Personalized AI-powered medical consultation system

Abstract

To provide an AI-powered medical interview system based on Oriental medicine that is highly convenient for patients and possesses sufficient accuracy in conducting medical interviews. [Solution] An artificial intelligence medical interview system that conducts a medical interview using artificial intelligence based on traditional Chinese medicine, comprising: an inquiry unit that asks questions to the patient regarding the patient's personal information and health condition; a patient information acquisition unit that acquires patient information in response to the questions asked by the inquiry unit; an artificial intelligence inquiry unit that analyzes the patient information acquired by the patient information acquisition unit and asks the most appropriate questions for diagnosis; an artificial intelligence analysis unit that analyzes the patient information acquired by the patient information acquisition unit and conventional knowledge of Oriental medicine and determines the optimal health condition; and a results display unit that displays the medical interview results analyzed by the artificial intelligence analysis unit.

Description

Technical Field

[0001] The present invention is a personalized artificial intelligence consultation system based on traditional Chinese medicine that uses artificial intelligence.

Background Art

[0002] In the process of diagnosing a disease, first, a consultation is conducted by medical staff such as a doctor, and based on the information of the examinee obtained from the consultation, an examination is performed, and a diagnosis is made based on the consultation result and the examination result.

[0003] Here, the information obtained by consultation includes personal information of the examinee, chief complaint, past history, diseases suffered in childhood, presence or absence of current medication, presence or absence of major diseases within the last year, etc.

[0004] In recent years, artificial intelligence technology has been applied to medical care. Specifically, there are systems in which artificial intelligence identifies a medical condition from the symptoms of a patient, and image recognition technology in which artificial intelligence diagnoses a medical condition based on a diagnostic image of the patient.

[0005] By using artificial intelligence in medical care as described above, it is possible to expect an improvement in the accuracy of disease diagnosis, stable diagnostic accuracy that does not depend on the physical condition of doctors, etc., and a reduction in the workload of medical staff.

[0006] In addition, a consultation system based on Western medicine that predicts a disease name based on the information input by an examinee is provided (Patent Document 1).

[0007] In addition, an automatic consultation system using artificial intelligence based on traditional Chinese medicine is known as a known technique (Patent Document 2).

[0008] In such a consultation using artificial intelligence, since artificial intelligence conducts the consultation, it can be carried out without limiting the consultation place, the time for chart input, etc., and has merits such as reducing the waiting time of the examinee in the medical institution and assisting in the judgment of the medical department where the examinee receives medical treatment.

[0009] Furthermore, because the examination is performed by artificial intelligence, the accuracy of the medical interview is high and stable, offering advantages such as preventing major medical errors that could occur due to human error. [Prior art documents] [Patent Documents]

[0010] [Patent Document 1] :Patent No. 6496876 [Patent Document 2] : JP 2021-47504 [Non-patent literature]

[0011] [Non-Patent Document 1] :Ryutaro Arita et al., "Identifying and Solving Challenges in Aiming for a Patient-Centered Automated Medical Interview System," Japanese Journal of Oriental Medicine, Vol. 69, No. 1, pp. 82-90, 2018. [Overview of the Initiative] [Problems that the invention aims to solve]

[0012] However, the medical interview system described in Patent Document 1 is a system that calculates the probability of disease based on the patient's answers, and does not use artificial intelligence or the like. It merely automates a general medical interview and cannot be said to have sufficient accuracy.

[0013] Furthermore, although the medical interview system described in Patent Document 2 is a medical interview system based on Oriental medicine using artificial intelligence, it does not incorporate improvements to the interview format or other measures to enhance the accuracy of the interview. Therefore, it is desirable to have a medical interview format that is as tailored as possible to the individual characteristics of each patient, thereby providing greater convenience for the patient.

[0014] Furthermore, conventional AI-powered medical interviews, while aiming to improve interview accuracy, currently require a large number of questions to identify diseases, demanding considerable time and effort from patients to input their answers. On the other hand, many patients are elderly and prone to illness, and many are unfamiliar with operating electronic devices. Therefore, it is desirable to provide a medical interview system that minimizes the number of questions, accurately identifies symptoms, is convenient for patients, and maintains high interview accuracy.

[0015] This invention aims to solve the above problems. [Means for solving the problem]

[0016] The present invention is an artificial intelligence medical interview system that performs a constitution assessment based on the knowledge of Oriental medicine and generates the next questions based on the patient's answers, comprising: an inquiry unit that asks the patient questions about the patient's personal information and health condition; a patient information acquisition unit that acquires patient information answered by the patient in response to the questions issued by the inquiry unit; an artificial intelligence inquiry unit that analyzes the patient information acquired by the patient information acquisition unit and asks the most appropriate questions for diagnosis; an artificial intelligence analysis unit that analyzes information regarding health maintenance methods and prescriptions for herbal medicines that are suitable for the patient's constitution based on the patient information acquired by the patient information acquisition unit and conventional knowledge of Oriental medicine; and a results display unit that displays the medical interview results analyzed by the artificial intelligence analysis unit, characterized in that it comprises: an inquiry unit that asks the patient questions about the patient's personal information and health condition; a patient information acquisition unit that acquires the patient information acquisition unit and conventional knowledge of Oriental medicine and analyzes information regarding health maintenance methods and prescriptions for herbal medicines that are suitable for the patient's constitution; and a results display unit that displays the medical interview results analyzed by the artificial intelligence analysis unit.

