Secondary hypertension multi-factor intelligent analysis system based on expert knowledge base

Abstract

This invention discloses an intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base, belonging to the field of big data analysis technology. The invention generates etiological nodes and abnormal state factors based on historical hypertension diagnoses, establishes an expert diagnostic knowledge graph by matching these nodes and factors, generates patient diagnostic factors based on real-time physical status data, inputs these factors into the expert diagnostic knowledge graph, obtains the cumulative weight values ​​of each etiological node through the matching results of patient diagnostic factors and abnormal state factors, determines the relationship between the cumulative weight values ​​and the judgment threshold, outputs an etiological analysis report based on the judgment results, corrects the etiological analysis report based on the patient's historical medical history, and finally generates a unified disease analysis report and medication recommendation report based on the risk probability of potential causes.

Description

Technical Field

[0001] This invention relates to the field of big data analytics, specifically to an intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base. Background Technology

[0002] Secondary hypertension has a complex etiology, encompassing dozens of diseases including renal, endocrine, vascular, genetic, metabolic, psychological, and drug-related factors. Its diagnosis is essentially a multi-step, multidisciplinary process of differentiation and exclusion. Current diagnostic techniques suffer from systemic and structural deficiencies and limitations. Diagnostic logic relies on human experience, resulting in poor standardization and accessibility. Existing diagnostic pathways heavily depend on the individual knowledge and experience of clinicians. In primary care or non-specialized medical institutions, the lack of hypertension subspecialty experts can easily lead to insufficient vigilance or misdiagnosis of secondary hypertension, causing a large number of cases to be simply classified as "primary hypertension," which violates the basic principle emphasized in the guidelines of "ruling out secondary hypertension first, then diagnosing primary hypertension." The diagnostic process is lengthy and fragmented, leading to inefficient patient management. The diagnosis of secondary hypertension is not a one-time test, but a multi-stage process involving screening, confirmation, and classification / localization. Under the current model, each stage relies on the doctor's manual memory and tracking, making it easy for patients to drop out during referrals, appointment scheduling, and waiting for results, resulting in diagnostic interruptions or delays. For complex cases involving multiple potential causes, the lack of systematic tools to help manage parallel examination clues and priorities makes the diagnostic and treatment process chaotic and inefficient. Therefore, this paper proposes an intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base. Summary of the Invention

[0003] The purpose of this invention is to provide an intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base, in order to address the shortcomings of the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: The intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base includes an expert knowledge base module, an intelligent analysis engine module, and an output and recommendation module. The expert knowledge base module is used to obtain historical diagnosis cases of hypertension, and then generate etiology nodes and abnormal state factors based on the historical diagnosis cases of hypertension. The etiology nodes and abnormal state factors are matched with each other to establish an expert diagnosis knowledge graph. The intelligent analysis engine module is used to acquire real-time physical status data, generate patient diagnostic factors based on the real-time physical status data, input the patient diagnostic factors into the expert diagnostic knowledge graph, obtain the cumulative weight value of each etiology node through the matching results of patient diagnostic factors and abnormal status factors, determine the relationship between the cumulative weight value and the judgment threshold, and output etiology analysis report based on the judgment result. The output and recommendation module is used to obtain the patient's historical medical records, revise the etiology analysis report based on the patient's historical medical records, and then generate a unified disease analysis report and medication recommendation report based on the risk probability of potential etiologies.

[0005] Furthermore, the aforementioned historical diagnosis of hypertension consists of the patient's name, identification number, range of multiple abnormal physical data related to the etiology of secondary hypertension, disease progression history, and information on the types and dosages of medications used at each stage of the disease.

[0006] Furthermore, the process of generating etiological nodes and abnormal state factors based on historical hypertension diagnoses includes: Several etiological nodes are established based on the etiological association information of secondary hypertension recorded in the historical diagnosis cases of hypertension, and each etiological node is assigned a unique number. The etiology node corresponds to a specific stage of a potential etiology. That is, a potential etiology may have multiple corresponding etiology nodes depending on its disease progression, and each etiology node is bound to a drug use node. The drug use node records the types and dosages of drugs routinely used in clinical practice at the corresponding stage of the disease. The drug types and dosages are directly extracted from the types and dosages of drugs used at each stage of the disease recorded in the historical diagnosis cases of hypertension. Furthermore, abnormal condition factors are established based on the range of abnormal physical data recorded in the historical diagnosis cases of hypertension. Each abnormal state factor corresponds to a range of physical state data. The numerical ranges or data types included in different abnormal state factors are not completely the same, and the same physical data can be divided into multiple non-overlapping abnormal state factors according to the degree of abnormality.

