Dynamic monitoring method and early warning system for endocrine hormone level

Abstract

The invention relates to an endocrine hormone level dynamic monitoring method and an early warning system, and relates to the technical field of hormone monitoring, the method comprises the following steps: determining physiological influence factors, influence diseases and influence drugs of a plurality of candidate endocrine hormones, and establishing a knowledge graph; determining a plurality of similar patients from the plurality of sample patients according to the knowledge graph and the physiological information, the disease information, the drug information and the plurality of target endocrine hormones of the to-be-detected patient; according to the endocrine hormone level dynamic monitoring data of the plurality of similar patients, determining correlation characteristics among the plurality of target endocrine hormones, determining at least one key endocrine hormone from the plurality of target endocrine hormones, and determining the dynamic monitoring frequency of the to-be-detected patient; according to the dynamic content of each key endocrine hormone, at least one key endocrine hormone and the dynamic monitoring frequency are dynamically adjusted, and the method has the advantage of improving the flexibility and accuracy of endocrine hormone level monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of hormone monitoring, and in particular to a method and an early warning system for dynamic monitoring of endocrine hormone levels. Background Art

[0002] The endocrine system plays a vital role in the human body, regulating multiple physiological functions including metabolism, immune response, growth and development, and emotions. With changes in lifestyle, increased environmental pollution, and an aging population, the incidence of endocrine system diseases (such as diabetes, thyroid disease, polycystic ovary syndrome, etc.) is on the rise. Effectively monitoring and managing the health of the endocrine system and effectively evaluating endocrine regulation strategies have become important topics in modern medicine.

[0003] While existing endocrine hormone monitoring technologies have played an important role in clinical applications, they still have some limitations. For example, they require frequent blood draws or the use of high-precision testing equipment, which is costly and limits their widespread clinical application. For example, mass spectrometry (LC-MS / MS), while highly sensitive and specific, is expensive and complex to operate. Frequent blood draws can cause discomfort to patients, reduce compliance, and compromise data integrity.

[0004] Therefore, it is necessary to provide a dynamic monitoring method and early warning system for endocrine hormone levels to improve the flexibility and accuracy of endocrine hormone level monitoring. Summary of the Invention

[0005] The present invention provides a method for dynamic monitoring of endocrine hormone levels, including: determining the physiological influencing factors, influencing diseases and influencing drugs of multiple candidate endocrine hormones, and establishing a knowledge graph; determining multiple similar patients from multiple sample patients based on the knowledge graph and the physiological information, disease information, drug information and multiple target endocrine hormones of the patients to be tested; determining the correlation characteristics between the multiple target endocrine hormones based on the dynamic monitoring data of the endocrine hormone levels of multiple similar patients; determining at least one key endocrine hormone from the multiple target endocrine hormones based on the correlation characteristics between the multiple target endocrine hormones; determining the dynamic monitoring frequency of the patient to be tested based on the knowledge graph, at least one key endocrine hormone and the physiological information, disease information and drug information of the patient to be tested; collecting the dynamic content of each key endocrine hormone based on the dynamic monitoring frequency of the patient to be tested; judging whether to adjust at least one key endocrine hormone and the dynamic monitoring frequency based on the dynamic content of each key endocrine hormone, and if so, collecting the dynamic content of each key endocrine hormone based on the adjusted at least one key endocrine hormone and the dynamic monitoring frequency.

[0006] Furthermore, the physiological influencing factors, influencing diseases, and influencing drugs of multiple candidate endocrine hormones are determined, including: for each candidate endocrine hormone, obtaining physiological information of multiple sample patients and the dynamic content of the candidate endocrine hormone; for each candidate physiological factor, calculating the nonlinear correlation coefficient between the candidate physiological factor and the candidate endocrine hormone based on the physiological information, disease information, drug information of multiple sample patients and the dynamic content of the candidate endocrine hormone; based on the nonlinear correlation coefficient between each candidate physiological factor and the candidate endocrine hormone, screening the physiological influencing factors of the candidate endocrine hormone; for each candidate disease, determining the diseased patient group and the unaffected patient group corresponding to the candidate disease, and calculating the disease impact value of the candidate disease on the candidate endocrine hormone based on the diseased patient group and the unaffected patient group corresponding to the candidate disease; based on the disease impact value of each candidate disease on the candidate endocrine hormone, screening the disease affecting the candidate endocrine hormone; for each candidate drug, determining the patient group and the patient group taking the candidate drug, and calculating the drug impact value of the candidate drug on the candidate endocrine hormone based on the diseased patient group and the unaffected patient group corresponding to the candidate drug; based on the drug impact value of each candidate drug on the candidate endocrine hormone, screening the drug affecting the candidate endocrine hormone.

[0007] Furthermore, based on the knowledge graph and the physiological information, disease information, drug information and multiple target endocrine hormones of the patient to be tested, multiple similar patients are determined from multiple sample patients, including: for each target endocrine hormone, based on the knowledge graph, the impact characteristics corresponding to the target endocrine hormone are extracted from the physiological information, disease information and drug information of the patient to be tested; for each sample patient, based on the impact characteristics corresponding to each target endocrine hormone of the patient to be tested and the impact characteristics corresponding to each target endocrine hormone of the sample patient, the similarity between the patient to be tested and the sample patient is calculated; based on the similarity between the patient to be tested and the sample patient, multiple similar patients are determined from multiple sample patients.