[0017] Furthermore, the present invention is a personalized artificial intelligence medical interview and diagnosis system according to claim 1, characterized in that the total number of questions that the artificial intelligence questioning unit asks for the diagnosis is in the range of 10 to 20.

[0018] Furthermore, the present invention is a personalized artificial intelligence medical interview and diagnosis system according to claim 2, characterized in that the questioning unit changes the content of the questions according to the gender, age, and physical constitution of each of the patients and asks questions to derive the optimal health condition tailored to each patient.

[0019] In addition, the present invention provides the personalize artificial intelligence interrogation diagnosis system according to claim 3, characterized in that the result display unit for displaying the optimal health state analyzes the recipient information within approximately 10 seconds and derives and displays the optimal health state.

[0020] In addition, the present invention provides the personalize artificial intelligence interrogation diagnosis system according to claim 4, characterized in that the display of the optimal health state by the result display unit reflects a SOAP summary summarizing the doctor's diagnosis result and re-diagnoses, enabling highly accurate diagnosis.

[0021] In addition, the present invention provides a personalized artificial intelligence interrogation method for performing interrogation by artificial intelligence based on traditional Chinese medicine, comprising: an interrogation step of issuing an interrogation about the health state of the recipient to the recipient; a response step of responding with recipient information, which is the recipient's response to the interrogation; and an output step of displaying an analysis result related to the health state based on the recipient information of the recipient. The interrogation step is characterized in that the artificial intelligence determines and proposes the content and number of interrogations based on the recipient information, and the output step is characterized in that the artificial intelligence analyzes and displays the analysis result based on the recipient information.

[0022] In addition, the present invention provides a program for causing a computer to execute the personalized artificial intelligence interrogation method according to claim 6.

Effect of the Invention

[0023] By using the artificial intelligence interrogation system based on traditional Chinese medicine according to the present invention, it is possible to know the health state of the recipient at the time of medical examination, the method of preventing diseases, or traditional Chinese medicines suitable for the treatment of one's own health state. In addition, since the artificial intelligence examines and formulates the interrogation questions, the number of interrogation questions is configured to be as small as possible, requiring little labor for the recipient to answer and maintaining sufficient interrogation accuracy. Therefore, it is an excellent mode in terms of convenience for the recipient.

Brief Description of the Drawings

[0024] [Figure 1] It is a schematic diagram showing the configuration of an artificial intelligence medical interview system according to an embodiment of the present invention. [Figure 2] It is a flowchart showing the flow of a medical interview of an artificial intelligence medical interview system according to an embodiment of the present invention. [Figure 3] It is a diagram showing a medical interview screen of an artificial intelligence medical interview system according to an embodiment of the present invention. [Figure 4] It is a diagram showing a medical interview screen of an artificial intelligence medical interview system according to an embodiment of the present invention. [Figure 5] It is a table showing a setting mode of provisional points according to an embodiment of the present invention. [Figure 6] It is a diagram showing a document for calculating provisional points according to an embodiment of the present invention. [Figure 7] It is a diagram showing a medical interview result display screen of an artificial intelligence medical interview system according to an embodiment of the present invention. [Figure 8] It is a schematic diagram showing the configuration of an artificial intelligence medical interview system according to an embodiment of the present invention.

Modes for Carrying Out the Invention

[0025] The present invention is an artificial intelligence medical interview system that performs a constitution assessment based on the knowledge of Oriental medicine and generates subsequent questions based on the patient's answers, comprising: an inquiry unit that asks the patient questions about the patient's personal information and health condition; a patient information acquisition unit that acquires patient information answered by the patient in response to the questions issued by the inquiry unit; an artificial intelligence inquiry unit that analyzes the patient information acquired by the patient information acquisition unit and asks the most appropriate questions for diagnosis; an artificial intelligence analysis unit that analyzes information regarding health maintenance methods and prescriptions for herbal medicines that are suitable for the patient's constitution based on the patient information acquired by the patient information acquisition unit and conventional knowledge of Oriental medicine; and a results display unit that displays the medical interview results analyzed by the artificial intelligence analysis unit, characterized in that it comprises: an inquiry unit that asks the patient questions about the patient's personal information and health condition; a patient information acquisition unit that acquires the patient information acquisition unit and conventional knowledge of Oriental medicine and analyzes information regarding health maintenance methods and prescriptions for herbal medicines that are suitable for the patient's constitution; and a results display unit that displays the medical interview results analyzed by the artificial intelligence analysis unit.

[0026] Furthermore, in traditional Chinese medicine, the four diagnostic methods—inquiry, observation, auscultation, and palpation—are generally known. This embodiment provides a system that substitutes for inquiry, one of these four diagnostic methods.

[0027] Furthermore, in the artificial intelligence medical interview system according to one embodiment of the present invention, big data such as medical data based on Oriental medicine and information on herbal medicines is used to train the artificial intelligence and create a trained model. Then, using the artificial intelligence equipped with this trained model, the actual diagnosis of the patient in this artificial intelligence medical interview system is performed in the form of a medical interview, and the artificial intelligence analyzes the results.