[0007] Furthermore, the process of matching causal nodes and abnormal state factors to establish an expert diagnostic knowledge graph includes: Each etiology node is associated and matched with abnormal state factors. The historical diagnosis cases of hypertension are traversed. If the disease stage corresponding to the etiology of secondary hypertension recorded in the historical diagnosis cases of hypertension exists at the same time as a certain abnormal body data range, a connection line is established between the etiology node and the abnormal state factor corresponding to the abnormal body data range, and the association weight value is set on the connection line. The correlation weight value is obtained by statistically analyzing the frequency of simultaneous occurrence of each etiological node and abnormal state factor in historical hypertension diagnoses, and dividing it by the total frequency of cases in which the etiological node occurs. Once all causal nodes and abnormal state factors have been matched and associated, an expert diagnostic knowledge graph is generated. The expert diagnostic knowledge graph is presented in a network structure, containing two main levels: the causal layer and the factor layer. The causal layer includes all causal nodes and their associated drug use nodes, while the factor layer includes all abnormal state factors.

[0008] Furthermore, the process of generating patient diagnostic factors based on real-time physical status data and inputting these factors into the expert diagnostic knowledge graph includes: Several patient diagnostic factors are generated based on the actual measurement values ​​of various physical data recorded in real-time physical status data. All patient diagnostic factors are mapped to the factor layer of the expert diagnostic knowledge graph. That is, the data range of the patient diagnostic factors is matched with the physical status data range corresponding to the abnormal state factors. If the actual measurement value of any physical data in the patient diagnostic factors falls completely within the numerical range specified by a certain abnormal state factor, then the abnormal state factor is marked as active. If the actual measurement value does not fall within the numerical range of any abnormal state factor, then no corresponding abnormal state factor is activated.

[0009] Furthermore, the process of obtaining the cumulative weight values ​​of each etiological node through the matching results of patient diagnostic factors and abnormal state factors includes: After all patient diagnostic factors and abnormal state factors are matched, cumulative weight values ​​are calculated based on the connection relationships between activated abnormal state factors and etiological nodes in the expert diagnostic knowledge graph. For each causal node, iterate through all the abnormal state factors connected to it. If an abnormal state factor belongs to the set of active abnormal state factors, then add the associated weight value on the connection line to the cumulative weight value of the causal node. If an abnormal state factor is not in an active state, then it is not included in the current cumulative weight value of the causal node.

[0010] Furthermore, the process of determining the relationship between the cumulative weight value and the judgment threshold, and outputting a pathogenesis analysis report based on the judgment result, includes: After the above calculations, each causal node obtains its own cumulative weight value. A first judgment threshold and a second judgment threshold are preset, where the first judgment threshold is less than the second judgment threshold. The cumulative weight value of each causal node is then compared with the judgment threshold respectively. If the cumulative weight value of the causal node is greater than or equal to the second judgment threshold, it is determined that the potential causal cause and its disease stage corresponding to the causal node are highly likely, and the potential causal cause is marked as a high-risk causal cause. If the cumulative weight value of the causal node is between the first judgment threshold and the second judgment threshold, then the potential causal node is judged to have a medium probability and is marked as a medium-risk causal node. If the cumulative weight value of the causal node is less than the first judgment threshold, it is determined that the current probability of the potential causal cause is low and it is marked as a low-risk causal cause. Based on the judgment results, a pathogenesis analysis report is generated. The pathogenesis analysis report records the risk probability of each potential pathogenesis, the corresponding disease stage, and the corresponding recommended medication information. The risk probability is presented as a percentage after being normalized according to the cumulative weight value. The disease stage information is extracted from the definition of the pathogenesis node itself, and the medication information is extracted from the drug type and dosage information recorded in the drug use node bound to the pathogenesis node.