[0008] Furthermore, based on the dynamic monitoring data of endocrine hormone levels of multiple similar patients, the correlation characteristics between multiple target endocrine hormones are determined, including: for any two target endocrine hormones, based on the dynamic monitoring data of endocrine hormone levels of multiple similar patients, the nonlinear correlation coefficients of the two target endocrine hormones are calculated, wherein the correlation characteristics between the multiple target endocrine hormones include the nonlinear correlation coefficients of any two target endocrine hormones.

[0009] Furthermore, based on the correlation characteristics between multiple target endocrine hormones, at least one key endocrine hormone is determined from multiple target endocrine hormones, including: for each target endocrine hormone, based on the dynamic monitoring data of endocrine hormone levels of multiple similar patients, determining the abnormal probability of the target endocrine hormone; based on the abnormal probability of each target endocrine hormone, determining at least one first key endocrine hormone; based on the correlation characteristics between the first key endocrine hormone and multiple target endocrine hormones, determining at least one second key endocrine hormone, wherein the at least one key endocrine hormone includes each first key endocrine hormone and each second key endocrine hormone.

[0010] Furthermore, based on the correlation characteristics between the first key endocrine hormone and multiple target endocrine hormones, at least one second key endocrine hormone is determined, including: for each target endocrine hormone, based on the nonlinear correlation coefficient of any two target endocrine hormones, determining the related target endocrine hormones of the target endocrine hormone; based on the related target endocrine hormones of each target endocrine hormone, establishing a relationship graph; for each node of the relationship graph, based on the abnormal probability of the first key endocrine hormone and the nonlinear correlation coefficient of any two target endocrine hormones, calculating the key value of the node; based on the key value of each node, determining the candidate key endocrine hormone; deduplicating the candidate key endocrine hormones and the first key endocrine hormone, and determining at least one second key endocrine hormone.

[0011] Furthermore, the dynamic monitoring frequency of the patient to be tested is determined based on the knowledge graph, at least one key endocrine hormone and the physiological information, disease information and drug information of the patient to be tested, including: determining the dynamic monitoring frequency of the patient to be tested based on the abnormal probability of each key endocrine hormone.

[0012] Furthermore, based on the dynamic content of each key endocrine hormone, it is determined whether to adjust at least one key endocrine hormone and the dynamic monitoring frequency, including: based on the dynamic content of each key endocrine hormone, determining whether there is at least one abnormal key endocrine hormone; if so, adjusting at least one key endocrine hormone and the dynamic monitoring frequency; if not, not adjusting at least one key endocrine hormone and the dynamic monitoring frequency.

[0013] Furthermore, adjusting at least one key endocrine hormone and the dynamic monitoring frequency includes: for each abnormal key endocrine hormone, determining a newly added key endocrine hormone based on the relationship map; and adjusting the dynamic monitoring frequency of the patient to be tested based on the adjusted abnormal probability of each key endocrine hormone.

[0014] The present invention provides a dynamic early warning system for endocrine hormone levels, which is used to implement the above-mentioned dynamic monitoring method for endocrine hormone levels, including: a map establishment module, which is used to determine the physiological influencing factors, influencing diseases and influencing drugs of multiple candidate endocrine hormones, and establish a knowledge map; a similarity search module, which is used to determine multiple similar patients from multiple sample patients based on the knowledge map and the physiological information, disease information, drug information and multiple target endocrine hormones of the patients to be tested; an association analysis module, which is used to determine the association characteristics between multiple target endocrine hormones based on the dynamic monitoring data of the endocrine hormone levels of multiple similar patients; a hormone determination module, which is used to determine at least one target endocrine hormone from multiple target endocrine hormones based on the association characteristics between multiple target endocrine hormones. A key endocrine hormone; a frequency determination module, used to determine the dynamic monitoring frequency of a patient to be tested based on a knowledge graph, at least one key endocrine hormone and the physiological information, disease information and drug information of the patient to be tested; a hormone monitoring module, used to collect the dynamic content of each key endocrine hormone based on the dynamic monitoring frequency of the patient to be tested; the hormone monitoring module is also used to determine whether to adjust at least one key endocrine hormone and the dynamic monitoring frequency based on the dynamic content of each key endocrine hormone, and if so, collect the dynamic content of each key endocrine hormone based on the adjusted at least one key endocrine hormone and the dynamic monitoring frequency; a dynamic early warning module, used to generate early warning information based on the dynamic content of each key endocrine hormone.

