Postmenopausal osteoporosis risk prediction and traditional chinese and western medicine intervention management system

Abstract

The present application relates to the technical field of health risk assessment, in particular to a postmenopausal osteoporosis risk prediction and traditional Chinese and western medicine intervention management system. The system quantifies the trend risk of bone mineral density by the deviation change of bone mineral density risk value in time sequence, combines with the initial detection of bone mineral density risk to divide groups and analyze and identify the group risk degree; the comprehensive metabolic risk influence degree is obtained by coupling the hormone drop fluctuation characteristics with the bone metabolism marker change; the osteoporosis comprehensive influence degree is determined by using the group comprehensive bone mineral density risk degree dynamic adjustment, combining with the external risk influence degree and the comprehensive metabolic risk influence degree; and the neural network model is trained based on the osteoporosis comprehensive influence degree of sample data to predict the risk. The present application considers the multidimensional influence of postmenopausal patient's internal hormone metabolism and behavior habit, constructs the risk model by multidimensional coupling evaluation value, improves the risk prediction accuracy, and provides more comprehensive and reliable data support for the bone health management of postmenopausal women.

Description

Technical Field

[0001] This invention relates to the field of health risk assessment technology, specifically to a postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system. Background Technology

[0002] Postmenopausal women are prone to osteoporosis due to a significant decrease in bone density, making their bones more fragile, especially increasing the risk of fractures in the spine, hip, and wrist. Because osteoporosis causes limited mobility and chronic pain, leading to longer-term physical discomfort, risk assessments are often conducted to slow its progression and facilitate early intervention.

[0003] Existing bone health management systems typically rely solely on static assessments based on single bone mineral density (BMD) test data. However, bone loss in postmenopausal women exhibits complex, non-linear characteristics, influenced by multidimensional drivers of endogenous hormonal fluctuations and exogenous behavioral patterns. Simply relying on static test data makes it difficult to capture the time-varying risk characteristics coupled with multiple factors, resulting in a lagging and one-sided quantitative assessment of individual bone health status. Summary of the Invention

[0004] To address the technical problem in existing technologies where relying solely on static detection data makes it difficult to capture time-varying risk characteristics involving multiple coupled factors, resulting in a lagging and one-sided quantitative assessment of individual bone health status, the present invention aims to provide a postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system. The specific technical solution adopted is as follows:

[0005] This invention provides a postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system, the system comprising:

[0006] The data acquisition module is used to acquire each patient's bone mineral density risk value, estradiol data, and bone metabolism marker data in continuous testing from the sample data, as well as the degree of external risk impact based on exercise habit performance assessment.

[0007] The bone mineral density risk analysis module is used to obtain the high-risk bone mineral density manifestation level based on the continuous changes of each patient's bone mineral density risk value during the testing process and the deviation of bone mineral density risk value between the beginning and end of the test; by combining the bone mineral density risk value of the initial test and the high-risk bone mineral density manifestation level, different patients are divided into risk groups and the overall bone mineral density risk level of the group is determined.

[0008] The metabolic risk analysis module is used to determine the hormone decline stage for each patient based on the decreasing trend of estradiol data during continuous monitoring; to obtain the hormone risk trend impact based on the decreasing trend of hormone decline stage and the high volatility of estradiol data after the decline stage; and to obtain the comprehensive metabolic risk impact based on the synchronous changes of bone metabolism marker data during the hormone decline stage, combined with the hormone risk trend impact.

[0009] The prediction module is used to determine the comprehensive impact of osteoporosis on each patient by weighting and fusing the comprehensive metabolic risk impact value and the external risk impact value of the patient's risk group. The neural network model is trained based on the comprehensive impact of osteoporosis to make predictions.

[0010] Furthermore, the method for obtaining the high-risk bone mineral density expression includes:

[0011] For each patient, the difference between the last test in the series of tests and the bone mineral density risk value in the first test is used as the initial and final risk performance.

[0012] The sequence of bone mineral density risk values ​​continuously measured in a patient over time is taken as the bone mineral density risk sequence; the difference sequence of the bone mineral density risk sequence is obtained; the proportion of negative difference values ​​in the difference sequence is taken as the decay proportion; the proportion of positive difference values ​​in the difference sequence is taken as the improvement proportion.

[0013] The sum of all negative difference values ​​in the difference sequence and the product of the decay percentage are used as the decay risk level; the sum of all positive difference values ​​in the difference sequence and the product of the improvement percentage are used as the improvement risk level; the sum of the decay risk level and the improvement risk level are negatively correlated and used as the persistence risk performance level.

