Return-to-post cancer surviver health management system

Abstract

The invention relates to the technical field of electric digital data processing, in particular to a post-returning cancer surviver health management system which comprises a health data segmentation module, a state feature modeling module, a health state transfer module, a health track prediction module and a health risk identification module. According to the method, the blood oxygen saturation and heart rate data are extracted and analyzed in real time, the data change rate and trend characteristics are combined, abnormal changes are accurately recognized, the health state is recognized in a segmented mode, the accuracy and sensitivity of health monitoring are effectively improved, the dynamic characteristic parameter set is further constructed through multi-parameter comprehensive analysis, and the accuracy and sensitivity of health monitoring are improved. The dynamic recognition and adjustment ability of the health state stage is improved by combining the change of the psychological stress index, the future health evolution path and the potential risk turning point are gradually calculated based on the state transition model and the prediction algorithm, visual prediction and early warning are provided for individual health, and the initiative and accuracy of health management are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic digital data processing, and in particular to a health management system for cancer survivors returning to work. Background Art

[0002] The technical field of electronic digital data processing includes technologies for realizing functions such as information collection, storage, processing, transmission, and display through electronic devices and digital means. Its core content includes using computers and related software and hardware systems to process and manage various types of data, and is widely applied in multiple fields such as industrial control, medical management, financial services, education and training, etc. The overall systematic feature of this technical field lies in optimizing the processing and utilization of information flows through methods such as algorithms, logical analysis, and data modeling, ensuring the accuracy, efficiency, and security of information processing, while promoting cross-field information sharing and collaborative work.

[0003] Among them, the health management system for cancer survivors returning to work refers to a health management platform specifically designed for patients who return to work after cancer treatment. The patent theme mainly focuses on the health condition tracking, risk assessment, and daily health management of cancer survivors. It uses data collection terminals to record individual physiological index data, conducts health status analysis through classification algorithms, and provides personalized health suggestions based on the analysis results. The technical matters involved in the system cover real-time monitoring of physiological data, risk early warning based on calculation models, and health information feedback, etc. It mainly obtains key health data of users through electronic sensing devices, completes data analysis by combining logical algorithms with fixed rules, and finally provides a health management plan for users through a visual interface.

[0004] Existing technologies mostly rely on static data analysis, lack the ability of continuous monitoring and real-time response, are difficult to capture subtle changes in health status in a timely manner, and may lead to lagging risk identification. At the same time, the analysis of individual differences is insufficient, and the general data processing mode cannot meet the complex and changeable health needs of cancer survivors, resulting in lack of pertinence in health suggestions. Existing systems have limited capabilities in data modeling and state prediction, and the identification of future health evolution trends and potential risk points is relatively rough, which may lead to lagging or inaccurate intervention measures, and this poses a relatively high risk to cancer survivors who require meticulous management. Summary of the Invention

[0005] The purpose of the present invention is to solve the deficiencies existing in the prior art, and to propose a health management system for cancer survivors returning to work.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solutions: A health management system for cancer survivors returning to work includes:

[0007] The health data segmentation module extracts the blood oxygen saturation and heart rate data of cancer survivors based on wearable device data, calculates the data changes within a continuous time period, compares the heart rate fluctuations with the health threshold, identifies abnormal sections, marks the abnormal changes in the blood oxygen trend, combines the analysis results of heart rate and blood oxygen data, identifies the health status stage, and establishes a health status segmentation interval;

[0008] The state feature modeling module analyzes the blood oxygen change trend and heart rate fluctuation amplitude of each section based on the health status segmentation interval, calculates the multi-parameter change rate, extracts the trend direction and fluctuation interval features, analyzes the change of the psychological stress index, and integrally constructs a segmented state dynamic feature parameter set;

[0009] The health status transition module analyzes the change relationship between the psychological stress index and the health status stage based on the segmented state dynamic feature parameter set, extracts the consistency of continuous stage features, calculates the state transition probability, adjusts the transition weight, and establishes a health status transition matrix;

[0010] The health trajectory prediction module gradually calculates the change probability of the blood oxygen trend and the psychological stress index within the future time period based on the health status transition matrix, combines the transition probability, predicts the evolution path of the health status, marks the potential risk turning points, and generates a health trajectory prediction path;

[0011] The health risk identification module extracts the high-probability risk stage features based on the health trajectory prediction path, analyzes the change trend of the psychological stress index within the risk stage, marks the key risk nodes, extracts the features before and after the key nodes, and generates health risk key point indicators.

[0012] The health status segmentation interval specifically refers to heart rate fluctuations, abnormal changes in blood oxygen trend, and health status stage. The segmented state dynamic feature parameter set includes trend direction, fluctuation interval features, and changes in the psychological stress index. The health status transition matrix specifically refers to state transition probability and transition weight. The health trajectory prediction path includes future blood oxygen trend, change probability of the psychological stress index, and potential risk turning points. The health risk key point indicators specifically refer to high-probability risk stage features, change trend of the psychological stress index, and key risk nodes.

[0013] As a further solution of the present invention, the steps for obtaining the health status segmentation interval are specifically as follows:

[0014] Based on wearable device data, by segmenting the blood oxygen saturation and heart rate data, calculating the change rate, identifying the sections with significant data changes within a continuous time period, extracting the segmented data and performing normalization calculation on the change rate, calculating the normalized change rate value according to the difference between the values before and after each data point, and establishing a marked interval, generating marked blood oxygen and heart rate change rate sections;

[0015] Using the marked blood oxygen and heart rate change rate segments, the difference between the heart rate fluctuation amplitude and the healthy threshold is calculated, the change rate of the data points in the marked segment is called to compare point by point with the preset threshold, all data point segments exceeding the threshold are retained and other segments are eliminated to obtain the abnormal change segment;

[0016] The direction and amplitude of the blood oxygen trend change in the abnormal change section are marked using the formula:

[0017]

[0018] Calculate the degree of abnormal trend change, extract and mark the segments where the degree of abnormal trend change is greater than the judgment threshold, and generate segments marking the blood oxygen trend change and heart rate fluctuation amplitude;

[0019] Among them, D represents the degree of abnormal trend change, H i represents the fluctuation amplitude of the i-th item of heart rate data, S i represents the rate of change of the i-th item of blood oxygen saturation, n represents the total number of data points in the segment, and the square root of the denominator is used to normalize the trend change;

[0020] Based on the segments that mark the blood oxygen trend changes and heart rate fluctuation amplitudes, combined with the blood oxygen trend and heart rate change direction of each segment, the segment change amplitudes are set with interval ranges, and the health status stages are divided according to the intersection of the blood oxygen and heart rate change ranges to generate health status segmented intervals.

