Medical inquiry system based on AI model

Abstract

The invention relates to the technical field of medical inquiry, in particular to a medical inquiry system based on an AI model. The system comprises an inquiry data arrangement module and an AI response text updating module. According to the invention, response items and associated response items in the AI model inquiry process are organized and generated through the inquiry data organizing module, the corresponding conventional medical inquiry process database is matched, and the disease type of the patient is predicted, so that the disease-associated diseases of the patient and the diseases of the associated diseases can be collected to the maximum extent, and the diagnosis efficiency is improved. Predicted diseases are proved through disease inversion of associated diseases, the prediction accuracy is further improved, prediction errors are reduced, comprehensive and multi-level inquiry information is provided for later treatment, the occurrence probability of hidden disease items associated with diagnosed diseases and the occurrence probability of response items are obtained through an AI response text updating module, and the diagnosis accuracy is improved. And updating the corresponding associated response items and the response sequence of the response items, thereby improving the prediction efficiency of the associated diseases.

Description

Technical Field

[0001] The present invention relates to the field of medical consultation technology, and in particular to a medical consultation system based on an AI model. Background Art

[0002] AI medical consultation refers to the use of artificial intelligence technology, through natural language processing, machine learning and other technical means to achieve automated and intelligent medical consultation services. It can help patients quickly obtain medical information, provide personalized health advice, and to a certain extent assist doctors in preliminary diagnosis and treatment.

[0003] Since the response text in AI medical consultation is obtained through consultation data from the actual medical industry, the response process is based on the patient's description and combined with the actual symptoms to predict the type of disease the patient suffers. However, different patients have different physical conditions, and the types of symptoms corresponding to different diseases are prone to overlap. At the same time, the occurrence of certain diseases is often accompanied by other related diseases, and these related diseases are related to specific geographical environments and special periods. In the prediction process, the results obtained by simply matching diseases with symptoms are often not accurate enough. Therefore, in the process of AI intelligent consultation, it is necessary to combine specific disease types, geographical environments and special periods to formulate matching response texts, accurately collect the corresponding symptoms of patients in all aspects, and improve prediction accuracy.

[0004] In order to address the above problems, a medical consultation system based on AI model is urgently needed. Summary of the invention

[0005] The purpose of the present invention is to provide a medical consultation system based on an AI model to solve the problems raised in the above background technology.

[0006] To achieve the above purpose, a medical consultation system based on an AI model is provided, including an industry consultation process information collection module, a risk content assessment module, and an AI consultation platform;

[0007] Among them, the industry consultation process information collection module is used to collect the manual consultation process of the medical industry, obtain the consultation process items under different disease states, and generate a conventional medical consultation process database;

[0008] The risk content assessment module combines the later treatment process and treatment results to obtain hidden disease items associated with the diagnosed disease, and combines the significant features of the hidden disease items to match the corresponding consultation content to generate a database of related disease consultation processes;

[0009] The AI ​​consultation platform includes a consultation data sorting module and an AI response text updating module. The consultation data sorting module combines a conventional medical consultation process database and a related disease consultation process database to sort out and generate response items and related response items in the AI ​​model consultation process;

[0010] The AI ​​response text update module is used to obtain the occurrence probability of hidden symptom items and the occurrence probability of response items associated with the confirmed symptoms, and update the corresponding associated response items and the response order of the response items.

[0011] As a further improvement of the technical solution, the process items in the industry consultation process information collection module include corresponding symptoms, symptom status and inquiry items;

[0012] The corresponding symptoms are other symptoms that the doctor asks the patient about based on the description of the symptoms provided by the patient and the examination results;

[0013] The disease state is the severity of the corresponding disease;

[0014] The query items are keywords for each symptom, and the symptom content is determined by the keywords.

[0015] As a further improvement of the technical solution, the method for generating a conventional medical consultation process database in the industry consultation process information collection module includes the following steps:

[0016] S101, acquiring keywords of corresponding symptoms of the predicted disease in combination with the query items;

[0017] S102, obtaining the symptom items provided by the patient to the consulting doctor and the symptom items asked by the consulting doctor about the patient through the keywords corresponding to the symptom;

[0018] S103. Collect historical medical data of the current predicted disease and calculate the associated probability Ap of the symptom items of the predicted disease probability , according to the disease item accompanying probability Ap probability Sort the consultation order for each disease item and establish a database of routine medical consultation processes.