[0014] By combining the initial and final risk performance scores and the continuous risk performance scores, a high-risk bone density performance score is obtained.

[0015] Furthermore, obtaining the patient risk group and determining the group's comprehensive bone mineral density risk score includes:

[0016] The initial bone mineral density risk value and the degree of high bone mineral density manifestation of each patient are combined to form a feature vector; the feature vector is used as a metric to cluster all patients using a clustering algorithm, and each cluster is regarded as the risk group of each patient;

[0017] For any given patient risk group, the overall population bone mineral density risk score is obtained by combining the initial bone mineral density risk scores and high-risk bone mineral density manifestations of all patients.

[0018] Furthermore, the method for determining the hormone decline phase includes:

[0019] For any given patient, obtain the slope of the estradiol data for each test during the continuous testing process; tests with a slope less than a preset decline threshold are considered significant decline tests, and the earliest and latest significant decline tests in time sequence are used as the boundary, and the continuous tests between the boundary and in between are considered as the hormone decline stage for that patient.

[0020] Furthermore, the method for obtaining the influence of hormone risk trends includes:

[0021] Based on the initial and final values ​​of estradiol data for each patient during the hormone decline phase and the length of the hormone decline phase, the degree of hormone decline trend for each patient is determined.

[0022] The analysis phase consisted of all tests performed on each patient after the hormone decline phase. Based on the standard deviation of estradiol data in all tests during the analysis phase, the degree of high hormone fluctuation trend for each patient was determined.

[0023] By combining the hormone decline trend and hormone high volatility trend for each patient, the hormone risk trend impact of each patient is obtained.

[0024] Furthermore, the method for obtaining the comprehensive metabolic risk impact includes:

[0025] During the hormone decline phase in each patient, bone metabolism marker data were continuously measured over time to form a metabolite sequence; the differential sequence of the metabolite sequence was obtained, and the sum of all differential values ​​in the differential sequence was used as the degree of synchronous dramatic change for each patient.

[0026] By combining the degree of synchronous dramatic change and the influence of hormonal risk trends for each patient, the comprehensive metabolic risk influence for each patient is obtained.

[0027] Furthermore, the method for obtaining the comprehensive impact of osteoporosis includes:

[0028] For any patient, the normalized value of the overall bone mineral density risk score of the patient's risk group is used as the weighting factor.

[0029] The impact of the overall metabolic risk was weighted using a weighting factor, and the impact of the external risk was weighted using the difference between a constant 1 and the weighting factor; the weighted sum was taken as the overall impact of osteoporosis on the patient.

[0030] Furthermore, the prediction based on the neural network model trained on the comprehensive impact of osteoporosis includes:

[0031] The neural network model is used to establish a nonlinear mapping relationship between the patient's overall osteoporosis impact and the actual clinical severity of osteoporosis; it is trained based on sample data, with the input of the neural network model being the overall osteoporosis impact and the output being the osteoporosis severity risk assessment value; prediction is performed using the trained neural network model.

[0032] Furthermore, the method for obtaining the bone mineral density risk value includes:

[0033] The T-value of bone mineral density test in each test is negatively correlated and normalized, and then used as the bone mineral density risk value for each test.

[0034] Furthermore, after the prediction is performed using the neural network model trained based on the comprehensive impact of osteoporosis, the method further includes:

[0035] Patients in the sample data are clustered according to their osteoporosis severity risk assessment values ​​to obtain the assessment risk group; the mean of all osteoporosis severity risk assessment values ​​in the assessment risk group is used as the population risk assessment value of the assessment risk group; the assessment risk group whose population risk assessment value is closest to the predicted value of the patient to be tested is used as the reference group of the patient to be tested.

[0036] The intervention plan for the predetermined number of patients whose osteoporosis severity risk assessment value in the reference group is closest to the predicted value of the patients to be tested will be used as the intervention reference plan for the patients to be tested.