[0021] As a further solution of the present invention, the step of acquiring the segment state dynamic feature parameter set is specifically as follows:

[0022] Analyze the blood oxygen change trend and heart rate fluctuation amplitude of each segment through the health status segmented interval, extract the blood oxygen change direction and heart rate fluctuation range of the data points in the interval, calculate the difference between the previous and next values ​​of each data point in the interval, judge the trend direction according to the difference, and calculate the trend change rate based on the average value of the difference, and establish the initial trend direction and fluctuation amplitude characteristics;

[0023] Utilizing the initial trend direction and fluctuation amplitude characteristics, calculating the change rates of multiple parameters in multiple intervals, performing point-to-point matching between the blood oxygen change direction and the heart rate fluctuation amplitude in each interval, calling the trend direction parameter and the fluctuation amplitude parameter to perform weighted difference calculation, and normalizing the weighted difference to generate a combined result of the multi-parameter change rates;

[0024] The combination results of the multi-parameter change rates are integrated with the changes in the psychological stress index and the formula is used:

[0025]

[0026] Calculate the dynamic parameters of the psychological stress index within the interval and generate a set of dynamic characteristic parameters for the segmented state;

[0027] Among them, P ij represents the j-th dynamic characteristic parameter in the i-th interval, R ij represents the multi-parameter change rate of the j-th heart rate fluctuation amplitude within the i-th interval, T ij represents the multi-parameter change rate of the j-th blood oxygen change direction within the i-th interval, W i represents the weight adjustment coefficient of the i-th interval, which is used to balance the data weights. i represents the serial number of the segmented interval of the health state, j represents the serial number of the characteristic parameter within the interval, the denominator part is the normalized feature matching degree, and the characteristic calculation range is adjusted.

[0028] As a further solution of the present invention, the steps for obtaining the health state transition matrix are specifically as follows:

[0029] Based on the set of dynamic characteristic parameters of the segmented state, analyze the change relationship between the psychological stress index and the health state stage, extract the intervals with consistent parameter changes in consecutive stages, group the data using the change rate and trend continuity, obtain the statistical change characteristics of the psychological stress index in each group, and generate data with consistent characteristics in consecutive stages;

[0030] Using the data with consistent characteristics in consecutive stages, calculate the state transition probability between health state stages, compare the change rate differences between different stages according to the change of the psychological stress index in each stage, adjust the initial transition weight between stages through the difference size, and perform normalization to form a transition probability matrix;

[0031] Adjust the weight of the transition probability matrix, referring to the psychological stress change and consistency characteristics between stages, and use the formula:

[0032]

[0033] Calculate the transition weight from stage i to stage j and generate a health state transition matrix;

[0034] Among them, T ij represents the state transition weight from stage i to stage j, showing the transition probability between two stages, P ij represents the initial transition probability from stage i to stage j, reflecting the probability obtained from directly observed data, V ij represents the difference in psychological stress change, referring to the influence of the psychological state change between stages on the transition probability, W ij is the adjustment weight during the transition process, which is used to modify the deviation in the original data.

[0035] As a further solution of the present invention, the steps for obtaining the health trajectory prediction path are specifically as follows:

[0036] Based on the health status transition matrix, calculate the blood oxygen trend and the change probability of the psychological stress index in the future time period, analyze the transition probability parameters of each state in the transition matrix, calculate the state change trend at the future time point by comparing the state probability distributions between stages, and generate a health status probability distribution;

[0037] Combined with the health status probability distribution, continuously predict the health status in the future time period using the transition probability, establish a prediction path for different health states according to the state transition values and time series changes at each time point, and iteratively calculate the state values at multiple time points in sequence to generate a health status evolution path;

[0038] According to the health status evolution path, mark potential risk turning points, and use the formula:

[0039]

[0040] Conduct risk assessment, calculate the risk assessment value at time t, and generate a health trajectory prediction path;

[0041] where, R(t) represents the risk assessment value at time t, reflecting the potential health risk at the target time point, p i represents the transition probability from state i - 1 to state i, reflecting the predictability of the state, t i represents the time required to reach state i, measuring the time span of state change, and n is the total number of reference states.

[0042] As a further solution of the present invention, the steps for obtaining the health risk key point indicators are specifically as follows:

[0043] Based on the health trajectory prediction path, extract the characteristics of the high - probability risk stage, analyze the change trend of the psychological stress index within the stage, calculate the point - by - point change rate of the psychological stress index, combine the trend line change to analyze the risk significance, and determine the risk stage to generate preliminary risk stage data;

[0044] Using the preliminary risk stage data, mark the key risk nodes, detect the local extreme points for the positions where the change rate of the psychological stress index fluctuates sharply, mark the rising and falling nodes in combination with the change trend direction, and eliminate errors through multiple trend verifications to generate a list of key risk nodes;

[0045] Analyze the characteristic changes before and after the list of key risk nodes, and use the formula:

[0046]

[0047] Calculate and generate the health risk key point indicators;

[0048] Among them, I(t) represents the key index of health risk at time t, indicating the comprehensive risk degree of the change in the psychological stress index. ΔP(t) represents the change amount of the psychological stress index at time t, reflecting the absolute value of the change, and σ P represents the overall standard deviation of the psychological stress index. The normalized data is used to eliminate the scale influence in the differentiation stage. W c is an adjustment coefficient used to balance the calculation deviation when the standard deviation is small.