[0011] Furthermore, the process of generating a unified disease analysis report and medication recommendation report based on the risk probability of potential causes includes: The patient's medical history includes the patient's medical history, description of physical signs, family history, past medication history, drug allergy history, and information on complications; If a patient's medical history contains a record of allergy to any of the ingredients in the recommended medication, then that medication will be removed from the recommended medication information and marked as contraindicated for allergies. If the patient's medical history records concomitant medications that interact with the recommended medication, the recommended medication will be replaced with a non-interacting alternative medication of the same class, or medication warning information will be added. If the patient's liver and kidney function indicators in their medical history suggest that the patient has impaired liver and kidney function, the dosage of the recommended medication should be reduced accordingly based on the standard adjustment guidelines. If the patient's medical history includes a record of pregnancy or breastfeeding, the corresponding medication contraindicated during breastfeeding should be removed and replaced with an alternative that is safe and available according to the safety classification. After completing the above corrections, a revised etiology and medication information table is generated. Each potential etiology in the table includes the medication name, single dose, frequency of administration, route of administration, and necessary remarks. The risk probability and disease stage information of each potential cause in the etiology analysis report are integrated with the revised etiology and medication information table to generate a unified disease analysis report and medication recommendation report. The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention establishes etiology nodes containing multiple potential causes and their different disease stages, as well as abnormal state factor nodes covering a range of abnormal physical state data. Etiology nodes and abnormal state factors are connected through association weight values, reflecting the strength of the directional influence of each abnormal physical state manifestation on a specific cause and disease stage. Simultaneously, the knowledge graph network structure can intuitively express the complex relationships of one cause with multiple effects and multiple causes with one effect, to some extent solving the problem that traditional single-cause screening models cannot handle situations involving intertwined multiple factors, and that diagnostic logic relies entirely on human experience.

[0012] 2. This invention achieves a closed loop from etiological inference to treatment recommendations by adaptively modifying the conventional medication regimen corresponding to the preliminary etiological analysis in combination with the patient's historical medical records, and finally generating a unified disease analysis report and medication recommendation report. Moreover, the recommended regimen fully considers the individual circumstances of the patient, improving the safety and pertinence of the treatment. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0014] Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Please see Figure 1 As shown, the intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base includes an expert knowledge base module, an intelligent analysis engine module, and an output and recommendation module. The expert knowledge base module is used to obtain historical diagnosis cases of hypertension, and then generate etiology nodes and abnormal state factors based on the historical diagnosis cases of hypertension. The etiology nodes and abnormal state factors are matched with each other to establish an expert diagnosis knowledge graph. The intelligent analysis engine module is used to acquire real-time physical status data, generate patient diagnostic factors based on the real-time physical status data, input the patient diagnostic factors into the expert diagnostic knowledge graph, obtain the cumulative weight value of each etiology node through the matching results of patient diagnostic factors and abnormal status factors, determine the relationship between the cumulative weight value and the judgment threshold, and output etiology analysis report based on the judgment result. The output and recommendation module is used to obtain the patient's historical medical records, correct the medication information in the etiology analysis report based on the patient's historical medical records, and then generate a unified disease analysis report and medication recommendation report based on the risk probability of potential etiologies.