[0037] The present invention has the following beneficial effects:

[0038] This invention constructs a multi-dimensional data fusion risk assessment framework. First, it quantifies the trend risk of bone mineral density (BMD) by utilizing the deviation changes in BMD risk values ​​over time, and then classifies and identifies the risk status of groups based on initial examination results. On one hand, it considers the influence of intrinsic driving factors by coupling changes in bone metabolism markers with the fluctuation characteristics of hormone levels to obtain a comprehensive metabolic risk impact, quantifying endogenous metabolic risk. On the other hand, it considers extrinsic behavioral factors by quantifying the extrinsic risk impact of exercise behavior, combining it with the comprehensive metabolic risk impact, and dynamically adjusting the group's comprehensive BMD risk to adaptively fuse intrinsic and extrinsic risks to obtain a comprehensive assessment of the overall impact of osteoporosis. Finally, based on the comprehensive impact of osteoporosis from sample data, a neural network model processes multi-dimensional feature vectors to achieve risk prediction, facilitating subsequent matching of intervention programs. This invention considers the multi-dimensional influence of patients' postmenopausal intrinsic hormone metabolism and behavioral habits, coupling assessment values ​​across multiple dimensions to construct a risk model, improving the accuracy of risk prediction, and providing more comprehensive and reliable data support for bone health management in postmenopausal women. Attached Figure Description

[0039] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a structural diagram of a postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system provided in one embodiment of the present invention. Detailed Implementation

[0041] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0042] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0043] The following description, in conjunction with the accompanying drawings, details the specific scheme of the postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system provided by this invention.

[0044] Please see Figure 1 The diagram shows a structural diagram of a postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system provided by an embodiment of the present invention. The system includes: a data acquisition module 101, a bone density risk analysis module 102, a metabolic risk analysis module 103, and a prediction module 104.

[0045] The data acquisition module 101 is used to acquire each patient's bone mineral density risk value, estradiol data, and bone metabolism marker data in continuous testing from the sample data, as well as the degree of external risk impact based on exercise habit performance assessment.

[0046] With increased health awareness, many postmenopausal women undergo regular checkups, including bone mineral density (BMD) testing via DEXA scanning. This is the standard method for diagnosing osteoporosis, allowing for early detection of decreased bone density and assessment of fracture risk. However, postmenopausal patients are affected by physiological and psychological factors. For example, the decrease in estrogen levels after menopause directly leads to an imbalance in bone metabolism, increased bone resorption, and reduced bone formation. Stress and anxiety also affect hormone levels and bone metabolism, thus exacerbating bone density loss. Therefore, a risk model is established by coupling multi-dimensional features to facilitate accurate early prediction and provide more reliable data support for intervention programs.

[0047] In this embodiment of the invention, to construct analytical correlations, a dataset of tests conducted on postmenopausal female patients is collected as sample data. From this sample data, bone mineral density risk values, estradiol data, and bone metabolism marker data for each test are obtained. Bone metabolism marker data can be reflected by the concentration of bone degradation products, such as CTX. It is understood that the time-series test data undergoes preprocessing, which may include data standardization and time-scale normalization to facilitate unified data analysis and remove the influence of dimensions. It should be noted that data preprocessing is a technique well-known to those skilled in the art; for example, the continuous testing timescale can be set to a monthly unit. Specific settings can be adjusted by the implementer and are not elaborated or limited here.

[0048] Bone mineral density (BMD) test results can be reflected by the T-score, which typically indicates normal bone mineral density. A T-score between -1 and -2.5 indicates osteopenia, while a T-score less than -2.5 indicates osteoporosis. Therefore, in this embodiment of the invention, the T-score from each BMD test is negatively correlated and normalized to serve as the BMD risk value for each test. This more directly reflects the degree of risk of reduced bone mineral density; a higher risk value indicates a more significant osteoporosis.

[0049] It should be noted that negative correlation mapping and normalization are both techniques well known to those skilled in the art. For example, negative correlation mapping can be in the form of inverse proportion or negative exponentiation with the natural constant as the base, and the choice of normalization can be linear normalization or standard normalization, etc. The specific method is not limited here.

[0050] In addition to focusing on the patient's internal metabolic effects, external factors such as the patient's daily exercise habits, diet, and lifestyle also play an important role in the assessment of osteoporosis. For example, regular weight-bearing exercise and appropriate physical activity help increase bone density and reduce bone loss, while lack of exercise or bad habits can accelerate bone loss and increase the risk of fractures.

[0051] Therefore, in this embodiment of the invention, the external impact on patients is obtained through a questionnaire survey and an evaluation scale. The final score of the evaluation scale reflects the adverse effect of exercise habits on bone mass changes, and this score is used as the external risk impact level. A higher external risk impact level indicates a worse exercise habit. It should be noted that the scale is designed by the implementer according to the specific implementation scenario, and no specific limitations are imposed here.