[0049] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0050] In the present invention, by extracting and analyzing the blood oxygen saturation and heart rate data in real time, combining the data change rate and trend characteristics, abnormal changes are accurately identified and the health status is identified in segments, effectively improving the accuracy and sensitivity of health monitoring. Further, through multi-parameter comprehensive analysis, a dynamic characteristic parameter set is constructed, and combined with the change of the psychological stress index, the dynamic recognition and adjustment ability of the health status stage is improved. Based on the state transition model and prediction algorithm, the future health evolution path and potential risk turning points are gradually calculated, providing visual prediction and early warning for individual health, and enhancing the initiative and accuracy of health management. The comprehensive application of segment identification, dynamic modeling and path prediction technologies strengthens the refinement and personalization of the health management of cancer survivors. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is the system flow chart of the present invention;

[0052] Figure 2 is the flow chart of the steps for obtaining the segmented interval of the health status of the present invention;

[0053] Figure 3 is the flow chart of the steps for obtaining the dynamic characteristic parameter set of the segmented state of the present invention;

[0054] Figure 4 is the flow chart of the steps for obtaining the health status transition matrix of the present invention;

[0055] Figure 5 is the flow chart of the steps for obtaining the predicted path of the health trajectory of the present invention;

[0056] Figure 6 is the flow chart of the steps for obtaining the key index of health risk of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0057] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present 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 only used to explain the present invention, and are not used to limit the present invention.

[0058] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by terms such as "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is 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 should not be construed as a limitation to the present invention. In addition, in the description of the present invention, the meaning of "a plurality of" is two or more, unless otherwise specifically defined.

[0059] Embodiment 1

[0060] Please refer to Figure 1 , a health management system for cancer survivors returning to work includes:

[0061] The health data segmentation module extracts the blood oxygen saturation and heart rate data of cancer survivors based on the data of wearable devices, calculates the data changes within a continuous time period, compares the heart rate fluctuations with the health threshold, identifies abnormal sections, marks the abnormal changes in the blood oxygen trend, combines the analysis results of heart rate and blood oxygen data, identifies the health status stage, and establishes a health status segmentation interval;

[0062] The state feature modeling module analyzes the blood oxygen change trend and heart rate fluctuation amplitude of each section based on the health status segmentation interval, calculates the multi-parameter change rate, extracts the trend direction and fluctuation interval features, analyzes the change of the psychological stress index, and integrally constructs a segmented state dynamic feature parameter set;

[0063] The health status transfer module analyzes the change relationship between the psychological stress index and the health status stage based on the segmented state dynamic feature parameter set, extracts the consistency of continuous stage features, calculates the state transfer probability, adjusts the transfer weight, and establishes a health status transfer matrix;

[0064] The health trajectory prediction module gradually calculates the change probability of the blood oxygen trend and the psychological stress index within the future time period based on the health status transfer matrix, combines the transfer probability, predicts the evolution path of the health status, marks the potential risk turning points, and generates a health trajectory prediction path;

[0065] The health risk identification module extracts the high-probability risk stage features based on the health trajectory prediction path, analyzes the change trend of the psychological stress index within the risk stage, marks the key risk nodes, extracts the features before and after the key nodes, and generates health risk key point indicators.

[0066] The specific segmented ranges of the health status are specifically the heart rate fluctuations, abnormal changes in the blood oxygen trend, and the health status stages. The set of dynamic characteristic parameters of the segmented status includes the trend direction, the characteristics of the fluctuation range, and the change in the psychological stress index. The health status transition matrix specifically refers to the state transition probability and the transition weight. The health trajectory prediction path includes the future blood oxygen trend, the change probability of the psychological stress index, and the potential risk turning points. The key indicators of health risks are specifically the characteristics of the high-probability risk stage, the change trend of the psychological stress index, and the key risk nodes.

[0067] Please refer to Figure 2 , and the steps for obtaining the segmented ranges of the health status are specifically as follows:

[0068] Based on the data from wearable devices, by segmenting the blood oxygen saturation and heart rate data, calculating the change rate, identifying the segments with significant data changes within consecutive time periods, extracting the segmented data and performing a normalization calculation on the change rate, calculating the normalized change rate value according to the difference between the values before and after each data point, and establishing a marked interval to generate the marked segments of the blood oxygen and heart rate change rates;

[0069] Divide the time periods of the collected blood oxygen saturation and heart rate data using a sliding window method. Each time period corresponding to a window is a sub-dataset. According to the difference between the values before and after the data points within the sub-dataset, calculate the absolute value of the change rate through the formula. The change rate formula is: R i =|X i+1 -X i |, where R i represents the change rate of the i-th point, and X i+1 and X i represent the data points at the current and the next moment respectively. By calculating the change rate of all data points point by point, perform a normalization operation on all the change rate values in each window. The normalization formula is: The normalized values are used to eliminate the windows with smaller change rates, and the windows with significant change rates are selected as the initial marked intervals; the marked segments of the blood oxygen and heart rate change rates are established through the above process.

[0070] Using the marked segments of the blood oxygen and heart rate change rates, calculate the difference value between the heart rate fluctuation amplitude and the health threshold, call the change rate of the data points in the marked segments and compare them point by point with the preset threshold, retain all the data point segments that exceed the threshold and eliminate the other segments to obtain the abnormal change segments;

[0071] Calculate the difference value between the heart rate fluctuation amplitude and the health threshold according to the change rate within the marked segments. First, calculate the average fluctuation amplitude within each marked segment. The average fluctuation amplitude formula is: where A k represents the average fluctuation amplitude of the k-th segment, and R iis the change rate value for each point within the section, and n is the total number of data points within the section; Let A k be compared point by point with the preset healthy threshold range T l and T u to retain the sections that meet the condition T l ≤A k ≤T u . These sections are regarded as potentially abnormally changing sections, and the sections that do not meet the conditions are excluded to obtain the abnormally changing sections.