[0019] As a further improvement of the technical solution, the hidden symptom items in the risk content assessment module include risk habits, treatment misunderstandings, induced symptoms and hidden symptoms;

[0020] The risk habits are the risk of concomitant diseases caused by the patient's living habits and the environment in which he lives;

[0021] The treatment error is that the patient does not follow the doctor's instructions to carry out the treatment program for the predicted disease;

[0022] The inducing condition is a disease associated with the current prediction of the disease routine;

[0023] The hidden disease is an associated genetic disease caused by the patient's family genetic history for the predicted disease.

[0024] As a further improvement of the technical solution, the method for generating a database of associated disease consultation processes in the risk content assessment module includes the following steps:

[0025] S201. Collect statistical data information of the medical industry to obtain the correlation probability Ac of each concomitant disease of the predicted disease under different hidden disease items probability ;

[0026] S202, setting the priority order of each hidden disease project;

[0027] S203, formulating the consultation contents of the accompanying diseases matched with each hidden disease in order of priority, and generating a database of consultation processes for related diseases.

[0028] As a further improvement of the present technical solution, the priority order of each hidden symptom item in S202 is inducing symptom>risk habit>hidden symptom>treatment misunderstanding.

[0029] As a further improvement of the technical solution, the method for updating the corresponding associated answer items and the answer order of the answer items in the AI ​​answer text update module comprises the following steps:

[0030] S401, predicting the disease currently suffered by the patient based on the description of the symptoms and the examination results provided by the patient;

[0031] S402, combining with the conventional medical consultation process database, obtaining various symptom items of the predicted disease, and eliminating the symptom items mentioned by the patient, and the remaining symptom items are sorted according to the accompanying probability Ap probability Sorting, matching the query order that generates the corresponding answer items;

[0032] S403, identifying the hidden disease items involved in the patient's medical consultation process, and generating a corresponding response sequence for the accompanying diseases according to the corresponding priorities and the probability of the accompanying diseases.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] In the medical consultation system based on the AI ​​model, the consultation data organization module organizes and generates the response items and related response items in the AI ​​model consultation process, matches the corresponding conventional medical consultation process database, and predicts the type of disease suffered by the patient, so as to maximize the collection of the patient's disease-related symptoms and symptoms of related diseases, and reversely prove the predicted disease through the symptoms of related diseases, further improve the prediction accuracy, reduce prediction errors, and provide comprehensive and multi-level consultation information for later treatment. The AI ​​response text update module obtains the probability of occurrence of hidden symptom items and response items associated with the confirmed disease, updates the corresponding related response items and the response order of the response items, thereby improving the prediction efficiency of related diseases. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is the overall structural diagram of the present invention;

[0036] Figure 2 A diagram showing the steps of a method for generating a conventional medical consultation process database according to the present invention;

[0037] Figure 3 A diagram showing the steps of a method for generating a database of consultation processes for associated symptoms according to the present invention;

[0038] Figure 4 It is a method step diagram for updating corresponding associated response items and the response sequence of the response items according to the present invention.

[0039] The meaning of each number in the figure is:

[0040] 10. Industry consultation process information collection module;

[0041] 20. Risk content assessment module;

[0042] 30. The AI ​​consultation platform includes a consultation data sorting module;

[0043] 40. AI response text update module. DETAILED DESCRIPTION

[0044] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] See also Figure 1 As shown, a medical consultation system based on an AI model is provided, including an industry consultation process information collection module 10, a risk content assessment module 20, and an AI consultation platform;

[0046] Among them, the industry consultation process information collection module 10 is used to collect the manual consultation process of the medical industry, obtain the consultation process items under different disease states, and generate a conventional medical consultation process database;

[0047] The risk content assessment module 20 combines the later treatment process and treatment results to obtain the hidden disease items associated with the diagnosed disease, and combines the significant features of the hidden disease items to match the corresponding consultation content to generate a database of related disease consultation processes;

[0048] The AI ​​consultation platform includes a consultation data sorting module 30 and an AI response text updating module 40. The consultation data sorting module 30 combines the conventional medical consultation process database and the related disease consultation process database to sort out and generate the response items and related response items in the AI ​​model consultation process;

[0049] The AI ​​response text updating module 40 is used to obtain the occurrence probability of hidden symptom items and the occurrence probability of response items associated with the confirmed symptoms, and update the corresponding associated response items and the response order of the response items.