[0028] [1. Configuration of the artificial intelligence medical interview system according to one embodiment of the present invention] First, the configurations of each part of the artificial intelligence medical interview according to one embodiment of the present invention will be described using Figure 1. As shown in Figure 1, the artificial intelligence medical interview system according to the present invention consists of a unit having a questioning function, an acquisition function, an analysis function, and a display function.

[0029] The unit with questioning capabilities consists of a questioning unit 1 and an artificial intelligence questioning unit 2. The unit with acquisition capabilities consists of a patient information acquisition unit 3. The unit with analysis capabilities consists of an analysis unit 4 and an artificial intelligence analysis unit 5. The unit with display capabilities consists of a result display unit 6. Of these, the artificial intelligence questioning unit 2 and the artificial intelligence analysis unit 5 are equipped with artificial intelligence.

[0030] Next, let's describe each part. First, Questioning Unit 1 is the unit that asks questions to the patient regarding their health condition. Unlike the AI ​​Questioning Unit 2, Questioning Unit 1 does not have artificial intelligence. Questioning Unit 1 is used to ask standard questions, such as the patient's gender and age.

[0031] Next, let's discuss the AI ​​Questioning Unit 2. The AI ​​Questioning Unit 2 functions when asking questions to the patient about their health condition, in addition to the standard phrases mentioned above. For example, based on the patient's answers, the AI ​​asks the most appropriate questions for the medical interview.

[0032] This artificial intelligence questioning unit 2 is equipped with artificial intelligence as its system, and it is trained on existing data of question content and question wording for specific answers as training data. Furthermore, this artificial intelligence questioning unit 2 functions in conjunction with the artificial intelligence analysis unit 5.

[0033] The learning data for this question, as detailed below, consists of questions about fatigue, loss of appetite, sleep, coldness, dry mouth, stiff shoulders, urge to urinate, constipation, excessive sweating, menstruation, etc.

[0034] Next, the patient information acquisition unit 3 will be described. The patient information acquisition unit 3 is a unit that acquires patient information such as gender and age, and information about their health condition, when a patient undergoing a medical interview using the medical interview system according to the present invention answers such information. In other words, the information acquired by this patient information acquisition unit 3 is the answer information to the questions devised by the questioning unit 1 or the artificial intelligence questioning unit 2.

[0035] Next, let's discuss the analysis unit 4. The analysis unit 4 is a unit that analyzes patient information without using artificial intelligence. This analysis unit 4 functions when the answer and the next action are determined based on the input information. One example of how this analysis unit 4 functions is that if the patient selects "yes" to a certain question, it prompts the system to ask a new question, and if the patient selects "no," it ends the question. In this way, the analysis unit 4 functions when providing standardized responses without requiring artificial intelligence or anything similar.

[0036] Next, let's discuss the artificial intelligence analysis unit 5. The artificial intelligence analysis unit 5 is the unit that uses artificial intelligence to analyze the input information entered by the patient. As will be explained in more detail later, the analysis by this artificial intelligence analysis unit 5 determines the probability of a patient's condition being based on traditional Chinese medicine, and generates questions necessary for estimating the patient's condition.

[0037] This artificial intelligence analysis unit 5 is equipped with a learning model that has been trained on pre-created training data, such as data on disease conditions based on conventional Oriental medicine and data based on conventional patient diagnosis results.

[0038] Next, we will describe the results display unit 6. The results display unit 6 is a display unit that shows the results of the medical questionnaire regarding the patient's health condition, which are analyzed and derived by artificial intelligence based on the patient information entered by the patient. This results display unit 6 displays a summary of the questionnaire, the probability of the estimated disease condition, health maintenance methods, and suitable herbal medicines.

[0039] [2. Flowchart of an AI-powered medical interview related to one embodiment of the present invention] Next, the flow of an artificial intelligence medical interview system according to one embodiment of the present invention will be described using Figure 2. As shown in Figure 2, the artificial intelligence medical interview system according to one embodiment of the present invention consists of the following steps: input of patient information by the patient (step S1), presentation of questions about health status by artificial intelligence (step S2), input of answers to the questions by the patient (step S3), calculation of a score for each constitution category based on the input information by artificial intelligence (step S4), prediction of disease name based on the score by artificial intelligence (step S5), and display of a summary result such as treatment methods and optimal herbal medicines for the predicted disease name (step S6).

[0040] Furthermore, between the AI ​​generating questions about the patient's health status (Step S2) and the patient inputting their answers to those questions (Step S3), the AI ​​questioning unit 2 makes appropriate judgments based on the patient's answers and presents suitable questions. Therefore, Steps S2 and S3 are repeated until the final question is reached.

[0041] Furthermore, within this AI-powered medical interview process, steps S1 (input of patient information by the patient) to S3 (input of answers to questions by the patient) will be referred to as the AI ​​medical interview step, steps S4 (calculation of scores for each constitution category based on the input information by AI) to S5 (prediction of disease name based on the AI ​​score) will be referred to as the constitution determination step, and step S6 (display of summary results such as health maintenance methods and optimal herbal medicines for the predicted disease name) will be referred to as the output step.