[0072] Mark the direction and amplitude of the blood oxygen trend change for the abnormally changing sections. Use the formula:

[0073]

[0074] Calculate the degree of abnormal trend change. Extract and mark the sections where the degree of abnormal trend change is greater than the determination threshold to generate sections marking the blood oxygen trend change and the heart rate fluctuation amplitude;

[0075] Among them, D represents the degree of abnormal trend change, H i represents the fluctuation amplitude of the i-th item of heart rate data, S i represents the change rate of the i-th item of blood oxygen saturation, n represents the total number of data points within the section, and the square sum and square root calculation in the denominator are used for normalizing the trend change;

[0076] Formula:

[0077]

[0078] The advantage of the formula is that by jointly considering the heart rate fluctuation amplitude and the blood oxygen change rate, introducing the difference between the two into the denominator normalization process, avoiding the influence of the absolute difference on the result, and at the same time enhancing the dynamic adaptability of the calculation by combining the number of data points n in the section, which helps to accurately extract the abnormally changing trend sections.

[0079] Detailed explanation of the formula and the derivation process of formula calculation:

[0080] Calculate the absolute value of the difference between the heart rate fluctuation amplitude H i and the blood oxygen saturation change rate S i . Calculate the sum of the absolute values of the differences for all data points within the section according to the formula. The calculation process is as follows:

[0081] First, calculate the value of the heart rate fluctuation amplitude H i . The formula is: H i =|HR i+1 -HR i |, where HR i+1 and HR i are the heart rate values at the current and next moments respectively;

[0082] Then, calculate the change rate S of blood oxygen saturation i value, and the formula is: where, SpO2 i+1 and SpO2 i are the blood oxygen values at the current and next moments respectively, and SpO2 max and SpO2 min are the maximum and minimum blood oxygen values within this section respectively;

[0083] Calculate the sum of absolute differences, and the formula is:

[0084] Calculate the normalization factor of the denominator, and the formula is:

[0085] Substitute Sum and Denominator into the formula to calculate the D value;

[0086] The calculation result is D = 0.65. This value indicates that there is a certain abnormal trend in the blood oxygen and heart rate fluctuations within the section. Extract and mark the sections where the degree of abnormal trend change is greater than the determination threshold of 0.5, and generate sections marking the blood oxygen trend change and heart rate fluctuation amplitude.

[0087] Parameter explanation: D represents the degree of abnormal trend change, H i represents the fluctuation amplitude of the i-th point in the heart rate data, S i represents the change rate of the i-th item of blood oxygen saturation, n represents the total number of data points within the section, Sum represents the total sum of absolute differences, Denominator represents the value used to normalize the trend change range in the denominator part, SpO2 max and SpO2 min represent the range of blood oxygen values respectively, and HR i represents the heart rate data point.

[0088] Based on the sections marking the blood oxygen trend change and heart rate fluctuation amplitude, combined with the blood oxygen trend and heart rate change direction of each section, set the range of the section change amplitude respectively, and divide the health status stage according to the intersection of the blood oxygen and heart rate change ranges to generate the health status segmentation range.

[0089] Classify the data points within each section according to the blood oxygen change direction and amplitude respectively. By calculating the average change rate and fluctuation amplitude of the blood oxygen saturation data within each section, the range setting formula is: C k =[min(R k ),max(R k )], where C k represents the change range of the k-th section, min(R k ) and max(R k) respectively represent the minimum and maximum values of the blood oxygen change rate within the section; the range of the heart rate change amplitude is set in the same way, and the intersection of the blood oxygen change range and the heart rate change range is used as the healthy state stage range, and the healthy state segmentation interval is generated through the above process.

[0090] Please refer to Figure 3 , and the steps for obtaining the dynamic characteristic parameter set of the segmentation state are specifically as follows:

[0091] Analyze the blood oxygen change trend and heart rate fluctuation amplitude of each section through the healthy state segmentation interval, extract the blood oxygen change direction and heart rate fluctuation range of the data points within the interval, calculate the difference between the values before and after each data point within the interval, judge the trend direction based on the magnitude of the difference, and calculate the trend change rate based on the average value of the differences to establish the initial trend direction and fluctuation amplitude characteristics;

[0092] For each section of blood oxygen data, call the difference before and after the data point for calculation, and use (O i+1 -O i ) to represent the difference of each data point and establish a difference array, and further calculate the average value of the differences Judge the blood oxygen trend direction of each section as positive (rising trend) or negative (falling trend) according to the average value result, and perform the same operation on the heart rate data. Use the difference before and after the heart rate (H i+1 -H i ) to calculate the average value of the differences Calculate the volatility in combination with its fluctuation range Normalize the comparison results of the positive and negative signs of the segmented trend direction and the volatility to generate the trend change rate and fluctuation amplitude characteristics. Finally, match the blood oxygen change direction and the heart rate fluctuation range with the trend positive and negative sign array and the normalized fluctuation amplitude to establish the initial trend direction and fluctuation amplitude characteristics.

[0093] Using the initial trend direction and fluctuation amplitude characteristics, calculate the multi-parameter change rate within multiple intervals, perform point-to-point matching on the blood oxygen change direction and heart rate fluctuation amplitude within each interval, call the trend direction parameter and the fluctuation amplitude parameter for weighted difference calculation, and generate the combined result of the multi-parameter change rate according to the normalization of the weighted difference;

[0094] Input the initial trend direction and fluctuation amplitude of each section into the function, calculate the average value (T avg ) and standard deviation (T std ) of the difference array of the data points of each section, and establish the change range interval [T avg -T std , T avg +T std, the change rate is normalized to [0, 1], and then the trend direction and fluctuation amplitude are further called. The weighted difference calculation formula for the absolute value of the trend direction and the fluctuation amplitude is integrated as (α·T + β·R), where α and β are adjustment coefficients, which are dynamically adjusted according to the change ranges of heart rate and blood oxygen data respectively. The multi-parameter change rate combination result is generated through the matching degree of the normalized eigenvalue of multiple parameters in each interval.