[0015] Compared with the existing technology, the dynamic monitoring method and early warning system of endocrine hormone levels provided in this specification have at least the following beneficial effects:

[0016] 1. Dynamically determine key endocrine hormones and monitoring frequencies through knowledge graphs and individual patient information (physiology, disease, medication, etc.) to avoid a "one-size-fits-all" fixed frequency. For example, when diabetic patients experience large fluctuations in insulin levels, the monitoring frequency can be automatically increased, while when hormone levels are stable, the monitoring frequency can be reduced to reduce resource waste. Based on the hormone-related characteristics of similar patients, the key hormones that have the greatest impact on the disease (such as insulin and cortisol) are screened out, the monitoring of non-critical hormones is reduced, and the efficiency of data interpretation is improved. By dynamically adjusting the monitoring frequency, patients can avoid discomfort caused by excessive blood sampling and improve compliance;

[0017] 2. Through nonlinear association analysis, synergistic or antagonistic effects between hormones can be identified, and hormones can be divided into "first key hormones" (with a high probability of direct abnormality) and "second key hormones" (indirectly affected by association characteristics), achieving precise allocation of monitoring resources. For example, insulin (the first key hormone) can be monitored first, and if it is abnormal, cortisol (the second key hormone) can be monitored in conjunction. Through nonlinear association analysis and key hormone screening, unnecessary hormone monitoring can be reduced, reducing patient burden and medical costs.

[0018] 3. Determine whether there are abnormal key endocrine hormones based on the dynamic content of each key endocrine hormone, and then decide whether to adjust the key endocrine hormones and the dynamic monitoring frequency. This judgment method based on actual test data can timely detect changes in the patient's endocrine hormone levels, accurately grasp the timing of adjustments, and ensure that the monitoring plan always matches the patient's condition. When there are abnormal key endocrine hormones, for each abnormal hormone, the newly added key endocrine hormones are determined based on the relationship map, and the dynamic monitoring frequency is adjusted according to the adjusted abnormal probability. This process not only improves the types of key endocrine hormones monitored, but also optimizes the monitoring frequency, so that the monitoring can change dynamically with the development of the patient's condition, more comprehensively and deeply reflect the patient's endocrine status, and provide doctors with more accurate and timely information, which helps to formulate more effective treatment plans, improve treatment effects, and improve patient prognosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] This specification will be further described in the form of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting, and in these embodiments, like numbers represent like structures, wherein:

[0020] Figure 1 is a flow chart of a method for dynamic monitoring of endocrine hormone levels shown in one embodiment of the present application;

[0021] Figure 2 is a schematic diagram of a knowledge graph shown in an embodiment of the present application;

[0022] Figure 3 is a schematic diagram of a relationship diagram shown in an embodiment of the present application;

[0023] Figure 4 It is a module diagram of the endocrine hormone level dynamic early warning system shown in one embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly introduces the drawings required for describing the embodiments.

[0025] Figure 1 FIG. 1 is a flow chart of a method for dynamically monitoring endocrine hormone levels shown in an embodiment of the present application. Figure 1 As shown, the method for dynamic monitoring of endocrine hormone levels may include the following process.

[0026] Step 110: determine the physiological influencing factors, influencing diseases, and influencing drugs of multiple candidate endocrine hormones, and establish a knowledge graph.

[0027] Figure 2is a schematic diagram of a knowledge graph shown in an embodiment of the present application, such as Figure 2 As shown in the figure, the knowledge graph intuitively displays the relationship between endocrine hormones, physiological factors, diseases and drugs through the connection of nodes and edges.

[0028] 1. Node type:

[0029] Endocrine hormones: such as insulin, thyroid hormone (T3 / T4), cortisol, growth hormone, etc.

[0030] Physiological influencing factors: including age, gender, blood pressure, BMI (body mass index), genetic background, environmental factors (such as light, temperature), lifestyle (such as diet, exercise, sleep), etc.

[0031] Affected diseases: such as diabetes, hypothyroidism, Cushing's syndrome, growth hormone deficiency, etc.

[0032] Affecting drugs: such as insulin injection, levothyroxine, mifepristone, metformin, etc.

[0033] 2. Edge type:

[0034] Influence edge: represents the influence of physiological factors on endocrine hormones (such as aging leading to decreased secretion of growth hormone).

[0035] Cause edge: Indicates the impact of the disease on endocrine hormone levels (such as diabetes, which affects insulin, glucagon, and cortisol levels).

[0036] Therapeutic side: Indicates the effect of drugs on endocrine hormone levels (such as methimazole and propylthiouracil, which inhibit thyroid peroxidase activity and reduce thyroid hormone synthesis).

[0037] Preferably, the physiological influencing factors, influencing diseases and influencing drugs of multiple candidate endocrine hormones are determined, including:

[0038] For each candidate endocrine hormone,

[0039] Obtain physiological information, disease information, drug information and dynamic content of candidate endocrine hormones of multiple sample patients;

[0040] For each candidate physiological factor (such as age, BMI, blood pressure, blood sugar, etc.), based on the physiological information of multiple sample patients and the dynamic content of candidate endocrine hormones, the nonlinear correlation coefficient between the candidate physiological factor and the candidate endocrine hormone is calculated. For example, based on the physiological information of the sample patients, the characteristic value of the sample patient in the candidate physiological factor is determined, and the characteristic value and hormone level of the sample patient in the candidate physiological factor are standardized to the interval [0,1] to eliminate the dimension effect. For dynamic data (such as hormone secretion curves), interpolation or time window segmentation is used to ensure the consistency of time points. The Z-score or IQR method is used to eliminate extreme values. The mutual information (MI) between the candidate physiological factor and the candidate endocrine hormone is calculated as the nonlinear correlation coefficient between the candidate physiological factor and the candidate endocrine hormone;