[0042] <About the AI-powered medical interview steps> Next, we will describe each step in order. First, we will describe the artificial intelligence consultation steps (steps S1 to S3) according to one embodiment of the present invention.

[0043] First, to receive a medical consultation using the artificial intelligence consultation system according to one embodiment of the present invention, this application is launched.

[0044] The hardware required to run this application can be a personal computer, a tablet, or a smartphone. Any electronic device capable of running this application is acceptable.

[0045] When this application is launched, a screen for entering patient information such as the patient's name, gender, age, and phone number will appear. The patient enters the information on this screen using an input device such as a stylus pen or keyboard (Step S1). This input device may also be voice input via a microphone or similar device. In this configuration, the medical interview can be conducted with less burden on elderly users and others.

[0046] Furthermore, input via the aforementioned input device may involve selecting buttons in a Yes or No format, or it may involve numerical input or direct text input. This input format is not limited to this patient information input screen, and the same applies to questions presented by the artificial intelligence.

[0047] Furthermore, the answer choices should be tailored to the question, and additional options such as "Neither agree nor disagree" or "Don't know" should be added.

[0048] Furthermore, this response format allows the patient to select multiple applicable answers. Specifically, for the question, "What concerns you about your health recently?", the patient can select all applicable items such as "I can't sleep enough," "I have no appetite," and "I feel very tired."

[0049] Furthermore, for example, the response to a question about the frequency of abdominal pain may be in a format that describes the frequency or severity, such as once a month or less, two to five times a month, or six to ten times a month.

[0050] Next, the questions about health status presented by the artificial intelligence in Step S2 are specifically divided into 10 to 20 questions, each categorized into nine major disease conditions.

[0051] The number of questions presented will vary as needed by the artificial intelligence analysis unit based on the patient's responses. However, it is also possible to pre-determine, for example, "only 10 questions will be presented." In this configuration, patients can more easily use this medical questionnaire system.

[0052] Next, the nine categories presented in the AI-powered medical interview step include, for example, fatigue, loss of appetite, insomnia, coldness, dry mouth, stiff shoulders, urge to urinate, bowel movements, and excessive sweating. These nine categories are not limited to this form. However, for a general diagnosis of the whole body, it is desirable that the questions include content that covers various parts of the body or basic symptoms.

[0053] Furthermore, it is advisable to further subdivide the questions into the nine categories mentioned above. Specifically, as shown in Figures 3 and 4, if the question "Do you often get thirsty?" is answered with "Yes," the next question will be "Do you carry a water bottle or thermos with you when you go out?" If the answer to this question is "Often," the next question will be "How often?"

[0054] The artificial intelligence analysis unit 5 will decide whether to ask more detailed questions about the content of these questions or what kind of questions to ask. In this case, the artificial intelligence analysis unit 5 will determine and ask questions that are most appropriate for diagnosing the symptoms based on the input information entered by the patient, the patient's gender, physical characteristics, etc. If the patient selects "No" to the first question in Figure 3, "Do you often get thirsty?", the content of the subsequent questions will change, or all subsequent questions related to "thirst" will be omitted.

[0055] Furthermore, the subsequent questions will differ depending on the individual, based on their answers to the initial question, "Do you often get thirsty?". At this time, the artificial intelligence will determine and ask the necessary questions based on the individual's previous responses.

[0056] Furthermore, the total number of questions for each category should fall within the range of 10 to 20 questions (derivative questions like those shown in Figure 4 above will be counted as one question).

[0057] Furthermore, questions that are presumed to be correlated will be combined into a single, unified question. For example, regarding shoulder stiffness and neck stiffness, it is assumed that patients with shoulder stiffness often also have neck stiffness. In such cases, the questions will be combined into one, or only one question will be asked. Therefore, patients will not need to give similar answers to similar questions, and the possibility of asking unnecessary questions will be minimized.

[0058] Other items that are expected to have a correlation include fatigue and lethargy. When determining whether or not such a correlation exists, the correlation between the two items in question and whether a single patient has both symptoms should be understood beforehand. Then, for example, if the correlation coefficient between the two items is 0.7 or higher, it is judged that there is similarity, and the problems are combined.

[0059] This method of identifying correlations involves pre-existing medical data from healthcare institutions being stored as big data and then analyzed by artificial intelligence.

[0060] In the medical interview system according to one embodiment of the present invention, unnecessary questions are successfully omitted in this manner. Therefore, even if the total number of questions is small, ranging from 10 to 20, only the necessary questions are asked, and similar questions with similar relationships are not presented. As a result, even with a very small number of questions, a sufficient diagnosis can be made without compromising the accuracy of the medical interview used to calculate the diagnostic result.

[0061] Furthermore, with a relatively small number of questions, such as 10 to 20, the burden on the patient to input the information is minimal. Therefore, even elderly people and others who are presumed to be users of this AI-powered medical questionnaire system can undergo the questionnaire without significant burden, and the risk of input errors by the patient can be minimized.