[0095] For the multi-parameter change rate combination result, the change of the psychological stress index is integrated, and the formula is used:

[0096]

[0097] Calculate the dynamic parameters of the psychological stress index within the interval, and generate a set of dynamic characteristic parameters for the segmented state;

[0098] Among them, P ij represents the jth dynamic characteristic parameter in the ith interval, R ij represents the multi-parameter change rate of the jth heart rate fluctuation amplitude within the ith interval, T ij represents the multi-parameter change rate of the jth blood oxygen change direction within the ith interval, W i represents the weight adjustment coefficient of the ith interval, which is used to balance the data weight. i represents the serial number of the segmented interval of the health state, j represents the serial number of the characteristic parameter within the interval, and the denominator part is the normalized feature matching degree, which adjusts the feature calculation range.

[0099] Formula:

[0100]

[0101] The benefit of the formula is that by integrating the multi-parameter change rates of the blood oxygen change direction and the heart rate fluctuation amplitude, and adding the weight adjustment coefficient W i , the comprehensive evaluation accuracy of the dynamic characteristic matching degree of parameters in different intervals is enhanced;

[0102] Detailed explanation of the formula and the derivation process of the formula calculation:

[0103] R ij represents the multi-parameter change rate of the jth heart rate fluctuation amplitude within the ith interval, and is calculated through the normalization formula Calculated; T ij represents the multi-parameter change rate of the jth blood oxygen change direction within the ith interval, and is calculated through the normalization formula Calculated; W i represents the weight adjustment coefficient of the ith interval, which is dynamically adjusted through the ratio of the fluctuation frequencies of the heart rate and blood oxygen in the historical statistical interval, and is set to fluctuate between (0.8, 1.2);

[0104] Assume that in a certain interval \(i = 1\), the data points are \(H=\{80, 85, 90\}\) and \(O=\{95, 98, 100\}\) respectively. The multi-parameter change rate of the heart rate fluctuation amplitude is calculated as follows:

[0105]

[0106] The multi-parameter change rate of the blood oxygen change direction is calculated as follows:

[0107]

[0108] Set the weight adjustment coefficient \(W\) i \(= 1.0\), substitute it into the formula for calculation:

[0109]

[0110] The result shows that the dynamic characteristic parameter \(P\) of the psychological stress index within the interval 11 \(= 0.428\), \(P\) 12 \(= 0.635\) reflects the dynamic degree of the matching between the heart rate and the blood oxygen change. Combining the dynamic characteristic values of all intervals can generate a segmented state dynamic characteristic parameter set.

[0111] Please refer to Figure 4 , the specific steps for obtaining the health state transition matrix are as follows:

[0112] Based on the segmented state dynamic characteristic parameter set, analyze the change relationship between the psychological stress index and the health state stage, extract the intervals with consistent parameter changes in the continuous stages, group the data using the change rate and trend continuity, obtain the statistical change characteristics of the psychological stress index for each group, and generate continuous stage feature consistency data;

[0113] Calculate the point-by-point change rate for each psychological stress index value. By recording the psychological stress index (the stress value measured by the wearable device) and the timestamp for each stage, obtain the psychological stress change data at each time point within the stage. The formula for calculating the change rate for the data within the stage is: where \(\Delta P\) i is the stress change rate at the \(i\)-th time point, \(P\) i+1 and \(P\) i are the psychological stress values at the current time point and the next time point respectively, \(T\) i+1 and \(T\) i are the corresponding timestamps. By calculation, obtain the psychological stress change rate sequence for each stage. Then, perform mean processing on the absolute values of this sequence to obtain the average change rate, which is used to measure the consistency of the continuous stages. Compare the stress change trends between different stages and screen out the intervals with lower change rates and smaller differences. Finally, use the mean change rate and the difference range to mark the intervals with consistent parameters within the stage, and generate continuous stage feature consistency data.

[0114] Using the consistent data of consecutive stage features, calculate the state transition probability between health status stages. According to the change of psychological stress index in each stage, compare the difference in the change rate between different stages, adjust the initial transition weight between stages according to the size of the difference, and perform normalization to form a transition probability matrix;

[0115] Adjust the initial transition weight between stages according to the size of the difference, record the average psychological stress in each stage (monitor by wearing a device, record the stress value and take the average), calculate the difference value of the change rate in each stage, and use the formula: Where ΔR ij is the difference in the change rate from stage i to stage j, are the average psychological stresses of stage i and stage j respectively. Perform normalization on the difference value. The normalization formula is: Where ΔR′ ij is the normalized difference value, n is the total number of stages. Adjust the normalized difference value and allocate the weight to the transfer relationship between each stage to generate a transition probability matrix.

[0116] Adjust the weight of the transition probability matrix. Refer to the psychological stress change and consistency characteristics between stages, and use the formula:

[0117]

[0118] Calculate the transition weight from stage i to stage j to generate a health status transition matrix;

[0119] Where, T ij represents the state transition weight from stage i to stage j, showing the transition probability between two stages, P ij represents the initial transition probability from stage i to stage j, reflecting the probability obtained from direct observation data, V ij represents the difference in psychological stress change. Refer to the influence of the psychological state change between stages on the transition probability, W ij is the adjustment weight in the transfer process, used to modify the deviation in the original data.

[0120] Formula:

[0121]

[0122] The advantage of the formula is that by combining the psychological stress difference value and the state consistency weight, it improves the accuracy and flexibility of the health status stage transition weight, and at the same time optimizes the transfer priority of different stages by adjusting the weight parameters;

[0123] Detailed explanation of the formula and the derivation process of the formula calculation:

[0124] P ijThe value of [[]] is obtained by a monitoring device to acquire the initial transition probabilities of two stages. For example, the probability of stage 1 is 0.3 and the probability of stage 2 is 0.2; V ij It is obtained through differential calculation that the pressure change value between stages is 0.1; W ij Adjust the weight according to the actual application. The weight setting from stage 1 to stage 2 is based on the stage importance and takes the value of 0.5; Substitute it into the formula for calculation:

[0125]

[0126] First step, calculate the absolute value of the numerator:

[0127] |0.3 - 0.1| = 0.2;

[0128] Second step, calculate the square root of the sum of squares of the denominator:

[0129]

[0130] Third step, calculate the weight adjustment part:

[0131]

[0132] Finally, calculate the transition weight:

[0133]

[0134] The result shows that the transition weight from stage 1 to stage 2 is 3.16. This weight reflects the optimization of the transition probability between the two stages under pressure change and consistency adjustment, indicating that the priority of the transition from stage 1 to stage 2 is relatively high. The result is used to optimize the weight distribution in the health state transition matrix to generate the health state transition matrix.