[0052] The bone mineral density risk analysis module 102 is used to obtain the high-risk bone mineral density manifestation degree based on the continuous change of each patient's bone mineral density risk value during the testing process and the deviation of bone mineral density risk value between the beginning and the end of the test; and to classify different patients by combining the initial bone mineral density risk value and the high-risk bone mineral density manifestation degree, thereby obtaining the patient risk group and determining the overall bone mineral density risk degree of the group.

[0053] Considering the risk of osteoporosis is not only based on current bone mineral density (BMD), but also on the rate of BMD loss, which reflects the trend of osteoporosis risk. Over a long period, theoretically, patients may be able to slow bone loss and improve BMD through medication and diet, as advised by their doctor. However, the effectiveness of medication is affected by individual differences, and the combined influence of long-term metabolism and other factors may lead to a decline in BMD test results.

[0054] Therefore, for patients exhibiting high-risk characteristics, the overall risk of bone mineral density loss after multiple testing phases can be reflected by analyzing the initial and final test results. Furthermore, if the patient's bone mineral density continues to decline uncontrollably across multiple testing phases despite the use of medication and other treatments, it indicates a higher level of severe risk. Therefore, combining the continuous changes in bone mineral density risk allows for a more accurate assessment of the severity of the risk.

[0055] Preferably, in this embodiment of the invention, the method for obtaining the high-risk bone mineral density profile includes:

[0056] First, for each patient, the difference between the last test in the continuous testing and the bone mineral density risk value in the first test is used as the initial-to-final risk performance score. The higher the initial-to-final risk performance score, the greater the overall degree of bone mineral density loss.

[0057] Furthermore, the sequence of bone mineral density (BMD) risk values ​​from continuous temporal monitoring of patients is used as the BMD risk sequence. A difference sequence is then obtained from this BMD risk sequence. This difference sequence represents the change between adjacent monitoring periods, reflecting the severity or improvement of BMD risk between two consecutive monitoring tests. The percentage of negative difference values ​​in the difference sequence is calculated as the decay percentage, reflecting the proportion of periods showing decay during continuous monitoring. Simultaneously, the percentage of positive difference values ​​in the difference sequence is calculated as the improvement percentage, reflecting the proportion of periods showing improvement during continuous monitoring.

[0058] The more persistent the decline phases, the higher the sustained state of bone mineral density loss and the greater the risk of serious damage. Therefore, the sum of all negative difference values ​​in the difference sequence, multiplied by the decline ratio, is used as the decline risk level. Conversely, the more persistent the improvement phases, the better the control of bone mineral density loss and the lower the risk of serious damage. The sum of all positive difference values ​​in the difference sequence, multiplied by the improvement ratio, is used as the improvement risk level.

[0059] By negatively mapping the sum of the decay risk and the improvement risk as a measure of persistent risk, the changes in risk between consecutive tests are used to specifically analyze the persistent decay or improvement of bone density between multiple tests. The decay risk is negative and the improvement risk is positive, so the smaller the sum, the higher the risk.

[0060] Finally, by combining the initial and final risk performance scores and the persistent risk performance scores, a high-risk bone density performance score is obtained. The overall changes are adjusted through persistent performance scores to conduct a more comprehensive risk assessment. In one specific embodiment of the present invention, the sum of the normalized value of the persistent risk performance score and the constant 1 is used as a persistent control factor. The product of the initial and final risk performance scores and the high-risk bone density performance score is used as the high-risk bone density performance score. The higher the high-risk bone density performance score, the higher the risk of overall bone density decline detected so far.

[0061] Further consideration was given to the initial bone mineral density (BMD) risk score, which characterizes the baseline bone mass reserve at the beginning of postmenopausal testing, i.e., the initial risk. This was combined with the high-risk BMD score, which represents dynamic changes in bone mass, to comprehensively characterize the risk profile of each patient. Due to significant inter-individual variability, clustering was used to group a large number of patients into groups with similar bone mass reserves and changes. The overall group BMD risk score, derived from group-wide analysis, characterizes the average severity of risk within this specific group.

[0062] In this embodiment of the invention, a feature vector is constructed by combining the initial bone mineral density risk value and the high-risk bone mineral density manifestation level of each patient. These two features are then used as a distance metric. A clustering algorithm is applied to all patients using this feature vector as the metric, and each cluster represents a risk group for each patient. Patients with similar feature vectors are grouped together. It should be noted that clustering algorithms are well-known techniques to those skilled in the art, and algorithms such as DBSCAN can be used; these will not be elaborated upon here.