[0062] Furthermore, the medical interview system according to one embodiment of the present invention is characterized in that the total number of questions that the artificial intelligence questioning unit asks, which are optimal for the diagnosis, is 10. This significantly reduces the burden on the patient compared to the 87 questions described in Non-Patent Literature 1 by Keio University et al.

[0063] Furthermore, since this medical questionnaire system consists of approximately 10 to 20 sub-questions, the medical questionnaire system according to one embodiment of the present invention does not thoroughly investigate a single symptom and strictly classify and identify the disease state. Rather, it is a medical questionnaire system that analyzes the patient's health condition from various perspectives, identifies the location of the cause of the discomfort, the name of the disease, etc., and classifies the disease state, particularly based on the perspective of Oriental medicine.

[0064] Furthermore, unlike conventional AI-based medical interviews using Western medicine, this AI-powered medical interview step can generate questions and conduct interviews targeting individuals ranging from those in good health to those experiencing minor ailments or the early stages of illness. Therefore, this AI-powered medical interview step is not limited to use in medical institutions; it can also be used, for example, during health checkups.

[0065] Furthermore, it can also be used by general users to check their own health status.

[0066] Furthermore, the questions presented in this AI-powered medical interview step utilize big data collected from past medical diagnostic information, which has been used to train the AI. The data incorporated as big data is not limited to conventional knowledge about disease conditions, but also includes knowledge of causal therapy based on traditional East Asian medicine. Therefore, this data also includes information on herbal medicines suitable for specific symptoms.

[0067] Furthermore, the content of subsequent questions will vary depending not only on the examinee's answers to the initial questions, but also on the examinee's personal information. This system works by having artificial intelligence make appropriate decisions for each question based on the examinee's personal information.

[0068] One example of this approach is that, if the patient is a woman in her 20s, the first question might be about menstruation. Furthermore, the content of subsequent questions would naturally differ depending on the patient's gender, age, physical condition, etc.

[0069] Furthermore, the questions in the AI-powered medical interview step will be as concise and to the point as possible, making them easily understandable to the patient.

[0070] Furthermore, as one aspect of the medical questionnaire screen, the left column 11 in Figure 4 displays the proposed question 14. The right column 12 in Figure 4 allows the patient to see the progress of their answers to the questions. Therefore, the patient can understand their own response status.

[0071] In the AI-powered medical interview step described above, the questioning unit 1 and AI questioning unit 2 of this AI-powered medical interview system are responsible for the questioning function, and the patient information acquisition unit 3 is responsible for acquiring patient information that the patient has answered.

[0072] <About the steps for determining your constitution> Next, we will discuss the constitution assessment steps (Steps S4 and S5). In the constitution assessment steps, the artificial intelligence determines and derives the constitution based on the answers to the questions entered in the previous AI medical interview step. The mechanism for determining the constitution by the AI ​​is to classify the constitution into seven categories based on the medical interview information. These seven categories are taken from the classification of constitutions in Oriental medicine and are: Qi deficiency, Qi stagnation, Qi reversal, Blood deficiency, Blood stasis, Water retention, and Body fluid deficiency.

[0073] Here, Qi deficiency, Qi stagnation, Qi reversal, Blood deficiency, Blood stasis, Water stagnation, and Body fluid deficiency are based on the Qi, Blood, and Body Fluid Differentiation of Traditional Chinese Medicine. These are categories used to classify health conditions by examining the state of Qi, Blood, and Body Fluids, which are the components of the human body.

[0074] To elaborate further, Qi refers to the energy that supports life activities, blood refers to what circulates throughout the body and provides nourishment, and water refers to bodily fluids involved in water metabolism and the immune system. The aforementioned Qi deficiency, Qi stagnation, Qi reversal, blood deficiency, blood stasis, water retention, and body fluid deficiency indicate that these Qi, blood, and water are insufficient (deficient) or stagnant.

[0075] In other words, Qi deficiency, Qi stagnation, and Qi reversal indicate states where the energy supporting vital activities is insufficient, stagnant, or circulating in reverse, respectively. Blood deficiency and blood stasis indicate states where there is insufficient blood or blood flow is stagnant, respectively. Water stagnation and fluid deficiency indicate states where bodily fluids are stagnant or deficient, respectively.

[0076] Next, we will describe the detailed mechanism of this constitution assessment. Figure 5 is a diagram showing the setting of provisional points related to constitution assessment. Note that Figure 5 is just one example showing some items and does not show all conditions such as "fatigue" or "qi deficiency". Also, the provisional points assigned to each item are only one form, and various setting scores are acceptable based on conventional Oriental medicine data.

[0077] As shown in Figure 5, for each of the categories of Qi deficiency, Qi stagnation, Qi reversal, Blood deficiency, Blood stasis, Water retention, and Body fluid deficiency (hereinafter referred to as Constitution Category 30), provisional points 32 have been assigned in advance for fatigue, loss of appetite, sleep, coldness, thirst, stiff shoulders, urge to urinate, constipation, excessive sweating, and menstrual symptoms (hereinafter referred to as Symptom Category 31) (Figure 5).

[0078] This provisional score is calculated in advance based on medical knowledge and is not changed based on the patient's constitution or their answers during the medical interview. The setting of this provisional score is determined by the relationship between each constitution and the associated symptoms.