[0041] Screening physiological influencing factors of the candidate endocrine hormones based on the nonlinear correlation coefficient between each candidate physiological factor and the candidate endocrine hormone, for example, taking the candidate physiological factors with a nonlinear correlation coefficient greater than a first threshold as the physiological influencing factors of the candidate endocrine hormones;

[0042] For each candidate disease, determining a diseased patient group and a non-diseased patient group corresponding to the candidate disease, and calculating a disease impact value of the candidate disease on the candidate endocrine hormone based on the diseased patient group and the non-diseased patient group corresponding to the candidate disease. Specifically, the non-diseased patient group includes sample patients who have not suffered from the candidate disease, and the diseased patient group includes sample patients who have suffered from the candidate disease. The proportion of sample patients with abnormal levels of the candidate endocrine hormone in the diseased patient group and the proportion of sample patients with abnormal levels of the candidate endocrine hormone in the non-diseased patient group can be calculated, and the difference between the two sample patient proportions is used as the disease impact value of the candidate disease on the candidate endocrine hormone;

[0043] Based on the disease impact value of each candidate disease on the candidate endocrine hormone, the affected diseases of the candidate endocrine hormone are screened. For example, the affected diseases whose disease impact value is greater than the disease impact value threshold can be used as the affected diseases of the candidate endocrine hormone;

[0044] For each candidate drug, determine the corresponding patient group taking the candidate drug and the corresponding patient group not taking the candidate drug. Based on the corresponding patient group with the disease and the patient group without the disease, calculate the drug effect value of the candidate drug on the candidate endocrine hormone. The method of calculating the drug effect value is similar to the method of calculating the disease effect value and will not be repeated here.

[0045] Based on the drug impact value of each candidate drug on the candidate endocrine hormone, the candidate endocrine hormone-influencing drugs are screened. Specifically, the method of screening the influencing drugs is similar to the method of screening the influencing diseases, which will not be repeated here.

[0046] Step 120 , determining multiple similar patients from multiple sample patients based on the knowledge graph and the physiological information, disease information, drug information, and multiple target endocrine hormones of the patient to be tested.

[0047] As a preferred step 120, specifically includes:

[0048] For each target endocrine hormone, based on the knowledge graph, the corresponding impact characteristics of the target endocrine hormone are extracted from the physiological information, disease information, and drug information of the patient to be tested. The physiological information covers the patient's age, gender, weight, height, blood pressure, heart rate and other basic physiological indicators. These indicators can affect the secretion, metabolism and effect of endocrine hormones. For example, aging may lead to decreased hormone secretion or decreased function; gender differences may affect the secretion pattern and intensity of hormone action. Disease information includes the various diseases suffered by the patient and their severity. Some diseases may directly affect the secretion or action of endocrine hormones. For example, patients with hyperthyroidism have elevated thyroid hormone levels in their bodies. Drug information includes the drugs currently being used by the patient and their dosage, usage, etc. Some drugs may affect the secretion, metabolism or effect of endocrine hormones, and abnormalities in endocrine hormones may also affect the efficacy or safety of the drugs. For example, glucocorticoids may suppress immune responses, affect glucose metabolism and electrolyte balance, and interact with endocrine hormones such as insulin. The impact features corresponding to the target endocrine hormones may include the characteristic values ​​of each physiological influencing factor, disease influencing factor, and drug influencing factor corresponding to the patient to be tested. Specifically, the physiological influencing factor, disease influencing factor, and drug influencing factor corresponding to the target endocrine hormone can be determined based on the knowledge graph, and then the impact features corresponding to the target endocrine hormone can be extracted from the physiological information, disease information, and drug information of the patient to be tested.

[0049] For each sample patient, based on the influence characteristics corresponding to each target endocrine hormone of the patient to be tested and the influence characteristics corresponding to each target endocrine hormone of the sample patient, the similarity between the patient to be tested and the sample patient is calculated. Specifically, based on the physiological influencing factors, influencing diseases and influencing drugs corresponding to the target endocrine hormones, the influence characteristics corresponding to the target endocrine hormones can be extracted from the physiological information, disease information and drug information of the sample patient, and then the similarity between the influence characteristics corresponding to the target endocrine hormones of the patient to be tested and the influence characteristics corresponding to the target endocrine hormones of the sample patient is calculated;

[0050] Based on the similarity between the patient to be detected and the sample patients, multiple similar patients are determined from the multiple sample patients. For example, sample patients whose similarity is greater than a similarity threshold can be regarded as similar patients.