[0063] For any patient risk group, the overall population bone mineral density risk score is obtained by combining the initial bone mineral density risk scores and high-risk bone mineral density manifestation scores of all patients. In one specific embodiment of the present invention, the mean of the initial bone mineral density risk scores of all patients in the patient risk group is multiplied by the mean of the high-risk bone mineral density manifestation scores of all patients to obtain the overall population bone mineral density risk score, which characterizes the severity of the overall situation of the group.

[0064] By conducting direct risk analysis on the bone mineral density data of patients after long-term regular monitoring, we can obtain the risk groups of patients and the comprehensive bone mineral density risk of the groups. The magnitude of the bone mineral density risk can provide a basis for subsequent osteoporosis risk prediction for patients.

[0065] The metabolic risk analysis module 103 is used to determine the hormone decline stage for each patient based on the decreasing trend of estradiol data during continuous monitoring; to obtain the hormone risk trend impact based on the decreasing trend of hormone decline stage and the high volatility of estradiol data after the decline stage; and to obtain the comprehensive metabolic risk impact based on the synchronous change of bone metabolism marker data during the hormone decline stage, combined with the hormone risk trend impact.

[0066] While bone mineral density (BMD) data directly reflects the risk of osteoporosis, its intrinsic driving factors are directly influenced by the patient's hormones and bone metabolism biomarkers. Even if a patient's current BMD risk is low, abnormal manifestations in intrinsic driving factors can greatly increase the likelihood of future osteoporosis, leading to a higher actual risk.

[0067] The sharp decline in estrogen levels in postmenopausal women is a significant cause of rapid bone loss, especially estradiol, a component of estrogen. Estradiol directly stimulates the proliferation and differentiation of osteoblasts to promote bone formation, and it also promotes intestinal calcium absorption and reduces renal calcium excretion. A rapid decline in estradiol levels in postmenopausal women typically indicates a rapid loss of bone protection and a potential disruption of bone metabolism balance.

[0068] Therefore, the first step is to determine the significant decline phase of estradiol data after menopause. In this embodiment of the invention, for any patient, the slope of the estradiol data measured each time during continuous testing is obtained. The smaller the slope, that is, the higher the absolute value of the slope, the more significant the decline trend. Detections with a slope less than a preset decline threshold are considered significant decline detections, reflecting that the detection is still in the rapid decline phase. The preset decline threshold can be set to -0.5, and the specific value can be adjusted by the implementer according to the specific scenario of the embodiment, without limitation.

[0069] Furthermore, using the earliest and latest significant decrease detections in time sequence as boundaries, the continuous detections between the boundaries were taken as the hormone decline stages of the patient, thus preliminarily determining the rapid decline stage of estradiol after menopause.

[0070] After the rapid decline of estradiol, there is a slow decrease and stabilization phase, during which the metabolic impact is smaller. However, if estradiol data shows high fluctuations after the rapid decline phase, it may indicate that the patient has comorbidities, intermittent medication use, or significant weight changes. This high fluctuation may mean that the bones are constantly adapting to the changing internal environment, which is detrimental to bone health. Therefore, the decline trend during the decline phase and the subsequent high fluctuations can be used to comprehensively reflect the degree of risk of intrinsically driven hormones affecting the patient's bone health.

[0071] Preferably, in this embodiment of the invention, the method for obtaining the influence of hormone risk trends includes:

[0072] First, based on the changes in estradiol data at the beginning and end of each patient's hormone decline phase and the length of the hormone decline phase, the hormone decline trend degree for each patient is determined. The faster the decline, the greater the risk. In a specific embodiment of the present invention, the ratio of the difference between the initial and final estradiol data in the decline phase to the length of the hormone decline phase is used as the decline trend degree.

[0073] Further analysis was conducted on all tests performed on each patient after the hormone decline phase. Based on the standard deviation of estradiol data in all tests during the analysis phase, the degree of high hormone volatility for each patient was determined. The larger the standard deviation, the higher the degree of volatility, which would further increase the risk to bone metabolism.

[0074] Finally, by combining the hormone decline trend and hormone high volatility trend for each patient, the hormone risk trend impact for each patient is obtained. In a specific embodiment of the present invention, the sum of the normalized value of the hormone high volatility trend and the constant 1 is used as the trend regulating factor. The product of the trend regulating factor and the hormone decline trend is used as the patient's hormone risk trend impact. The larger the hormone risk trend impact, the higher the degree of hormone influence on bone mass changes and the greater the intrinsic driving risk.