[0079] Furthermore, one method for setting these provisional points is based on the Qi-Blood-Water Score devised by Terazawa Katsutoshi et al. (hereinafter referred to as Terazawa et al.), which is known as publicly available technology. The Qi-Blood-Water Score devised by Terazawa et al. is shown in Figure 6. As shown in Figure 6, provisional points are assigned to each item in each of the 30 constitutional categories, and to each item in each symptom category.

[0080] This provisional scoring system is based on knowledge derived from traditional East Asian medicine. Therefore, items with a higher degree of relevance will have higher provisional scores.

[0081] As shown in Figure 5, for each constitution in the constitution category 30, the total value of these provisional points is pre-set to be approximately 40, based on Figure 6. Note that the symptom category 31 in Figure 5 has 9 items, and new provisional points have been set for related content based on the provisional points in Figure 6. Therefore, the data in Figure 6 and the data in Figure 5 do not perfectly match.

[0082] Next, based on the questionnaire results, the artificial intelligence analysis unit 5 automatically makes a judgment for each item in the symptom category, such as fatigue and loss of appetite, for each constitution category, based on the answers to the questionnaire in the artificial intelligence questionnaire step, and calculates a point within the range of 0 to 1.

[0083] The way these points are allocated is, for example, for a question like "Do you have an appetite recently?", the symptoms of loss of appetite are assigned as follows: "Yes" = 0 points, "No" = 1 point, "Don't know" = 0.5 points, etc. This point assignment is automatically determined by the artificial intelligence analysis unit 5 based on conventional big data.

[0084] At this point, the total value directly relates to the probability of having that constitution; the closer to 0, the better the condition of the symptoms, and the closer to 1, the worse the condition of the symptoms. For each of the nine constitution categories, the sum of the values ​​obtained by multiplying each hypothetical point by each point is calculated. The category with the largest sum represents the patient's constitution.

[0085] Furthermore, if there are multiple questions on the same topic, points are assigned as a sum for all related questions. This point assignment mechanism is illustrated using Figure 4 as an example. For a question about dry mouth as shown in Figure 4, an answer of "Often, and more than 10 times a month" would be worth 1 point, an answer of "Often, and 1 to 5 times a month" would be worth 0.7 points, an answer of "Never" would be worth 0 points, and an answer of "Don't know" would be worth 0.3 points. The setting, calculation, and distribution of these points are all determined and set by artificial intelligence based on big data and the responses of the test takers.

[0086] Then, for each constitution, points are assigned to each item in the symptom category, and the resulting sum is converted to 100%. In this case, since each item in the constitution category is set to a maximum total score of approximately 40 points, the total score is multiplied by 7 / 3 to convert it to approximately 100%. This value converted to 100% represents the likelihood of each constitution of the patient and will be displayed as the diagnostic result in the subsequent output step.

[0087] Furthermore, the total score does not need to be 40 points. It is acceptable to expand on the aforementioned nine items, or change the distribution of provisional points, and then convert them to a 100% score.

[0088] In the constitution assessment step described in detail above, the artificial intelligence analysis unit 5 performs analysis based on the patient's information.

[0089] <About the Output Step> Next, following the AI-powered medical interview, an output step is provided to inform the patient of the most suitable herbal medicine and lifestyle habits. This output step informs the patient of the most suitable herbal medicine and lifestyle habits based on the results of the symptom diagnosis.

[0090] This output step is performed by the results display unit 6 after the artificial intelligence analysis unit 5 makes a decision based on the patient's response data in the artificial intelligence medical interview step.

[0091] Figure 7 shows one way in which the medical interview results are displayed in this output step. As shown in Figure 7, the display of the medical interview results includes a title field 40, a summary field 41, a field for diagnostic possibility in Oriental medicine 42, a field for disease name and diagnostic possibility 43, a field for herbal medicine and efficacy 44, a field for health maintenance methods tailored to one's constitution 45, a field for additional notes 46, a QR code field 47, and so on.

[0092] Furthermore, the format of the results display is not limited to the format shown in Figure 7. It is also acceptable to include photographs, diagrams, or flowcharts in addition to the text shown in Figure 7, to introduce the herbal medicine and health maintenance methods.

[0093] Furthermore, the content of the diagnostic results notification is not limited to these items. Specifically, it may include advising the patient to the most appropriate medical department, for example, by saying, "To receive a more detailed examination, please visit the XX department."

[0094] Let's discuss each item. First, in the title field 40, the Japanese and Chinese Artificial Intelligence Medical Questionnaire and its explanation will be written. It is acceptable to include a disclaimer such as, "The results of this questionnaire may vary depending on the individual's constitution and symptoms."

[0095] Next, let's discuss the summary section 41. This section displays a summary of the interview results. It is acceptable to add advice regarding lifestyle habits, such as "Try to eat a balanced diet and get moderate exercise," in this section.

[0096] Next, we will discuss the Oriental Medicine Diagnosis Possibility column 42. The Oriental Medicine Diagnosis Possibility column 42 displays the diagnostic possibility from an Oriental medicine perspective, based on the AI's diagnosis results. The column may include the primary and secondary diagnoses, and the probability of each diagnosis result may be displayed in percentage format.