[0135] Please refer to Figure 5 , and the specific steps for obtaining the health trajectory prediction path are as follows:

[0136] Based on the health state transition matrix, calculate the blood oxygen trend and the change probability of the psychological stress index in the future time period, analyze the transition probability parameters of each state in the transition matrix, and calculate the state change trend at the future time point by comparing the state probability distributions between stages to generate the health state probability distribution;

[0137] First, extract the transition probability value P of each stage in the health state transition matrix ij and the corresponding initial state value S i , and obtain the state distribution probability at the future time point through step-by-step transition accumulation calculation. For example, for the transition probability P ij , it can be gradually deduced in the form of matrix multiplication. Assume that the current health state is S i, whose next state S corresponding in the transition matrix i+1 has a probability value of P ij ·S i , where P ij is the transition probability from state i to state j. Iteratively update this result in sequence to calculate the state distribution trend within multiple time points. Then, for the analysis of the change rates of blood oxygen trend and psychological stress index, the health parameter values of each state can be combined, and using the change rate calculation formula gradually analyze the change rates of blood oxygen values or psychological stress index under two consecutive states. For example, when the blood oxygen value pointed to by the state transition probability P ij is 95% in the initial state and 90% in the final state, the change rate is That is, the blood oxygen decline rate corresponding to this state is 5.26%. The calculation of the psychological stress index can be carried out for change rate analysis in the same way. Through the accumulation of the change rates of all states, the comprehensive change probability distribution in the future time period is obtained, and the health state probability distribution is generated.

[0138] Combined with the health state probability distribution, use the transition probability to continuously predict the health state in the future time period. According to the state transition values and time series changes at each time point, establish a prediction path for different health states, and iteratively calculate the state values at multiple time points in sequence to generate a health state evolution path;

[0139] First, extract all the health states and their probability values at the current time point from the health state probability distribution. For example, S 1 = 40%, S 2 = 35%, S 3 = 25%. According to the transition probability P ij in the transition matrix, gradually update the state probability distribution at each time point. For example, the state distribution at time point t + 1 is calculated from the probability distribution at time point t and the transition matrix. For example, the probability value of state S 2 at time point t + 1 is S 1 ·P 12 + S 2 ·P 22 + S 3 ·P 32 . Calculate the state distribution probabilities at multiple future time points in sequence. At the same time, mark the state with the highest probability value in the state distribution at each time point as the main health state path. Combining time series analysis, calculate the change trend of each state path within a continuous time period. For example, through the time series trend formula calculate the trend scores of each state path, where x i is the state probability value at each time point, and t iis the time value at the corresponding time point. By combining the accumulation of this trend score and the accumulation of probability values, the evolution path of the health state is obtained. At the same time, the time point state values in each evolution path are corrected to ensure the logical consistency of state changes, and the health state evolution path is generated.

[0140] According to the health state evolution path, potential risk turning points are marked, and the formula is used:

[0141]

[0142] to conduct risk assessment, calculate the risk assessment value at time t, and generate the health trajectory prediction path;

[0143] where R(t) represents the risk assessment value at time t, reflecting the potential health risk at the target time point, and p i represents the transition probability from state i - 1 to state i, reflecting the predictability of the state, and t i represents the time required to reach state i, measuring the time span of state changes, and n is the total number of reference states.

[0144] Formula:

[0145]

[0146] The advantage of the formula is that by non-linearly processing the comprehensive influence of the cumulative transition probability value and the time series on the risk assessment value, the sensitivity and accuracy of the risk assessment are enhanced, and at the same time, it is ensured that the results of the risk values fluctuate within a reasonable range.

[0147] Detailed explanation of the formula and the derivation process of formula calculation:

[0148] First, calculate the product of the transition probability p i and the time series t i . Assume that the current state contains three stages with n = 3, which are p 1 = 0.4, p 2 = 0.3, p 3 = 0.2, and the corresponding time values are t 1 = 1, t 2 = 2, t 3 = 3. Then the cumulative sum calculation is:

[0149]

[0150] Next, substitute the result into the formula to calculate the risk assessment value R(t), that is

[0151] Calculate the exponential part e -1.6 ≈ 0.2019;

[0152]

[0153] The risk assessment value is R(t) = 0.8324.

[0154] This result indicates that the risk assessment value at the current time point t is 0.8324, indicating that the potential health risk is at a relatively high level. Combining with the risk turning points marked on the health status evolution path, this result can be further used to judge the potential risk area and generate a predicted health trajectory path.

[0155] Please refer to Figure 6 , and the specific steps for obtaining the key indicators of health risk are as follows:

[0156] Based on the predicted health trajectory path, extract the characteristics of the high-probability risk stage, analyze the change trend of the psychological stress index within the stage, calculate the point-by-point change rate of the psychological stress index, combine the trend line change to analyze the risk significance, and determine the risk stage to generate preliminary risk stage data;

[0157] First, screen out the change values of the psychological stress index for each time period from the predicted health trajectory path, calculate the rate of change of stress point by point as the initial input for data analysis. Then, compare the rate of change of stress for each time period with the mean of the historical rate of change of stress. Determine the risk significance according to the interval division of the historical mean, where the significance is defined by the multiple threshold of the standard deviation of the stress rate greater than the mean. For example, define the time period with a change rate exceeding twice the standard deviation of the mean as the high-risk stage. Subsequently, summarize all high-risk stages into preliminary risk stage data, and combine the continuous time points of the risk stage and the trend characteristics of the stress change rate to eliminate the time periods with insufficient significance to ensure the accuracy of the risk stage data and generate preliminary risk stage data.