[0050] When used specifically, during the AI ​​intelligent consultation process, first, in order to enrich the response text content in the AI ​​consultation model, it is necessary to collect the manual consultation process of the medical industry through the industry consultation process information collection module 10, obtain the process items under different symptom states, and generate a conventional medical consultation process database, that is, after the patient provides the symptom description and examination results and other information to the consulting doctor, the consulting doctor predicts the type of disease the current patient suffers from through his own judgment, and provides it to the patient again as the consultation content based on the historical manifestation of the disease, and obtains patient feedback to further cite the consulting doctor's speculation. The consultation content involved is the process item;

[0051] Since some diseases are accompanied by other related diseases as they develop or become more serious, and these diseases are often still in the early stages, the corresponding symptoms are not obvious enough, causing the patient not to mention them during the consultation process. In order to remind the patient, the patient needs to be provided with the unclear symptoms for recollection based on further consultation. It is necessary to obtain the hidden symptom items associated with the confirmed symptoms through the risk content assessment module 20 in combination with the later treatment process and treatment results, and match the corresponding consultation content in combination with the significant features of the hidden symptom items to generate a database of related symptom consultation processes, that is, to provide the patient with the significant features of the related diseases that may occur as the consultation content, so as to help the patient recall whether the situation involved in the consultation content has occurred, and to prevent omissions in the consultation, which may affect the later treatment effect.

[0052] Therefore, the corresponding AI consultation platform collects data from the related disease consultation process database and the conventional medical consultation process database, and organizes and generates response items and related response items in the AI ​​model consultation process through the consultation data organization module 30. That is, in the AI ​​intelligent consultation process, the patient's disease type is predicted through the description of the disease and the examination results and other information provided by the patient, and the regular response items are provided to the patient as regular response items, and the patient's feedback information is obtained. Finally, the patient's related disease type is predicted based on the feedback information and the symptom description and examination results and other information provided in the early stage, and the corresponding related response items are matched in the related disease consultation process database, and provided to the user as secondary consultation information, and the patient's secondary feedback information is obtained, so as to maximize the collection of the patient's disease-related symptoms and the symptoms of related diseases, and reversely prove the predicted disease through the symptoms of related diseases, further improve the prediction accuracy, reduce the prediction error, and provide comprehensive and multi-level consultation information for later treatment.

[0053] It is worth noting that since some symptoms match more disease types, the corresponding associated disease types will also increase. Therefore, when the AI ​​model completes the disease prediction work, multiple hidden symptom items of related diseases will be generated in the corresponding response text. If all of them are interviewed, it will not only lead to a long consultation time, but also cause patients' disgust. Therefore, in order to solve the above problem, the AI ​​response text update module 40 obtains the probability of occurrence of hidden symptom items associated with confirmed symptoms and the probability of occurrence of response items, and updates the corresponding associated response items and the response order of response items, that is, the response order of corresponding symptoms is matched according to the probability of occurrence of hidden symptom items associated with the actual confirmed symptoms. When the prediction result is obtained, further inquiries will be stopped, thereby improving the prediction efficiency of associated diseases.

[0054] In addition, the process items in the industry consultation process information collection module 10 include corresponding symptoms, symptom status, and inquiry items;

[0055] The corresponding symptoms are the other symptoms that the consulting doctor asks the patient about based on the description of the symptoms provided by the patient and the examination results, that is, the symptoms that the patient has not provided, but these symptoms are significant symptoms in the disease predicted by the consulting doctor, and it is necessary to determine whether these symptoms appear to further determine the predicted results;

[0056] The disease state is the severity of the corresponding disease, for example, determined by the pain level;

[0057] The inquiry items are keywords for each symptom, and the symptom content is determined by the keywords.

[0058] Further, such as Figure 2As shown, the method for generating a conventional medical consultation process database in the industry consultation process information collection module 10 includes the following steps:

[0059] S101. Acquire keywords corresponding to symptoms of the predicted disease in combination with the query items;

[0060] S102, obtaining the symptom items provided by the patient to the consulting doctor and the symptom items asked by the consulting doctor about the patient through the keywords corresponding to the symptom;

[0061] S103. Collect historical medical data of the current predicted disease and calculate the associated probability Ap of the symptom items of the predicted disease probability , according to the disease item accompanying probability Ap probability Sort the consultation order for each disease item and establish a database of routine medical consultation processes.