[0051] By way of example only, the impact characteristics of the target endocrine hormone (e.g., insulin) of the patient to be tested may include:

[0052] 1. Characteristic values ​​of physiological influencing factors:

[0053] Age: 45 (middle-aged, possibly with increased insulin resistance)

[0054] Gender: Male (men may have lower insulin sensitivity than women)

[0055] Weight: 80 kg (overweight may increase insulin resistance)

[0056] Height: 175cm (used to calculate BMI and assess obesity level)

[0057] Blood pressure: 130 / 85 mmHg (high blood pressure may affect insulin sensitivity)

[0058] Heart rate: 75 beats / min (normal heart rate range, little effect on insulin)

[0059] 2. Impact on disease characteristic values:

[0060] Diabetes type: Type 2 diabetes (insulin resistance and insufficient insulin secretion)

[0061] Duration of disease: 5 years (long-term hyperglycemia may lead to decreased insulin secretion function)

[0062] Complications: diabetic nephropathy (impaired renal function may affect insulin metabolism)

[0063] 3. Impact on drug characteristic values:

[0064] Medications currently being used: Metformin (lowers blood sugar and improves insulin sensitivity)

[0065] Dosage: 500 mg twice daily (dose may affect insulin levels)

[0066] Specifically, the similarity between the patient to be tested and the sample patient can be calculated according to the following formula:

[0067]

[0068] in, is the similarity between the patient to be tested and the i-th sample patient, 、 and is the preset weight, 、 and greater than 0, , The mth target endocrine hormone corresponding to the i-th sample patient The characteristic values ​​of physiological factors, The mth target endocrine hormone corresponding to the patient to be tested The characteristic values ​​of physiological factors, The mth target endocrine hormone corresponding to the i-th sample patient The characteristic value of the mth physiological influencing factor is the same as the characteristic value of the mth target endocrine hormone corresponding to the patient to be tested. The cosine similarity of the eigenvalues ​​of the physiological influencing factors, is the total number of physiological influencing factors corresponding to the mth target endocrine hormone, The mth target endocrine hormone corresponding to the i-th sample patient The characteristic values ​​that affect the disease, The mth target endocrine hormone corresponding to the patient to be tested The characteristic values ​​that affect the disease, The mth target endocrine hormone corresponding to the i-th sample patient The characteristic value of the disease is the same as the characteristic value of the mth target endocrine hormone corresponding to the patient to be tested. The cosine similarity of the eigenvalues ​​that affect the disease, is the total number of diseases affected by the mth target endocrine hormone, The mth target endocrine hormone corresponding to the i-th sample patient The characteristic values ​​of the drug are affected. The mth target endocrine hormone corresponding to the patient to be tested The characteristic values ​​of the drug are affected. The mth target endocrine hormone corresponding to the i-th sample patient The characteristic value of the drug that affects the patient to be tested corresponds to the mth target endocrine hormone The cosine similarity of the characteristic values ​​of the drugs that affect the drug, is the total number of influencing drugs corresponding to the mth target endocrine hormone, is the total number of target endocrine hormones.

[0069] It can be understood that the above formula measures the similarity between feature vectors in each dimension through cosine similarity and directly accumulates them to achieve multi-dimensional measurement of the similarity between the patient to be tested and the sample patient. The cosine similarity of each physiological factor characteristic value is calculated one by one, and the similarity of all hormones is accumulated to directly measure the overall similarity between the patient to be tested and the sample patients in physiological factors. The same method is used to measure the overall similarity between the patient to be tested and the sample patients in diseases and drugs. Finally, the similarity of each dimension is weighted to obtain the comprehensive similarity between the patient to be tested and the sample patients.

[0070] Step 130 : determining correlation features between multiple target endocrine hormones based on the dynamic monitoring data of endocrine hormone levels of multiple similar patients.

[0071] Preferably, step 130 specifically includes:

[0072] For any two target endocrine hormones, a nonlinear correlation coefficient between the two target endocrine hormones is calculated based on the dynamic monitoring data of the endocrine hormone levels of multiple similar patients, wherein the association characteristics between the multiple target endocrine hormones include the nonlinear correlation coefficient between the two target endocrine hormones. For example, the mutual information between the two target endocrine hormones can be calculated based on the hormone levels of the two target endocrine hormones at different time points of multiple similar patients, as the nonlinear correlation coefficient between the two target endocrine hormones.

[0073] Step 140 : determining at least one key endocrine hormone from the multiple target endocrine hormones based on the correlation characteristics between the multiple target endocrine hormones.

[0074] Preferably, step 140 specifically includes:

[0075] For each target endocrine hormone, the abnormal probability of the target endocrine hormone is determined based on the dynamic monitoring data of the endocrine hormone levels of multiple similar patients. For example, the proportion of abnormal target endocrine hormone levels in multiple similar patients can be calculated as the abnormal probability of the target endocrine hormone;

[0076] Determine at least one first key endocrine hormone based on the abnormal probability of each target endocrine hormone. For example, a target endocrine hormone whose abnormal probability is greater than an abnormal probability threshold may be used as the first key endocrine hormone.

[0077] At least one second key endocrine hormone is determined based on the correlation features between the first key endocrine hormone and multiple target endocrine hormones, wherein the at least one key endocrine hormone includes each first key endocrine hormone and each second key endocrine hormone.