[0158] Use the preliminary risk stage data to mark the key risk nodes, detect the local extreme points for the positions where the change rate of the psychological stress index fluctuates sharply, mark the rising and falling nodes in combination with the change trend direction, and eliminate errors through multiple trend verifications to generate a list of key risk nodes;

[0159] By analyzing the rate of change of pressure at each time point within the risk stage, the extreme value points of the rate of change are identified. These extreme value points are the local maximum or minimum of the rate of change. First, the rate of change of pressure at each time point is smoothed using a sliding window to weaken the noise fluctuations in the data. Then, based on the relative value of the rate of change within the sliding window, the change direction of each local extreme value point is marked. A sharp increase in the rate of change is marked as a risk increase node, and a sharp decrease in the rate of change is marked as a risk decrease node. By repeatedly detecting and verifying the trend consistency between the smoothed result and the original data, mislabeled nodes or noise nodes are removed. Finally, the extreme value points that have passed multiple verifications are marked as key risk nodes, and a list of key risk nodes is generated.

[0160] Analyze the characteristic changes before and after the list of key risk nodes, and use the formula:

[0161]

[0162] Calculate and generate the key index of health risk;

[0163] Among them, I(t) represents the key index of health risk at time t, indicating the comprehensive risk degree of the change of the psychological stress index. ΔP(t) represents the change amount of the psychological stress index at time t, reflecting the absolute value of the change. σ P represents the overall standard deviation of the psychological stress index. Normalize the data to eliminate the scale effect of the differentiation stage. W c is an adjustment coefficient used to balance the calculation deviation when the standard deviation is small.

[0164] Formula:

[0165]

[0166] The advantage of the formula is that by squaring the change amount of pressure to emphasize the significance of the change value, and at the same time combining the standard deviation for normalization, adding the adjustment coefficient W c to balance the calculation deviation that may occur when the standard deviation is small, thereby improving the sensitivity and accuracy to abnormal risk changes.

[0167] Detailed explanation of the formula and the derivation process of the formula calculation:

[0168] First, the change amount ΔP(t) of the psychological stress index is obtained through pressure monitoring data collection. This value is obtained by calculating the difference of the psychological stress index point by point. For example, at t = 1, assume P(1) = 80, and at t = 2, assume P(2) = 100, then:

[0169] ΔP(2) = P(2) - P(1) = 100 - 80 = 20;

[0170] Next, by analyzing the changing trend of the psychological stress index over the entire time period, the overall standard deviation σ of the stress index is calculated P , which is calculated by the formula , where P i is the stress index value at each time point, is the mean of the stress indices at all time points. Assuming P = [80, 100, 120, 90], then:

[0171]

[0172] The calculation results are as follows:

[0173]

[0174] Finally, an adjustment coefficient W c is set, which is dynamically adjusted according to the fluctuation range of the stress change rate. Its value depends on the size of the analysis window and the historical fluctuation range of the stress index. For example, W c = 10. Substitute the above parameters into the formula:

[0175]

[0176] This result shows that the stress change at time point t = 2 is highly significant, and the magnitude of the result I(t) is related to the absolute value and fluctuation range of the stress change, which can be used to further mark the key points of health risks and generate key indicators of health risks.

[0177] The above are only the preferred embodiments of the present invention, and do not limit the present invention in other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A health management system for cancer survivors returning to work, characterized in that: The system comprises: The health data segmentation module extracts the blood oxygen saturation and heart rate data of cancer survivors based on wearable device data, calculates the data changes in continuous time periods, compares the heart rate fluctuations with the health threshold, identifies abnormal segments, marks abnormal changes in blood oxygen trends, combines the heart rate and blood oxygen data analysis results, identifies the health status stage, and establishes the health status segmentation interval; The state feature modeling module analyzes the blood oxygen change trend and heart rate fluctuation amplitude of each segment based on the health state segmentation interval, calculates the multi-parameter change rate, extracts the trend direction and fluctuation interval characteristics, analyzes the change of the psychological stress index, and integrates and constructs the segmented state dynamic feature parameter set; The health state transfer module analyzes the changing relationship between the psychological stress index and the health state stage based on the segmented state dynamic feature parameter set, extracts the consistency of continuous stage features, calculates the state transfer probability, adjusts the transfer weight, and establishes a health state transfer matrix; The health trajectory prediction module gradually calculates the blood oxygen trend and the probability of change of the psychological stress index in the future time period based on the health state transfer matrix, and predicts the health state evolution path by combining the transfer probability, marking the potential risk turning point, and generating the health trajectory prediction path; The health risk identification module extracts the characteristics of the high-probability risk stage based on the health trajectory prediction path, analyzes the changing trend of the psychological stress index within the risk stage, marks the key risk nodes, extracts the characteristics before and after the key nodes, and generates health risk key point indicators.

2. The health management system for cancer survivors returning to work according to claim 1, characterized in that: The health status segmented intervals specifically include heart rate fluctuations, abnormal changes in blood oxygen trends, and health status stages. The segmented state dynamic feature parameter set includes trend direction, fluctuation interval characteristics, and changes in the psychological stress index. The health status transfer matrix specifically refers to the state transition probability and transfer weight. The health trajectory prediction path includes future blood oxygen trends, probability of changes in the psychological stress index, and potential risk turning points. The health risk key point indicators specifically include high probability risk stage characteristics, psychological stress index change trends, and key risk nodes.

3. The health management system for cancer survivors returning to work according to claim 2, characterized in that: The steps for obtaining the health status segment interval are specifically as follows: Based on the wearable device data, the blood oxygen saturation and heart rate data are segmented, the change rate is calculated, and the segments with significant data changes in the continuous time period are identified. The segmented data is extracted and the change rate is normalized. According to the difference between the previous and next values ​​of each data point, the change rate normalization value is calculated and the marked interval is established to generate the marked blood oxygen and heart rate change rate segments; Using the marked blood oxygen and heart rate change rate segments, the difference between the heart rate fluctuation amplitude and the healthy threshold is calculated, the change rate of the data points in the marked segment is called to compare point by point with the preset threshold, all data point segments exceeding the threshold are retained and other segments are eliminated to obtain the abnormal change segment; The direction and amplitude of the blood oxygen trend change in the abnormal change section are marked using the formula: Calculate the degree of abnormal trend change, extract and mark the segments where the degree of abnormal trend change is greater than the judgment threshold, and generate segments marking the blood oxygen trend change and heart rate fluctuation amplitude; Among them, D represents the degree of abnormal trend change, H i represents the fluctuation amplitude of the i-th item of heart rate data, S i represents the rate of change of the i-th item of blood oxygen saturation, n represents the total number of data points in the segment, and the square root of the denominator is used to normalize the trend change; Based on the segments that mark the blood oxygen trend changes and heart rate fluctuation amplitudes, combined with the blood oxygen trend and heart rate change direction of each segment, the segment change amplitudes are set with interval ranges, and the health status stages are divided according to the intersection of the blood oxygen and heart rate change ranges to generate health status segmented intervals.