[0062] In specific use, since some diseases have a lot of corresponding symptoms, and different patients have different numbers of symptoms, in order to reduce the content of the later AI consultation, it is necessary to first combine the inquiry items to obtain the keywords of the corresponding symptoms of the predicted disease, and obtain the symptom items provided by the patient to the consulting doctor and the symptom items asked by the consulting doctor through the keywords of the corresponding symptoms. The disease items provided by the patient do not need to be repeatedly asked by the consulting doctor. At the same time, in order to improve the efficiency of the inquiry, it is necessary to collect the historical medical data of the current predicted disease and calculate the probability Ap of the symptom items of the predicted disease. probability , according to the disease item accompanying probability Ap probability Sort the consultation order of each symptom item and establish a database of routine medical consultation process. It is worth noting that the judgment of a certain disease is determined by the determination of multiple matching symptom items, medical device inspection results and doctor analysis and judgment. Therefore, for different diseases, the matching symptom items that meet the prediction results are diverse and different in different periods and environments. Therefore, in the later AI consultation process, the symptom items that need to be predicted as a successful match are defined by the doctor, and the definition method is through the accompanying probability Ap probability The ranking is selected from the front to the back, and the specific number is determined by the doctor.

[0063] Further, the hidden symptom items in the risk content assessment module 20 include risk habits, treatment misunderstandings, induced symptoms, and hidden symptoms;

[0064] Among them, risk habits refer to the risk of concomitant diseases in the patient's living habits and the environment in which he lives. For example, a patient who smokes for a long time may be predicted to have a cold due to coughing. Due to his long-term smoking, the probability of pneumonia will increase.

[0065] Treatment errors are when patients do not follow the treatment items prescribed by doctors for the predicted diseases, such as using folk remedies, which can easily lead to other related diseases;

[0066] The precipitating disease is a disease associated with the current conventional prediction of the disease;

[0067] Hidden diseases are related genetic diseases caused by the patient's family genetic history for the predicted disease.

[0068] Specifically, Figure 3 As shown, the method for generating a database of associated disease consultation processes in the risk content assessment module 20 includes the following steps:

[0069] S201. Collect statistical data information of the medical industry to obtain the correlation probability Ac of each concomitant disease of the predicted disease under different hidden disease items probability ;

[0070] S202, setting the priority order of each hidden disease project;

[0071] S203, formulating the consultation contents of the accompanying diseases matched with each hidden disease in order of priority, and generating a database of consultation processes for related diseases.

[0072] In addition, the priority order of each hidden disease item in S202 is inducing disease > risky habits > hidden disease > treatment misunderstanding.

[0073] In specific use, in the process of establishing the associated disease consultation process database, due to the different types of hidden disease items, the corresponding concomitant disease types will be different. The hidden disease items involved in the present invention include risk habits, treatment misunderstandings, induced diseases and hidden diseases. At the same time, the probabilities of matching concomitant disease types under different hidden disease items are different. In order to further determine the consultation order of concomitant diseases, it is first necessary to collect statistical data information of the medical industry and obtain the associated probability Ac of each concomitant disease of the predicted disease under different hidden disease items. probability At the same time, when multiple hidden disease items appear in the patient's consultation, it is necessary to determine the consultation order of each hidden disease item in advance, that is, to define the priority order of each hidden disease item. The priority order of each hidden disease item in the present invention is induced disease > risk habit > hidden disease > treatment misunderstanding. That is, when the patient's consultation process involves both induced disease and treatment misunderstanding, it is necessary to firstly consult the concomitant disease caused by the induced disease. The consultation order is determined by the corresponding association probability Ac. probability Determine, the associated probability Ac probabilityThe higher the priority, the earlier the corresponding consultation order. Then, the concomitant diseases caused by treatment errors will be consulted. Finally, the consultation content of the concomitant diseases matching each hidden disease will be formulated in order of priority, and a database of related disease consultation process will be generated.

[0074] Further, such as Figure 4 As shown, the method for updating the corresponding associated response items and the response order of the response items in the AI ​​response text updating module 40 includes the following steps:

[0075] S401, predicting the disease currently suffered by the patient based on the description of the symptoms and the examination results provided by the patient;

[0076] S402, combining with the conventional medical consultation process database, obtaining various symptom items of the predicted disease, and eliminating the symptom items mentioned by the patient, and the remaining symptom items are sorted according to the accompanying probability Ap probability Sorting, matching the query order that generates the corresponding answer items;

[0077] S403, identifying the hidden disease items involved in the patient's medical consultation process, and generating a corresponding response sequence for the accompanying diseases according to the corresponding priorities and the probability of the accompanying diseases.