[0075] When estrogen levels are stable, bone resorption and bone formation are in a relatively balanced state. However, when hormone levels drop rapidly, this balance is disrupted. Therefore, during the rapid decline of estradiol, the concentrations of bone metabolism markers change drastically, and this change is directly driven by the decline in estradiol.

[0076] If, during the simultaneous analysis of bone metabolism markers in patients, the concentrations of these markers also exhibit a rapid changing trend, it indicates that the bone metabolism markers are more consistent with the risk profile influenced by hormones. Specifically, the protective effect of estradiol disappears, and the inhibition of osteoclasts is relieved. Osteoclasts are massively activated, their activity increases dramatically, and they begin to rapidly break down type I collagen in bone, resulting in a large amount of degradation products being released into the bloodstream, causing their concentration to surge.

[0077] Therefore, in this embodiment of the invention, by combining the synchronous and dramatic changes in bone metabolism marker data, during the hormone decline phase of each patient, bone metabolism marker data are continuously detected over time to form a metabolite sequence, and the differential sequence of the metabolite sequence is obtained. The sum of all differential values ​​in the differential sequence is used as the synchronous and dramatic change degree of each patient. The greater the synchronous and dramatic change degree, the more consistent it is with the characteristics of bone metabolism markers being affected by hormones, the greater the degree of intrinsic driving factors, and the greater the impact on the risk of osteoporosis.

[0078] Therefore, by combining the synchronous drastic change degree and the influence degree of hormone risk trend for each patient, the comprehensive metabolic risk influence degree for each patient is obtained. In a specific embodiment of the present invention, the product of the synchronous drastic change degree and the influence degree of hormone risk trend for each patient is taken as the comprehensive metabolic risk influence degree for the patient.

[0079] The prediction module 104 is used to determine the comprehensive impact of osteoporosis on each patient by weighting and fusing the comprehensive metabolic risk impact value and the external risk impact value of the patient's risk group; and to train a neural network model for prediction based on the comprehensive impact of osteoporosis.

[0080] The overall bone mineral density risk of the patient's risk group is used as the basic metric for multidimensional assessment. When the overall bone mineral density risk of the patient's risk group is higher, the patient may face a higher risk of osteoporosis, the root cause and urgency of bone mineral density loss are stronger, the possibility of reversal by external factors is lower, and the internal driving factors of physiological processes are dominant. Therefore, priority should be given to the risk impact formed by the internal driving factors.

[0081] Conversely, when the overall bone mineral density risk level of the patient's risk group is lower, the driving force of intrinsic factors is weaker, and more attention can be paid to the impact of external risks on the patient. The intervention and regulation effect based on external factors such as exercise habits will be more prominent.

[0082] Therefore, in this embodiment of the invention, for any patient, the normalized value of the overall bone mineral density risk score of the patient's risk group is used as a weighting factor. The overall metabolic risk impact is weighted using this weighting factor, and the difference between a constant 1 and the weighting factor is used to weight the external risk impact. The weighted sum is taken as the patient's overall osteoporosis impact score, comprehensively considering the multi-dimensional impact risk of osteoporosis on the patient. As an example, the expression for the overall osteoporosis impact score is:

[0083] In the formula, This represents the overall impact of osteoporosis on the patient. Represented as weighting factors; This is expressed as the overall metabolic risk impact. This is expressed as the degree of external risk impact.

[0084] Thus, neural network training can be performed using sample data to predict osteoporosis risk. A non-linear mapping relationship between the overall impact of osteoporosis on a patient and the actual clinical severity of osteoporosis can be established through the neural network model. In this embodiment, all collected sample data are assessed for osteoporosis severity risk. After processing all data, the training and validation sets are divided in a 7:3 ratio. The neural network is trained using the training set samples. The input to the neural network model is a feature vector of overall osteoporosis impact, bone mineral density risk value, and external risk impact. The output is the osteoporosis severity risk assessment value. The loss function is the cross-entropy function. Gradient descent is used to train until the damage function converges. The robustness of the training results is verified using the validation set, thus obtaining the trained neural network model.