4. The health management system for cancer survivors returning to work according to claim 3, characterized in that: The steps for obtaining the segment state dynamic feature parameter set are specifically as follows: Analyze the blood oxygen change trend and heart rate fluctuation amplitude of each segment through the health status segmented interval, extract the blood oxygen change direction and heart rate fluctuation range of the data points in the interval, calculate the difference between the previous and next values ​​of each data point in the interval, judge the trend direction according to the difference, and calculate the trend change rate based on the average value of the difference, and establish the initial trend direction and fluctuation amplitude characteristics; Utilizing the initial trend direction and fluctuation amplitude characteristics, calculating the change rates of multiple parameters in multiple intervals, performing point-to-point matching between the blood oxygen change direction and the heart rate fluctuation amplitude in each interval, calling the trend direction parameter and the fluctuation amplitude parameter to perform weighted difference calculation, and normalizing the weighted difference to generate a combined result of the multi-parameter change rates; The combination results of the multi-parameter change rates are integrated with the changes in the psychological stress index and the formula is used: Calculate the dynamic parameters of the psychological stress index within the interval and generate a segmented state dynamic characteristic parameter set; Among them, P ij represents the jth dynamic characteristic parameter in the i-th interval, R ij represents the multi-parameter change rate of the jth heart rate fluctuation amplitude in the i-th interval, T ij Represents the multi-parameter change rate of the j-th blood oxygen change direction in the i-th interval, W i Represents the weight adjustment coefficient of the i-th interval, which is used to balance the data weight. i represents the serial number of the health status segment interval, j represents the serial number of the feature parameter in the interval, and the denominator normalizes the feature matching degree to adjust the feature calculation range.

5. The health management system for cancer survivors returning to work according to claim 4, characterized in that: The steps for obtaining the health state transfer matrix are specifically as follows: Based on the segmented state dynamic characteristic parameter set, the relationship between the psychological stress index and the health state stage is analyzed, the intervals with consistent parameter changes in the continuous stages are extracted, the data are grouped using the change rate and trend continuity, the statistical change characteristics of each group of psychological stress index are obtained, and the continuous stage characteristic consistency data is generated; Utilizing the continuous stage feature consistency data, the state transition probability between health state stages is calculated, and according to the change of the psychological stress index in each stage, the difference in the change rate between the differentiated stages is compared, and the initial transfer weight between the stages is adjusted according to the difference, and normalized to form a transition probability matrix; The weight of the transition probability matrix is ​​adjusted, referring to the psychological stress changes and consistency characteristics between stages, using the formula: Calculate the transfer weight from stage i to stage j and generate the health state transfer matrix; Among them, T ij represents the state transition weight from stage i to stage j, showing the transition probability between the two stages, P ij represents the initial transition probability from stage i to stage j, reflecting the probability obtained by direct observation data, V ij represents the difference in psychological pressure change, referring to the impact of psychological state changes between stages on the transition probability, W ij is the adjustment weight in the transfer process, used to modify the bias in the original data.

6. The health management system for cancer survivors returning to work according to claim 5, characterized in that: The steps for obtaining the health trajectory prediction path are specifically as follows: Based on the health state transition matrix, the blood oxygen trend and the probability of change of the psychological stress index in the future time period are calculated, the transition probability parameters of each state in the transition matrix are analyzed, and the state change trend at the future time point is calculated by comparing the state probability distribution between stages to generate the health state probability distribution; Combined with the health status probability distribution, the transition probability is used to continuously predict the health status in the future time period, and a prediction path of differentiated health status is established according to the state transition value and time series change at each time point, and the state values ​​at multiple time points are iteratively calculated in turn to generate a health status evolution path; According to the health status evolution path, mark the potential risk turning points, using the formula: Conduct risk assessment, calculate the risk assessment value at time t, and generate a health trajectory prediction path; Among them, R(t) represents the risk assessment value at time t, reflecting the potential health risk at the target time point, p i represents the transition probability from state i-1 to state i, reflecting the predictability of the state, t i It represents the time required to reach state i, measuring the time span of state change, and n is the total number of reference states.

7. The health management system for cancer survivors returning to work according to claim 6, characterized in that: The steps for obtaining the health risk key point indicators are specifically as follows: Based on the health trajectory prediction path, extract the high probability risk stage characteristics, analyze the change trend of the psychological stress index within the stage, calculate the point-by-point change rate of the psychological stress index, analyze the risk significance in combination with the trend line change, and determine the risk stage to generate preliminary risk stage data; Using the preliminary risk stage data, key risk nodes are marked, local extreme point detection is performed for locations where the rate of change of the psychological stress index fluctuates sharply, rising and falling nodes are marked in combination with the direction of the change trend, errors are eliminated through multiple trend verifications, and a list of key risk nodes is generated; Analyze the characteristic changes before and after the key risk node list, using the formula: Calculate and generate key health risk indicators; Among them, I(t) represents the key health risk index at time t, indicating the comprehensive risk level of the change in the psychological stress index, ΔP(t) represents the change in the psychological stress index at time t, reflecting the absolute value of the change, and σ P represents the overall standard deviation of the psychological stress index, normalizing the data to eliminate the scale effect of the differentiation stage, W c is an adjustment factor used to balance the calculated deviation when the standard deviation is smaller.

Images