[0017] Furthermore, the working principle of the present invention will be illustrated below through embodiments: The expert knowledge base module extracts all historical hypertension diagnosis cases from the hypertension case database. These historical hypertension diagnosis cases consist of the patient's name, identification number, range of multiple abnormal physical data related to the etiology of secondary hypertension, disease progression history, and information on the types and dosages of medications used at each stage of the disease. The range of abnormal physical data includes, but is not limited to, blood pressure range, aldosterone-renin ratio range, blood potassium level range, plasma renin activity range, and cortisol rhythm range. Several etiological nodes are established based on the etiological association information of secondary hypertension recorded in the historical diagnosis cases of hypertension, and each etiological node is assigned a unique number. It should be noted that the causal nodes correspond to a specific stage of a potential cause. That is, a potential cause may have multiple corresponding causal nodes depending on its disease progression. For example, the potential cause "primary aldosteronism" may correspond to causal nodes at different disease stages such as the early stage of normal blood potassium, the intermediate stage of hypokalemia, and the late stage of target organ damage. "Cushing's syndrome" may correspond to multiple causal nodes such as the subclinical Cushing's stage and the typical stage of hypercortisolism. Furthermore, each causal node is associated with a drug use node. The drug use node records the types and dosages of drugs routinely used in clinical practice at the corresponding stage of the disease. The drug types and dosages are directly extracted from the types and dosages of drugs used at each stage of the disease recorded in the historical diagnosis cases of hypertension. Furthermore, abnormal condition factors are established based on the range of abnormal physical data recorded in the historical diagnosis cases of hypertension. Each abnormal condition corresponds to a range of body condition data. For example, abnormal conditions are associated with a systolic blood pressure range of 160-179 mmHg, diastolic blood pressure range of 100-109 mmHg, aldosterone-renin ratio greater than 30, and blood potassium levels below 3.5 mmol / L. The numerical ranges or data types included in different abnormal state factors are not completely the same, and the same physical data can be divided into multiple non-overlapping abnormal state factors according to the degree of abnormality. For example, blood potassium content can be divided into abnormal state factors such as "low potassium level 1", "low potassium level 2", and "low potassium level 3". Each etiology node is associated and matched with abnormal state factors. The historical diagnosis cases of hypertension are traversed. If the disease stage corresponding to the etiology of secondary hypertension recorded in the historical diagnosis cases of hypertension exists at the same time as a certain abnormal body data range, a connection line is established between the etiology node and the abnormal state factor corresponding to the abnormal body data range, and the association weight value is set on the connection line. The correlation weight value is obtained by statistically analyzing the frequency of simultaneous occurrence of each etiological node and abnormal state factor in historical hypertension diagnoses, and dividing it by the total frequency of cases in which the etiological node occurs. The correlation weight value is a floating-point number between 0 and 1. The higher the correlation weight value, the stronger the correlation between the abnormal state factor and the causal node. The expert diagnostic knowledge graph contains multiple potential causes connected to the same abnormal state factor. For example, the abnormal state factor "aldosterone-renin ratio greater than 30" can be simultaneously connected to the cause node "mid-stage of primary aldosteronism" and the cause node "progressive stage of renovascular hypertension". At the same time, there are also cases where a potential cause is connected to multiple abnormal state factors. For example, the cause node "typical stage of Cushing's syndrome" can be simultaneously connected to multiple abnormal state factors such as "disappearance of cortisol rhythm", "signs of central obesity", and "persistently elevated systolic blood pressure". Once all causal nodes and abnormal state factors have been matched and associated, an expert diagnostic knowledge graph is generated. The expert diagnostic knowledge graph is presented in a network structure, containing two main levels: the causal layer and the factor layer. The causal layer includes all causal nodes and their associated drug use nodes, while the factor layer includes all abnormal state factors. At the same time, the expert knowledge base module sends the expert diagnostic knowledge graph to the intelligent analysis engine module.

[0018] It should be noted that whenever there is a new historical diagnosis of hypertension, the expert knowledge base module automatically triggers the update process. First, it determines whether there are any new etiological nodes or abnormal state factors. If the corresponding etiological nodes or abnormal state factors already exist in the expert diagnostic knowledge graph, then only the associated weight values ​​of the affected connection lines are recalculated. If there are any newly added causal nodes or new abnormal state factors that are not included, the construction of the newly added causal nodes or new abnormal state factors shall be completed first, and then the corresponding connection lines and associated weight values ​​shall be established to complete the iterative update of the expert diagnostic knowledge graph.