[0085] Furthermore, a neural network model is used to predict the risk of osteoporosis in postmenopausal patients, providing risk warnings. At the same time, to improve the efficiency of subsequent intervention and management, in this embodiment of the invention, patients in the sample data are clustered according to the osteoporosis severity risk assessment value to obtain the risk assessment group. The mean of all osteoporosis severity risk assessment values ​​in the risk assessment group is used as the group risk assessment value of the risk assessment group. Clustering based on the assessment results of the sample data reflects the patient group under the same risk.

[0086] After obtaining the predicted values ​​for the patients to be tested, the risk assessment group whose population risk assessment value is closest to the predicted value is used as the reference group for the patients to be tested. In other words, the risk assessment group with the smallest difference between the population risk assessment value and the predicted value is used as the reference group. The historical treatment plans of patients in the reference group (both traditional Chinese medicine and Western medicine) can provide historical evidence for the patients to be tested, and further screening of more similar patient plans can be used as a reference. Here, the difference is the absolute value of the difference between the data.

[0087] In this embodiment of the invention, the intervention plans of a predetermined number of patients in the reference group whose osteoporosis severity risk assessment values ​​are closest to the predicted values ​​are used as the intervention reference plans for the patients to be tested. That is, the difference between the osteoporosis severity risk assessment value and the predicted value for each patient in the reference group is calculated, the differences are arranged in ascending order, and the patients in the reference group corresponding to the top predetermined number of differences are selected as reference patients. The intervention plans of these reference patients are the intervention reference plans for the patients to be tested, providing a reference for the plan and improving the efficiency of subsequent interventions. The predetermined number can be set to 7, and the specific value can be adjusted by the implementer and is not limited here.

[0088] In summary, this invention constructs a multi-dimensional data fusion risk assessment framework. First, it quantifies the trend risk of bone mineral density (BMD) by utilizing the deviation changes in BMD risk values ​​over time, and then classifies and identifies the risk status of groups based on initial examination results. On one hand, it considers the influence of intrinsic driving factors by coupling changes in bone metabolism markers with the fluctuation characteristics of hormone levels to obtain a comprehensive metabolic risk impact, quantifying endogenous metabolic risk. On the other hand, it considers extrinsic behavioral factors by quantifying the extrinsic risk impact of exercise behavior, combining it with the comprehensive metabolic risk impact, and using the group's comprehensive BMD risk as a dynamic regulating factor to adaptively fuse intrinsic and extrinsic risks to obtain a comprehensive assessment of the overall impact of osteoporosis. Finally, based on the comprehensive impact of osteoporosis from sample data, a neural network model processes multi-dimensional feature vectors to achieve risk prediction, facilitating subsequent matching of intervention programs. This invention considers the multi-dimensional influence of patients' postmenopausal intrinsic hormone metabolism and behavioral habits, coupling assessment values ​​across multiple dimensions to construct a risk model, improving the accuracy of risk prediction, and providing more comprehensive and reliable data support for bone health management in postmenopausal women.

[0089] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0090] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system, characterized in that, The system includes: The data acquisition module is used to acquire each patient's bone mineral density risk value, estradiol data, and bone metabolism marker data in continuous testing from the sample data, as well as the degree of external risk impact based on exercise habit performance assessment. The bone mineral density risk analysis module is used to obtain the high-risk bone mineral density manifestation level based on the continuous changes of each patient's bone mineral density risk value during the testing process and the deviation of bone mineral density risk value between the beginning and end of the test; by combining the bone mineral density risk value of the initial test and the high-risk bone mineral density manifestation level, different patients are divided into risk groups and the overall bone mineral density risk level of the group is determined. The metabolic risk analysis module is used to determine the hormone decline stage for each patient based on the decreasing trend of estradiol data during continuous monitoring; to obtain the hormone risk trend impact based on the decreasing trend of hormone decline stage and the high volatility of estradiol data after the decline stage; and to obtain the comprehensive metabolic risk impact based on the synchronous changes of bone metabolism marker data during the hormone decline stage, combined with the hormone risk trend impact. The prediction module is used to determine the overall impact of osteoporosis on each patient by weighting and fusing the overall metabolic risk impact and the external risk impact based on the overall bone mineral density risk value of the patient's risk group; and to train a neural network model for prediction based on the overall impact of osteoporosis. The methods for determining the hormone decline phase include: For any given patient, obtain the slope of the estradiol data for each test during the continuous testing process; tests with a slope less than a preset decline threshold are considered significant decline tests, and the earliest and latest significant decline tests in time sequence are used as the boundary, and the continuous tests between the boundary and in between are considered as the hormone decline stage for that patient.