[0078] Preferably, determining at least one second key endocrine hormone based on the correlation characteristics between the first key endocrine hormone and multiple target endocrine hormones includes:

[0079] For each target endocrine hormone, determining a correlated target endocrine hormone of the target endocrine hormone based on a nonlinear correlation coefficient between any two target endocrine hormones, for example, two target endocrine hormones whose nonlinear correlation coefficients are greater than a nonlinear correlation coefficient threshold are mutually correlated target endocrine hormones;

[0080] Establish a relationship map based on the related target endocrine hormones of each target endocrine hormone;

[0081] For each node in the relationship graph, the key value of the node is calculated based on the abnormal probability of the first key endocrine hormone and the nonlinear correlation coefficient of any two target endocrine hormones;

[0082] Determine a candidate key endocrine hormone based on the key value of each node. For example, the target endocrine hormone corresponding to the node whose key value is greater than the key value threshold can be used as the candidate key endocrine hormone;

[0083] The candidate key endocrine hormones and the first key endocrine hormones are deduplicated to determine at least one second key endocrine hormone.

[0084] Specifically, Figure 3 is a schematic diagram of a relationship diagram shown in an embodiment of the present application, such as Figure 3 As shown, in the relationship graph, the nodes of two target endocrine hormones that are related to each other are connected by an edge. The length of the edge can represent the size of the nonlinear correlation coefficient. The shorter the edge, the larger the nonlinear correlation coefficient.

[0085] For example, the key value of a node can be calculated according to the following formula:

[0086]

[0087] in, is the key value of the i-th node, is the total number of the first key endocrine hormones in the relevant target endocrine hormones of the i-th node, is the nonlinear correlation coefficient between the target endocrine hormone corresponding to the i-th node and the k-th first key endocrine hormone among the related target endocrine hormones of the i-th node, is the abnormal probability of the kth first key endocrine hormone among the relevant target endocrine hormones of the i-th node.

[0088] It can be understood that through the nonlinear correlation coefficient and abnormal probability The weighted summation of the endocrine hormone nodes is used to quantify the importance of each endocrine hormone node in the global network, and a few key endocrine hormones are extracted from a large number of target endocrine hormones to reduce the complexity of subsequent analysis. This reduces unnecessary endocrine hormone level analysis while ensuring the accuracy and comprehensiveness of dynamic monitoring of endocrine hormone levels.

[0089] Step 150 , determining the dynamic monitoring frequency of the patient to be tested based on the knowledge graph, at least one key endocrine hormone, and the physiological information, disease information, and drug information of the patient to be tested.

[0090] Preferably, step 150 specifically includes:

[0091] Based on the probability of abnormality of each key endocrine hormone, the dynamic monitoring frequency of the patient to be tested is determined.

[0092] Specifically, the dynamic monitoring frequency of the patient to be tested can be determined according to the following formula:

[0093]

[0094] in, is the dynamic monitoring frequency of the patient to be tested, is the fundamental frequency, is the abnormal probability of the qth key endocrine hormone, is the total number of key endocrine hormones.

[0095] As can be understood, the above formula, based on the mean probability of abnormality for each key endocrine hormone, determines the frequency of dynamic monitoring. The greater the mean probability of abnormality for each key endocrine hormone, the higher the dynamic monitoring frequency for the patient to be tested, thereby reducing the monitoring burden on low-risk patients and concentrating resources on high-risk patients. When high-risk hormones are abnormal, the monitoring interval can be shortened (for example, from once daily to once hourly) to capture early changes.

[0096] Step 160 : collecting the dynamic content of each key endocrine hormone according to the dynamic monitoring frequency of the patient to be tested.

[0097] Step 170, based on the dynamic content of each key endocrine hormone, determine whether to adjust at least one key endocrine hormone and the dynamic monitoring frequency. If so, collect the dynamic content of each key endocrine hormone based on the adjusted at least one key endocrine hormone and the dynamic monitoring frequency.

[0098] Preferably, judging whether to adjust at least one key endocrine hormone and the dynamic monitoring frequency based on the dynamic content of each key endocrine hormone includes:

[0099] Based on the dynamic content of each key endocrine hormone, determine whether there is at least one abnormal key endocrine hormone. For example, if the dynamic content of a key endocrine hormone exceeds the reference range, it is an abnormal key endocrine hormone;

[0100] If so, adjust at least one key endocrine hormone and the frequency of dynamic monitoring;

[0101] If not, do not adjust at least one key endocrine hormone and the frequency of dynamic monitoring.

[0102] Preferably, adjusting at least one key endocrine hormone and the frequency of dynamic monitoring include:

[0103] For each abnormal key endocrine hormone, a newly added key endocrine hormone is determined based on the relationship map. For example, all target endocrine hormones related to the abnormal key endocrine hormone that are not abnormal key endocrine hormones can be used as newly added key endocrine hormones.

[0104] Based on the adjusted probability of abnormality of each key endocrine hormone, the dynamic monitoring frequency of the patients to be tested is adjusted.

[0105] For example, the following formula can be used:

[0106]

[0107] in, is the adjusted dynamic monitoring frequency of the patients to be tested, The abnormal value of the abnormal key endocrine hormone can be determined based on the difference between the dynamic content of the abnormal key endocrine hormone and the reference range. It is the total amount of abnormal key endocrine hormones.

[0108] It can be understood that the above formula can dynamically adjust the dynamic monitoring frequency of patients to be tested through abnormal values ​​of abnormal key endocrine hormones, perform high-frequency testing on abnormal patients, and improve the real-time monitoring of abnormal hormone levels.