[0019] Furthermore, the intelligent analysis engine module collects real-time physical status data of patients through the hospital information system interface, bedside monitoring equipment data interface and mobile medical terminal. The collected data items are consistent with the types of physical status data corresponding to the abnormal status factors, including but not limited to blood pressure value, aldosterone-renin ratio, blood potassium content, plasma renin activity, cortisol rhythm data, adrenal imaging characteristic parameters, etc. Several patient diagnostic factors are generated based on the actual measurement values ​​of various body data recorded in real-time body status data. Each patient diagnostic factor consists of the corresponding body data item name and the data range to which its measurement value belongs. For example, the patient diagnostic factors are "systolic blood pressure 172 mmHg (belonging to the range of 160-179 mmHg)" and "blood potassium content 3.2 mmol / L (belonging to the second level of hypokalemia)". All patient diagnostic factors are mapped to the factor layer of the expert diagnostic knowledge graph. This involves matching the data range of patient diagnostic factors with the data range of physical state corresponding to abnormal state factors. The matching process includes: if the actual measurement value of any physical data in the patient diagnostic factors falls completely within the numerical range specified by an abnormal state factor, then the abnormal state factor is marked as active; if the actual measurement value does not fall within the numerical range of any abnormal state factor, then no corresponding abnormal state factor is activated. After all patient diagnostic factors and abnormal state factors have been matched, cumulative weight calculations are performed based on the connection relationships between activated abnormal state factors and etiological nodes in the expert diagnostic knowledge graph; the calculation process includes: For each causal node, iterate through all the abnormal state factors connected to it. If an abnormal state factor belongs to the set of active abnormal state factors, then add the associated weight value on the connection line to the cumulative weight value of the causal node. If an abnormal state factor is not in an active state, then it is not included in the current cumulative weight value of the causal node. After the above calculations, each causal node obtains its own cumulative weight value. A first judgment threshold and a second judgment threshold are preset, where the first judgment threshold is less than the second judgment threshold. The cumulative weight value of each causal node is then compared with the judgment threshold respectively. If the cumulative weight value of the causal node is greater than or equal to the second judgment threshold, it is determined that the potential causal cause and its disease stage corresponding to the causal node are highly likely, and the potential causal cause is marked as a high-risk causal cause. If the cumulative weight value of the causal node is between the first judgment threshold and the second judgment threshold (inclusive of the first judgment threshold), then the potential causal node is judged to have a medium probability and is marked as a medium-risk causal node. If the cumulative weight value of the causal node is less than the first judgment threshold, it is determined that the current probability of the potential causal cause is low and it is marked as a low-risk causal cause. Based on the above judgment results, a pathogenesis analysis report is generated. The pathogenesis analysis report records the risk probability of each potential pathogenesis, the corresponding disease stage, and the corresponding recommended medication information. The risk probability is normalized according to the cumulative weight value and presented in the form of a percentage. The disease stage information is extracted from the definition of the pathogenesis node itself, and the medication information is extracted from the drug type and dosage information recorded in the drug use node bound to the pathogenesis node. The content structure of the etiology analysis report is as follows: high-risk etiologies (arranged in descending order of cumulative weight value), medium-risk etiologies (arranged in descending order of cumulative weight value), and low-risk etiologies (arranged in descending order of cumulative weight value); for each etiology, a description of the current stage of the disease and preliminary recommended medication are listed; The intelligent analysis engine module sends the generated etiology analysis report, along with corresponding real-time physical status data and patient diagnostic factor information, to the output and recommendation module.

[0020] Furthermore, the output and recommendation module establishes a data interface with the hospital's electronic medical record system, which can retrieve the corresponding patient's historical medical records based on the patient's identification number. The patient's medical history includes the patient's medical history, description of physical signs, family history, past medication history, drug allergy history, and information on complications; The output and recommendation module then reviews and corrects the preliminary recommended medication information listed in the etiology analysis report item by item based on the patient's historical medical records. The process includes: If a patient's medical history contains a record of allergy to any of the ingredients in the recommended medication, then that medication will be removed from the recommended medication information and marked as contraindicated for allergies. If the patient's medical history records concomitant medications that interact with the recommended medication, the recommended medication will be replaced with a non-interacting alternative medication of the same class, or medication warning information will be added. If the patient's liver and kidney function indicators in their medical history suggest that the patient has impaired liver and kidney function, the dosage of the recommended medication should be reduced accordingly based on the standard adjustment guidelines, and the reasons for the adjustment should be recorded. If the patient's medical history shows a history of pregnancy or breastfeeding, the corresponding medications contraindicated during pregnancy / breastfeeding should be removed and replaced with alternatives that are safe and available according to the safety classification. After completing the above corrections, a revised etiology and medication information table is generated. Each potential etiology in the table includes the medication name, single dose, frequency of administration, route of administration, and necessary remarks. The risk probability and disease stage information of each potential cause in the etiology analysis report are integrated with the revised etiology medication information table to generate a unified disease analysis report and medication recommendation report. The unified disease analysis report includes the following parts: The first part is a summary of the patient's basic information and test data, including the patient's name, identification number, examination time, and a summary of key abnormal physical data; The second part is the risk etiology stratification assessment, which displays the risk probability of each potential etiology and the corresponding disease stage according to high-risk etiology, medium-risk etiology, and low-risk etiology, and also includes a graphical display of the cumulative weight value. The third part provides recommendations for differential diagnosis, suggesting further diagnostic tests for high-risk and medium-risk causes, such as adrenal vein blood sampling, low-dose dexamethasone suppression test, and renal vascular ultrasound. The medication recommendation report includes revised medication regimens for each risk stratified etiology, and indicates officially recommended medications and alternative recommended medications; the officially recommended medications are formulated for etiologies with the highest risk probability and their disease stages, while the alternative recommended medications are formulated for etiologies with medium risk, for clinicians' reference. If the risk probability of a high-risk cause is significantly higher than that of other causes (e.g., more than 50% and a difference of more than 20% from the second-ranked cause), the medication recommendation report should clearly suggest initiating empirical treatment according to the treatment plan for that cause, and monitoring key indicators during treatment to further verify the diagnosis. The output and recommendation module sends the generated unified disease analysis report and medication recommendation report to the clinical physician workstation system, and also saves them in the patient's electronic medical record. It should be noted that when patients subsequently provide new examination data or treatment effect feedback data, the intelligent analysis engine module can rerun the analysis process based on the updated real-time physical status data. The output and recommendation module can also update the medication recommendation report based on the new analysis results, forming a dynamic iterative closed loop for auxiliary decision-making.

