baseline value was recorded as the average daily rate of change.

6. A health management system for chronic disease patients based on big data according to claim 1, characterized in that, The construction of a risk correlation matrix between patient behavior and the external environment based on the environmental exposure data includes: Multiple environmental exposure factors and patient behavioral activity records were extracted from the environmental exposure data. The environmental exposure factors included air quality index, temperature and humidity levels, and noise decibel levels. The patient behavioral activity records included physical activity intensity, sleep duration, and dietary content. For each health outcome associated with a chronic disease, a set of relevant physical and laboratory indicators are defined as target variables for risk association. Multivariate correlation analysis and regression analysis were used to calculate the statistical association strength coefficient between each environmental exposure factor and each patient behavioral activity record on each target variable. All the calculated statistical association strength coefficients are organized into a multi-dimensional numerical table, namely the risk association matrix, according to the environmental exposure factor, the behavioral activity record rows, and the target variable as the column. The coefficients in the risk correlation matrix are standardized to make the coefficients of different target variables comparable.

7. A health management system for chronic disease patients based on big data according to claim 6, characterized in that, The method employs multivariate correlation analysis and regression analysis to calculate the statistical association strength coefficient between each environmental exposure factor and each patient behavioral activity record on each target variable, including: For each specific pair of environmental exposure factors and target variables, historical data pairs of environmental exposure factors and target variables at the same time point are collected. Calculate the correlation coefficient between the historical data series of the environmental exposure factors and the historical data series of the target variable; A linear regression model was established with the environmental exposure factors as independent variables and the target variable as the dependent variable, and the regression coefficients were obtained by fitting the model. The correlation coefficient and the regression coefficient are weighted together to obtain the statistical association strength coefficient of the environmental exposure factor on the target variable; For each combination of patient behavior activity records and target variables, the steps from data collection to comprehensive weighting are repeated.

8. A health management system for chronic disease patients based on big data according to claim 7, characterized in that, The generation of the phased health intervention plan includes: The health intervention program includes medication adjustment recommendations, a list of behavioral pattern optimizations, and a follow-up visit plan; Based on the risk event type indicated in the disease risk warning signal, a set of corresponding intervention measures is matched from a pre-set intervention knowledge base; Based on the patient's current long-term medication records, a compatibility check is performed on the medication recommendations in the intervention set to screen out medication adjustment recommendations without conflicts; Based on the patient's personal daily routine preferences and the recorded behavioral activities, specific execution times, frequencies, and intensities are set for the behavioral intervention recommendations in the intervention set, forming a list of behavioral pattern optimizations; Based on the monitoring needs of key indicators indicated in the trend evolution graph of the test indicators, and the urgency of the disease risk warning signals, a follow-up plan including the next test items and the recommended follow-up time is formulated. The medication adjustment recommendations, behavioral pattern optimization list, and follow-up visit plan are integrated into a phased health intervention plan document with a clear implementation cycle.

9. The chronic disease patient health management system based on big data according to claim 1, characterized in that, The steps for constructing the chronic disease progression prediction model include: Collect a complete set of health monitoring data for a historical chronic disease patient population. The set of health monitoring data includes continuous vital sign monitoring data, periodic test data, environmental exposure data, and corresponding disease progression event records. The continuous vital signs monitoring data are processed by time series segmentation to extract the feature vector of vital signs fluctuation patterns for each patient at different time periods; The periodic test data are processed to extract trend features and generate a descriptor for the trajectory of test index changes for each patient. The environmental exposure data and patient behavioral activity records are subjected to feature encoding processing to obtain an environmental behavior feature matrix; The feature vector of the vital sign fluctuation pattern, the descriptor of the trajectory of the test index change, and the environmental behavior feature matrix are subjected to multi-source feature fusion processing to generate a comprehensive feature representation. Each patient sample is labeled with a disease progression status tag based on the disease progression event records; The time-series deep learning network is trained using the comprehensive feature representation and disease progression status labels. The network parameters are then optimized using the backpropagation algorithm to obtain the trained chronic disease progression prediction model.

Images