[0109] Figure 4 This is a module diagram of a dynamic early warning system for endocrine hormone levels shown in an embodiment of the present application. Figure 4 As shown, the endocrine hormone level dynamic early warning system can include a map building module, a similarity search module, an association analysis module, a hormone determination module, a frequency determination module, a hormone monitoring module and a dynamic early warning module.

[0110] A graph building module is used to determine the physiological influencing factors, diseases, and drugs of multiple candidate endocrine hormones and to build a knowledge graph;

[0111] The similarity search module is used to identify multiple similar patients from multiple sample patients based on the knowledge graph and the physiological information, disease information, drug information and multiple target endocrine hormones of the patients to be tested;

[0112] The association analysis module is used to determine the association characteristics between multiple target endocrine hormones based on the dynamic monitoring data of endocrine hormone levels of multiple similar patients;

[0113] a hormone determination module, configured to determine at least one key endocrine hormone from a plurality of target endocrine hormones based on correlation features among the plurality of target endocrine hormones;

[0114] The frequency determination module is configured to determine a dynamic monitoring frequency of the patient to be detected according to the knowledge graph, the at least one key endocrine hormone, and physiological information, disease information, and drug information of the patient to be detected.

[0115] The hormone monitoring module is configured to collect the dynamic content of each key endocrine hormone according to the dynamic monitoring frequency of the patient to be detected.

[0116] The hormone monitoring module is further configured to determine whether to adjust the at least one key endocrine hormone and the dynamic monitoring frequency according to the dynamic content of each key endocrine hormone, and if so, collect the dynamic content of each key endocrine hormone according to the adjusted at least one key endocrine hormone and the dynamic monitoring frequency.

[0117] The dynamic early warning module is configured to generate early warning information according to the dynamic content of each key endocrine hormone.

[0118] The endocrine hormone level dynamic early warning system can be used to execute the endocrine hormone level dynamic monitoring method described above, and thus will not be described here again.

[0119] Finally, it should be understood that the embodiments described in the specification are only used to illustrate the principles of the embodiments of the specification. Other variations can also belong to the scope of the specification. Therefore, as an example but not limitation, alternative configurations of the embodiments of the specification can be considered as consistent with the teachings of the specification. Accordingly, the embodiments of the specification are not limited to the embodiments explicitly introduced and described in the specification.

Claims

1. A method for dynamic monitoring of endocrine hormone levels, characterized in that: include: Identify the physiological influencing factors, diseases affected, and drugs affected by multiple candidate endocrine hormones, and establish a knowledge graph; Identify multiple similar patients from multiple sample patients based on the knowledge graph and the physiological information, disease information, drug information, and multiple target endocrine hormones of the patients to be tested; Determine the correlation characteristics between multiple target endocrine hormones based on dynamic monitoring data of endocrine hormone levels of multiple similar patients; identifying at least one key endocrine hormone from the multiple target endocrine hormones based on correlation characteristics among the multiple target endocrine hormones; Determine the dynamic monitoring frequency of the patient to be tested based on the knowledge graph, at least one key endocrine hormone, and the patient's physiological information, disease information, and drug information; Collect the dynamic content of each key endocrine hormone according to the dynamic monitoring frequency of the patient to be tested; Based on the dynamic content of each key endocrine hormone, determine whether to adjust at least one key endocrine hormone and the dynamic monitoring frequency. If so, collect the dynamic content of each key endocrine hormone based on the adjusted at least one key endocrine hormone and the dynamic monitoring frequency.

2. The method for dynamic monitoring of endocrine hormone levels according to claim 1, characterized in that: Identify the physiological influencers, diseases, and drugs of multiple candidate endocrine hormones, including: For each candidate endocrine hormone, Obtain physiological information, disease information, drug information and dynamic content of candidate endocrine hormones of multiple sample patients; For each candidate physiological factor, based on the physiological information of multiple sample patients and the dynamic content of the candidate endocrine hormone, the nonlinear correlation coefficient between the candidate physiological factor and the candidate endocrine hormone is calculated; Screening physiological influencing factors of candidate endocrine hormones based on the nonlinear correlation coefficient between each candidate physiological factor and the candidate endocrine hormone; For each candidate disease, determining a diseased patient group and a non-diseased patient group corresponding to the candidate disease, and calculating a disease impact value of the candidate disease on the candidate endocrine hormone based on the diseased patient group and the non-diseased patient group corresponding to the candidate disease; Screening the diseases affected by the candidate endocrine hormones based on the disease impact value of each candidate disease on the candidate endocrine hormones; For each candidate drug, determining a corresponding group of patients taking the candidate drug and a corresponding group of patients not taking the candidate drug, and calculating a drug effect value of the candidate drug on the candidate endocrine hormone based on the corresponding groups of patients with and without the disease; Based on the drug impact value of each candidate drug on the candidate endocrine hormone, the candidate endocrine hormone-influencing drugs are screened.

3. The method for dynamic monitoring of endocrine hormone levels according to claim 1, characterized in that: Based on the knowledge graph and the physiological information, disease information, drug information, and multiple target endocrine hormones of the patient to be tested, multiple similar patients are identified from multiple sample patients, including: For each target endocrine hormone, based on the knowledge graph, the corresponding impact characteristics of the target endocrine hormone are extracted from the physiological information, disease information, and drug information of the patient to be tested; For each sample patient, the similarity between the patient to be tested and the sample patient is calculated based on the impact characteristics corresponding to each target endocrine hormone of the patient to be tested and the impact characteristics corresponding to each target endocrine hormone of the sample patient; Based on the similarity between the patient to be detected and the sample patients, a plurality of similar patients are determined from the plurality of sample patients.