[0097] Next, the disease name and diagnosis possibility column 43 will be described. In this column, the corresponding disease conditions based on traditional Chinese medicine are described. Also, the probabilities of each disease condition analyzed by the artificial intelligence analysis unit 5 according to the aforementioned provisional points and points are also noted. Then, the disease condition with the highest possibility is taken as the primary diagnosis, and the next highest possibility is displayed as the secondary diagnosis.

[0098] Next, the Chinese herbal medicine and efficacy column 44 will be described. In this column, the optimal Chinese herbal medicine for the disease condition based on the diagnosis result of artificial intelligence is displayed. As one aspect of this notation, for example, "Ho Chu Ekki To (Hochu-ekki-to), effective for general malaise and weakness" etc.

[0099] Also, in this column, it may be a mode of also noting the medication style such as the way of taking Chinese herbal medicine.

[0100] Next, the health preservation method column 45 according to constitution will be described. In this column, it may be a mode of describing advice on lifestyle habits such as diet and exercise based on the analyzed disease name and diagnosis possibility.

[0101] Next, the additional note column 46 will be described. In this column, cautions regarding the diagnosis result are described. As an example of this, "The above diagnosis result is only a general theory. Please consult a doctor or pharmacist and receive appropriate treatment." etc.

[0102] Next, the QR code column 47 will be described. In the QR code column 47, the contents such as the interview information and summary are displayed in the form of a QR code (registered trademark). Therefore, when the patient of this artificial intelligence interview system receives a diagnosis from medical staff such as a doctor or pharmacist after this interview, the medical staff can read this QR code and use it for their own interview.

[0103] Since the disease condition of the patient can be grasped in such a format, this interview system can reduce the burden of the medical work of medical staff. Also, it can lead to a reduction in the time for recording in the patient's medical record.

[0104] Furthermore, the content displayed on the QR code may include, but is not limited to, information from the medical interview, a summary, diagnostic possibility, illness and its diagnostic possibility, herbal medicine and its effects, and health maintenance methods. For example, information such as the patient's medical history, major illnesses in the past, and whether or not they are currently taking medication may be collected in advance during the medical interview and made available for reading via QR code.

[0105] In addition to the results of the medical interview in this output step, the results of this AI medical interview system may also be transcribed from voice input into text, and the AI ​​may perform a summarization process of the medical interview results before outputting them. This approach eliminates the effort of digitizing the medical interview results from information written on paper, and allows the interviewer to conduct the interview while directly viewing the patient's face. As a result, a high-quality medical interview can be conducted.

[0106] Here, the output method for the summary may be as follows: it may be displayed in a summary field on the screen of an electronic terminal such as a personal computer, as described above; or it may be as an audio output that reads the summary results aloud; or it may be as an output of electronic data on paper. With such methods, the preparation of paper medical records can be easily carried out.

[0107] In this case, one example of how artificial intelligence can summarize the results of a medical interview is to create the summary using SOAP summarization, which is a common method for recording information in electronic medical records.

[0108] Let's discuss SOAP summarization in detail in this case. S (subjective), or subjective information, refers to the information the patient is reporting. Specifically, this includes things like "I'm having trouble breathing" or "My stomach hurts."

[0109] Furthermore, O (objective) refers to objective information obtained from the diagnostic side through tests, diagnoses, etc. Specifically, this includes things like "blood pressure" and "pulse rate."

[0110] Furthermore, A (assessment), or evaluation, is the result of analyzing the symptoms based on the content of S and O mentioned earlier.

[0111] Furthermore, P (plan), or the plan, is the treatment strategy based on A mentioned above.

[0112] By performing SOAP summaries on such patient responses, healthcare professionals can easily understand the patient's symptoms.

[0113] Furthermore, the display of the optimal health status in the results display unit can be displayed with greater accuracy by converting the information from the doctor and patient during the consultation into audio text and performing a re-diagnosis based on the SOAP summary text.

[0114] In the output step described above, the results display unit 6, which has a display function, displays the results of the medical interview.

[0115] As described above, the artificial intelligence-based medical interview according to one embodiment of the present invention can appropriately diagnose and treat minor ailments that are not considered diseases. Furthermore, it can help reduce the workload of medical professionals and, as a result, provide high-quality medical interviews.

[0116] [3. System comprising artificial intelligence medical interview according to one embodiment of the present invention] Next, we will describe the system comprising an artificial intelligence medical interview according to one embodiment of the present invention. As shown in Figure 8, the artificial intelligence medical interview system according to one embodiment of the present invention consists of an administrator terminal 51, a public communication network 52, a patient terminal 53, a medical interview server 54, and the like.

[0117] Here, the administrator terminal 51 is an administrator terminal for managing the artificial intelligence medical interview system. This administrator terminal 51 is used by the physician, and the results can be viewed only on the administrator terminal. The patient terminal 53 is a terminal used by patients of this artificial intelligence medical interview system when they actually visit a clinic, and includes electronic devices such as personal computers, tablet devices, and smartphones.