2. The postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system according to claim 1, characterized in that, The methods for obtaining the high-risk bone mineral density expression include: For each patient, the difference between the last test in the series of tests and the bone mineral density risk value in the first test is used as the initial and final risk performance. The sequence of bone mineral density risk values ​​continuously measured in a patient over time is taken as the bone mineral density risk sequence; the difference sequence of the bone mineral density risk sequence is obtained; the proportion of negative difference values ​​in the difference sequence is taken as the decay proportion; the proportion of positive difference values ​​in the difference sequence is taken as the improvement proportion. The sum of all negative difference values ​​in the difference sequence and the product of the decay percentage are used as the decay risk level; the sum of all positive difference values ​​in the difference sequence and the product of the improvement percentage are used as the improvement risk level; the sum of the decay risk level and the improvement risk level are negatively correlated and used as the persistence risk performance level. By combining the initial and final risk performance scores and the continuous risk performance scores, a high-risk bone density performance score is obtained.

3. The postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system according to claim 1, characterized in that, The process of obtaining the patient risk group and determining the group's comprehensive bone mineral density risk includes: The initial bone mineral density risk value and the degree of high bone mineral density manifestation of each patient are combined to form a feature vector; the feature vector is used as a metric to cluster all patients using a clustering algorithm, and each cluster is regarded as the risk group of each patient; For any given patient risk group, the overall population bone mineral density risk score is obtained by combining the initial bone mineral density risk scores and high-risk bone mineral density manifestations of all patients.

4. The postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system according to claim 1, characterized in that, The methods for obtaining the impact of hormone risk trends include: Based on the initial and final values ​​of estradiol data for each patient during the hormone decline phase and the length of the hormone decline phase, the degree of hormone decline trend for each patient is determined. The analysis phase consisted of all tests performed on each patient after the hormone decline phase. Based on the standard deviation of estradiol data in all tests during the analysis phase, the degree of high hormone fluctuation trend for each patient was determined. By combining the hormone decline trend and hormone high volatility trend for each patient, the hormone risk trend impact of each patient is obtained.

5. The postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system according to claim 1, characterized in that, The methods for obtaining the comprehensive metabolic risk impact include: During the hormone decline phase in each patient, bone metabolism marker data were continuously measured over time to form a metabolite sequence; the differential sequence of the metabolite sequence was obtained, and the sum of all differential values ​​in the differential sequence was used as the synchronous drastic change degree for each patient. By combining the degree of synchronous dramatic change and the influence of hormonal risk trends for each patient, the comprehensive metabolic risk influence for each patient is obtained.

6. The postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system according to claim 1, characterized in that, The method for obtaining the comprehensive impact of osteoporosis includes: For any patient, the normalized value of the overall bone mineral density risk score of the patient's risk group is used as the weighting factor. The impact of the overall metabolic risk was weighted using a weighting factor, and the impact of the external risk was weighted using the difference between a constant 1 and the weighting factor; the weighted sum was taken as the overall impact of osteoporosis on the patient.

7. The postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system according to claim 1, characterized in that, The prediction based on the neural network model trained on the comprehensive impact of osteoporosis includes: The neural network model is used to establish a nonlinear mapping relationship between the patient's overall osteoporosis impact and the actual clinical severity of osteoporosis. It is trained based on sample data. The input of the neural network model is the feature vector of overall osteoporosis impact, bone mineral density risk value and external risk impact, and the output is the osteoporosis severity risk assessment value. Prediction is performed through the trained neural network model.

8. The postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system according to claim 1, characterized in that, The method for obtaining the bone mineral density risk value includes: The T-value of bone mineral density test in each test is negatively correlated and normalized, and then used as the bone mineral density risk value for each test.

9. The postmenopausal osteoporosis risk prediction and integrated traditional Chinese and Western medicine intervention management system according to claim 7, characterized in that, After the prediction is performed using the neural network model trained based on the comprehensive impact of osteoporosis, the following steps are also included: Patients in the sample data are clustered according to their osteoporosis severity risk assessment values ​​to obtain the assessment risk group; the mean of all osteoporosis severity risk assessment values ​​in the assessment risk group is used as the population risk assessment value of the assessment risk group; the assessment risk group whose population risk assessment value is closest to the predicted value of the patient to be tested is used as the reference group of the patient to be tested. The intervention plan for the predetermined number of patients whose osteoporosis severity risk assessment value in the reference group is closest to the predicted value of the patients to be tested will be used as the intervention reference plan for the patients to be tested.

Images