4. The method for dynamic monitoring of endocrine hormone levels according to any one of claims 1 to 3, characterized in that: Based on the dynamic monitoring data of endocrine hormone levels of multiple similar patients, the correlation characteristics between multiple target endocrine hormones are determined, including: For any two target endocrine hormones, the nonlinear correlation coefficient of the two target endocrine hormones is calculated based on the dynamic monitoring data of endocrine hormone levels of multiple similar patients, wherein the association characteristics between the multiple target endocrine hormones include the nonlinear correlation coefficient of the two target endocrine hormones.

5. The method for dynamic monitoring of endocrine hormone levels according to claim 4, characterized in that: According to the correlation characteristics among the multiple target endocrine hormones, at least one key endocrine hormone is identified from the multiple target endocrine hormones, including: For each target endocrine hormone, the abnormal probability of the target endocrine hormone is determined based on the dynamic monitoring data of endocrine hormone levels of multiple similar patients; determining at least one first key endocrine hormone based on the probability of abnormality of each target endocrine hormone; At least one second key endocrine hormone is determined based on the correlation features between the first key endocrine hormone and multiple target endocrine hormones, wherein the at least one key endocrine hormone includes each first key endocrine hormone and each second key endocrine hormone.

6. The method for dynamic monitoring of endocrine hormone levels according to claim 5, characterized in that: At least one second key endocrine hormone is determined based on the correlation characteristics between the first key endocrine hormone and multiple target endocrine hormones, including: For each target endocrine hormone, determining the correlated target endocrine hormone of the target endocrine hormone based on the nonlinear correlation coefficient between any two target endocrine hormones; Establish a relationship map based on the related target endocrine hormones of each target endocrine hormone; For each node in the relationship graph, the key value of the node is calculated based on the abnormal probability of the first key endocrine hormone and the nonlinear correlation coefficient of any two target endocrine hormones; Based on the key value of each node, determine the candidate key endocrine hormones; The candidate key endocrine hormones and the first key endocrine hormones are deduplicated to determine at least one second key endocrine hormone.

7. The method for dynamic monitoring of endocrine hormone levels according to claim 5, characterized in that: Determine the dynamic monitoring frequency of the patient to be tested based on the knowledge graph, at least one key endocrine hormone, and the patient's physiological information, disease information, and drug information, including: Based on the probability of abnormality of each key endocrine hormone, the dynamic monitoring frequency of the patient to be tested is determined.

8. The method for dynamic monitoring of endocrine hormone levels according to claim 6, characterized in that: Based on the dynamic content of each key endocrine hormone, determine whether to adjust at least one key endocrine hormone and the dynamic monitoring frequency, including: Based on the dynamic content of each key endocrine hormone, determine whether there is at least one abnormal key endocrine hormone; If so, adjust at least one key endocrine hormone and the frequency of dynamic monitoring; If not, do not adjust at least one key endocrine hormone and the frequency of dynamic monitoring.

9. The method for dynamic monitoring of endocrine hormone levels according to claim 8, characterized in that: Adjust at least one key endocrine hormone and the frequency of dynamic monitoring, including: For each abnormal key endocrine hormone, newly added key endocrine hormones are identified based on the relationship map; Based on the adjusted probability of abnormality of each key endocrine hormone, the dynamic monitoring frequency of the patients to be tested is adjusted.

10. Dynamic early warning system for endocrine hormone levels, characterized by: A method for dynamically monitoring endocrine hormone levels according to any one of claims 1 to 9, comprising: A graph building module is used to determine the physiological influencing factors, diseases, and drugs of multiple candidate endocrine hormones and to build a knowledge graph; The similarity search module is used to identify multiple similar patients from multiple sample patients based on the knowledge graph and the physiological information, disease information, drug information and multiple target endocrine hormones of the patients to be tested; The association analysis module is used to determine the association characteristics between multiple target endocrine hormones based on the dynamic monitoring data of endocrine hormone levels of multiple similar patients; a hormone determination module, configured to determine at least one key endocrine hormone from a plurality of target endocrine hormones based on correlation features among the plurality of target endocrine hormones; A frequency determination module is used to determine the dynamic monitoring frequency of the patient to be tested based on the knowledge graph, at least one key endocrine hormone, and the physiological information, disease information, and drug information of the patient to be tested; The hormone monitoring module is used to collect the dynamic content of each key endocrine hormone according to the dynamic monitoring frequency of the patient to be tested; The hormone monitoring module is further configured to determine, based on the dynamic content of each key endocrine hormone, whether to adjust at least one key endocrine hormone and the dynamic monitoring frequency; if so, collect the dynamic content of each key endocrine hormone based on the adjusted at least one key endocrine hormone and the dynamic monitoring frequency; The dynamic early warning module is used to generate early warning information based on the dynamic content of each key endocrine hormone.

Images