[0118] As shown in Figure 8, the patient accesses the medical interview server 54 via the public communication network 52 through their patient terminal 53. They can then answer questions and receive medical consultation using this artificial intelligence medical interview system.

[0119] The medical interview server 54 also includes a control unit, a memory unit, an analysis unit, an artificial intelligence analysis unit, an interface unit, and the like.

[0120] Furthermore, when running this artificial intelligence medical interview system on a single computer or other electronic device without using a public communication network, it is preferable to run it on a computer or other electronic device equipped with a GPU (Graphics Processing Units). However, it can also be run on a computer or other electronic device equipped with a CPU (Central Processing Units).

[0121] [4. Effects and benefits of a medical interview system according to one embodiment of the present invention] A summary of the results obtained by the artificial intelligence medical interview system according to one embodiment of the present invention can be used by physicians during counseling or by pharmacists during medication guidance.

[0122] Furthermore, the artificial intelligence medical interview system according to one embodiment of the present invention can be used, for example, in the medical examination of patients in hospitals.

[0123] Alternatively, instead of treating patients who have already developed symptoms, it could be used in general corporate health checkups, for example, on people in the early stages of illness, to prevent the worsening of their condition through the use of Oriental medicine.

[0124] Furthermore, the artificial intelligence medical interview system according to one embodiment of the present invention is configured to have a high processing speed by the artificial intelligence analysis unit, and can derive analysis results within 10 seconds of the patient finishing their answers.

[0125] Furthermore, as mentioned above, the artificial intelligence medical interview system according to one embodiment of the present invention has a small number of questions, ranging from 10 to 20, yet it carefully selects the necessary questions and integrates similar questions that have a correlation. As a result, it has a small number of questions and reduces the burden on the patient, while still maintaining sufficient accuracy in the medical interview. In this respect, the artificial intelligence medical interview system according to one embodiment of the present invention is highly convenient.

[0126] Furthermore, the artificial intelligence medical interview system according to one embodiment of the present invention may be configured to read out the interview results by voice. In this case, the text output from the aforementioned interview results may be read out by the artificial intelligence.

[0127] The embodiments described above are merely examples. Inventions that do not differ significantly from these embodiments, such as minor modifications to the configuration, shall be considered to fall within the same technical scope as the present invention. [Explanation of Symbols]

[0128] 1. Questioning Section 2. Artificial Intelligence Questioning Section 3. Patient Information Acquisition Department 4 Analysis section 5. Artificial Intelligence Analysis Department 6 Display section 14 Questions 30 Body Type Categories 31 Symptom Categories 32 Provisional points 51 Administrator terminal 52 Public telecommunications network 53 Patient terminal 54 Medical Questionnaire Server

Claims

1. An artificial intelligence medical interview system that performs a constitution assessment based on the knowledge of Oriental medicine and generates the next questions based on the patient's answers, An inquiry unit that asks the patient questions regarding the patient's personal information and health condition, A patient information acquisition unit acquires patient information obtained by the patient in response to the question issued by the questioning unit, The artificial intelligence analyzes the recipient information acquired by the patient information acquisition unit and asks the most appropriate questions for diagnosis. Based on the patient information acquired by the patient information acquisition unit and prior knowledge of Oriental medicine, the artificial intelligence analyzes information regarding health maintenance methods and prescriptions of herbal medicines that are suitable for the patient's constitution. A results display unit that displays the medical interview results analyzed by the aforementioned artificial intelligence analysis unit, To be prepared A personalized artificial intelligence medical interview and diagnosis system characterized by the following features.

2. The total number of questions that the artificial intelligence questioning unit asks, which is optimal for the diagnosis, is within the range of 10 to 20 questions. Characterized by, The personalized artificial intelligence medical interview and diagnosis system according to claim 1.

3. The questioning unit and the artificial intelligence questioning unit modify the content of the questions according to each patient's gender, age, and physical constitution, thereby deriving the optimal health condition tailored to each patient. Characterized by, The personalized artificial intelligence medical interview and diagnosis system according to claim 2.

4. The results display unit that displays the optimal health status analyzes the recipient information within approximately 10 seconds after the recipient has finished answering all the questions, and derives the optimal health status and displays the results. Characterized by, The personalized artificial intelligence medical interview and diagnosis system according to claim 3.

5. The display of the optimal health status in the results display unit can be made highly accurate by reflecting the SOAP summary, which summarizes the doctor's medical results, and performing a re-diagnosis. A personalized artificial intelligence medical interview and diagnosis system according to claim 4, characterized by the above.

6. An AI-powered medical interview method that uses artificial intelligence to conduct interviews based on traditional Chinese medicine. A questioning step in which questions are asked of the patient about the patient's health condition, The patient answers the aforementioned questions, providing patient information in a response step, An output step that displays the analysis results regarding the health status based on the patient information of the aforementioned patient, It consists of, The aforementioned questioning step is initiated by artificial intelligence, which determines the content and number of questions based on the patient information, Furthermore, the output step involves artificial intelligence analyzing the patient information and displaying the analysis results. A personalized AI-powered medical interview method characterized by the following features.

7. A program for causing a computer to execute the personalized artificial intelligence medical interview method described in claim 6.

Images