[0021] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A secondary hypertension multi-cause intelligent analysis system based on an expert knowledge base, characterized in that, It includes an expert knowledge base module, an intelligent analysis engine module, and an output and recommendation module; The expert knowledge base module is used to obtain historical diagnosis cases of hypertension, and then generate etiology nodes and abnormal state factors based on the historical diagnosis cases of hypertension. The etiology nodes and abnormal state factors are matched with each other to establish an expert diagnosis knowledge graph. The intelligent analysis engine module is used to acquire real-time physical status data, generate patient diagnostic factors based on the real-time physical status data, input the patient diagnostic factors into the expert diagnostic knowledge graph, obtain the cumulative weight value of each etiology node through the matching results of patient diagnostic factors and abnormal status factors, determine the relationship between the cumulative weight value and the judgment threshold, and output etiology analysis report based on the judgment result. The output and recommendation module is used to obtain the patient's historical medical records, revise the etiology analysis report based on the patient's historical medical records, and then generate a unified disease analysis report and medication recommendation report based on the risk probability of potential etiologies.

2. The intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base according to claim 1, characterized in that, The hypertension history diagnosis case consists of the patient's name, identification number, range of multiple abnormal physical data related to the cause of secondary hypertension, disease progression history, and information on the types and dosages of medications used at each stage of the disease.

3. The intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base according to claim 2, characterized in that, The process of generating etiological nodes and abnormal state factors based on the history of hypertension diagnosis includes: Several etiological nodes are established based on the etiological association information of secondary hypertension recorded in the historical diagnosis cases of hypertension, and each etiological node is assigned a unique number. The etiology node corresponds to a specific stage of a potential etiology. That is, a potential etiology may have multiple corresponding etiology nodes depending on its disease progression, and each etiology node is bound to a drug use node. The drug use node records the types and dosages of drugs routinely used in clinical practice at the corresponding stage of the disease. The drug types and dosages are directly extracted from the types and dosages of drugs used at each stage of the disease recorded in the historical diagnosis cases of hypertension. Furthermore, abnormal condition factors are established based on the range of abnormal physical data recorded in the historical diagnosis cases of hypertension. Each abnormal state factor corresponds to a range of physical state data. The numerical ranges or data types included in different abnormal state factors are not completely the same, and the same physical data can be divided into multiple non-overlapping abnormal state factors according to the degree of abnormality.