[0078] When used specifically, in the process of determining the response order of the response items, the patient and the symptom description mentioned must first be considered. The symptom description here is the response item. That is, the symptom description provided by the patient and the examination results are combined to predict the disease of the current patient, and the various symptom items of the predicted disease are obtained in combination with the conventional medical consultation process database, and the symptom items mentioned by the patient are eliminated. The remaining symptom items are ranked according to the accompanying probability Ap. probability Sorting, and finally matching to generate the query sequence of corresponding answer items;

[0079] In the process of determining the response order of associated response items, since the associated response items are related to the hidden disease items, it is necessary to identify the hidden disease items involved in the patient's consultation process, and generate the corresponding response order of the accompanying diseases according to the corresponding priority and the probability of the accompanying diseases, that is, the hidden disease items involved are identified through the feedback of the patient's consultation results, and the order of the associated response items is determined. The associated response items here are the accompanying diseases matched by the hidden disease items.

[0080] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A medical consultation system based on an AI model, characterized by: It includes an industry consultation process information collection module (10), a risk content assessment module (20) and an AI consultation platform; The industry consultation process information collection module (10) is used to collect the manual consultation process of the medical industry, obtain consultation process items under different disease states, and generate a conventional medical consultation process database; The risk content assessment module (20) combines the later treatment process and treatment results to obtain hidden disease items associated with the diagnosed disease, and combines the significant features of the hidden disease items to match the corresponding consultation content to generate a database of related disease consultation processes; The AI ​​consultation platform comprises a consultation data arrangement module (30) and an AI response text update module (40). The consultation data arrangement module (30) combines a conventional medical consultation process database and a related disease consultation process database to arrange and generate response items and related response items in the AI ​​model consultation process. The AI ​​response text updating module (40) is used to obtain the occurrence probability of hidden symptom items and the occurrence probability of response items associated with the confirmed symptoms, and to update the corresponding associated response items and the response order of the response items.

2. The medical consultation system based on the AI ​​model according to claim 1, characterized in that: The process items in the industry consultation process information collection module (10) include corresponding symptoms, symptom status and inquiry items; The corresponding symptoms are other symptoms that the doctor asks the patient about based on the description of the symptoms provided by the patient and the examination results; The disease state is the severity of the corresponding disease; The query items are keywords for each symptom, and the symptom content is determined by the keywords.

3. The medical consultation system based on the AI ​​model according to claim 2, characterized in that: The method for generating a conventional medical consultation process database in the industry consultation process information collection module (10) comprises the following steps: S101, acquiring keywords of corresponding symptoms of the predicted disease in combination with the query items; S102, obtaining the symptom items provided by the patient to the consulting doctor and the symptom items asked by the consulting doctor about the patient through the keywords corresponding to the symptom; S103. Collect historical medical data of the current predicted disease and calculate the associated probability Ap of the symptom items of the predicted disease probability , according to the disease item accompanying probability Ap probability Sort the consultation order for each disease item and establish a routine medical consultation process database.

4. The medical consultation system based on the AI ​​model according to claim 1, characterized in that: The hidden symptom items in the risk content assessment module (20) include risk habits, treatment misunderstandings, induced symptoms and hidden symptoms; The risk habits are the risk of concomitant diseases caused by the patient's living habits and the environment in which he lives; The treatment error is that the patient does not follow the doctor's instructions to carry out the treatment program for the predicted disease; The inducing condition is a disease associated with the current prediction of the disease routine; The hidden disease is an associated genetic disease caused by the patient's family genetic history for the predicted disease.

5. The medical consultation system based on the AI ​​model according to claim 4, characterized in that: The method for generating a database of associated disease consultation procedures in the risk content assessment module (20) comprises the following steps: S201. Collect statistical data information of the medical industry to obtain the correlation probability Ac of each concomitant disease of the predicted disease under different hidden disease items probability ; S202, setting the priority order of each hidden disease project; S203, formulating the consultation contents of the accompanying diseases matched with each hidden disease in order of priority, and generating a database of consultation processes for related diseases.

6. The medical consultation system based on the AI ​​model according to claim 5, characterized in that: The priority order of each hidden symptom item in S202 is inducing symptom>risk habit>hidden symptom>treatment misunderstanding.

7. The medical consultation system based on the AI ​​model according to claim 1, characterized in that: The method for updating the corresponding associated response items and the response order of the response items in the AI ​​response text updating module (40) comprises the following steps: S401, predicting the disease currently suffered by the patient based on the description of the symptoms and the examination results provided by the patient; S402, combining with the conventional medical consultation process database, obtaining various symptom items of the predicted disease, and eliminating the symptom items mentioned by the patient, and the remaining symptom items are sorted according to the accompanying probability Ap probability Sorting, matching the query order that generates the corresponding answer items; S403, identifying the hidden disease items involved in the patient's medical consultation process, and generating a corresponding response sequence for the accompanying diseases according to the corresponding priorities and the probability of the accompanying diseases.

Images