4. The intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base according to claim 3, characterized in that, The process of establishing an expert diagnostic knowledge graph by matching causal nodes and abnormal state factors includes: Each etiology node is associated and matched with abnormal state factors. The historical diagnosis cases of hypertension are traversed. If the disease stage corresponding to the etiology of secondary hypertension recorded in the historical diagnosis cases of hypertension exists at the same time as a certain abnormal body data range, a connection line is established between the etiology node and the abnormal state factor corresponding to the abnormal body data range, and the association weight value is set on the connection line. The correlation weight value is obtained by statistically analyzing the frequency of simultaneous occurrence of each etiological node and abnormal state factor in historical hypertension diagnoses, and dividing it by the total frequency of cases in which the etiological node occurs. Once all causal nodes and abnormal state factors have been matched and associated, an expert diagnostic knowledge graph is generated. The expert diagnostic knowledge graph is presented in a network structure, containing two main levels: the causal layer and the factor layer. The causal layer includes all causal nodes and their associated drug use nodes, while the factor layer includes all abnormal state factors.

5. The intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base according to claim 4, characterized in that, The process of generating patient diagnostic factors based on real-time physical status data and inputting these factors into an expert diagnostic knowledge graph includes: Several patient diagnostic factors are generated based on the actual measurement values ​​of various physical data recorded in real-time physical status data. All patient diagnostic factors are mapped to the factor layer of the expert diagnostic knowledge graph. That is, the data range of the patient diagnostic factors is matched with the physical status data range corresponding to the abnormal state factors. If the actual measurement value of any physical data in the patient diagnostic factors falls completely within the numerical range specified by a certain abnormal state factor, then the abnormal state factor is marked as active. If the actual measurement value does not fall within the numerical range of any abnormal state factor, then no corresponding abnormal state factor is activated.

6. The intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base according to claim 5, characterized in that, The process of obtaining the cumulative weight values ​​of each etiological node based on the matching results of patient diagnostic factors and abnormal state factors includes: After all patient diagnostic factors and abnormal state factors are matched, cumulative weight values ​​are calculated based on the connection relationships between activated abnormal state factors and etiological nodes in the expert diagnostic knowledge graph. For each causal node, iterate through all the abnormal state factors connected to it. If an abnormal state factor belongs to the set of active abnormal state factors, then add the associated weight value on the connection line to the cumulative weight value of the causal node. If an abnormal state factor is not in an active state, then it is not included in the current cumulative weight value of the causal node.

7. The intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base according to claim 6, characterized in that, The process of determining the relationship between the cumulative weight value and the judgment threshold, and outputting a pathogenesis analysis report based on the judgment result, includes: After the above calculations, each causal node obtains its own cumulative weight value. A first judgment threshold and a second judgment threshold are preset, where the first judgment threshold is less than the second judgment threshold. The cumulative weight value of each causal node is then compared with the judgment threshold respectively. If the cumulative weight value of the causal node is greater than or equal to the second judgment threshold, it is determined that the potential causal cause and its disease stage corresponding to the causal node are highly likely, and the potential causal cause is marked as a high-risk causal cause. If the cumulative weight value of the causal node is between the first judgment threshold and the second judgment threshold, then the potential causal node is judged to have a medium probability and is marked as a medium-risk causal node. If the cumulative weight value of the causal node is less than the first judgment threshold, it is determined that the current probability of the potential causal cause is low and it is marked as a low-risk causal cause. Based on the judgment results, a pathogenesis analysis report is generated. The pathogenesis analysis report records the risk probability of each potential pathogenesis, the corresponding disease stage, and the corresponding recommended medication information. The risk probability is presented as a percentage after being normalized according to the cumulative weight value. The disease stage information is extracted from the definition of the pathogenesis node itself, and the medication information is extracted from the drug type and dosage information recorded in the drug use node bound to the pathogenesis node.

8. The intelligent analysis system for multiple causes of secondary hypertension based on an expert knowledge base according to claim 7, characterized in that, The process of generating a standardized disease analysis report and medication recommendation report based on the risk probability of potential causes includes: If the patient's medical history records concomitant medications that interact with the recommended medication, the recommended medication will be replaced with a non-interacting alternative medication of the same class, or medication warning information will be added. After completing the above corrections, a revised etiology and medication information table is generated. Each potential etiology in the table includes the medication name, single dose, frequency of administration, route of administration, and necessary remarks. The risk probability and disease stage information of each potential cause in the etiology analysis report are integrated with the revised etiology medication information table to generate a unified disease analysis report and medication recommendation report.

Images