Program, information processing method, and information processing apparatus.

Abstract

The present invention provides a program, information processing method, and information processing device that output responses to a patient using a language model based on assessment information from the attending physician and related healthcare professionals. [Solution] In a collective intelligence formation system for physicians, the attending physician terminal obtains patient information from the server, receives assessment information from selected medical professionals and the attending physician according to the obtained patient information, obtains second assessment information related to the assessment information from the responsible physician, and outputs a response to the patient by providing the obtained assessment information and the second assessment information to a language model.

Description

Technical Field

[0001] The present invention relates to a program, an information processing method, and an information processing apparatus.

Background Art

[0002] In recent years, the development of technologies for providing treatment information to patients has been actively promoted. For example, Patent Document 1 discloses a medical support system that determines the appropriateness of the current treatment policy for a patient based on guidelines related to the patient's treatment and patient information, and generates recommended treatment information for the patient.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, the invention according to Patent Document 1 has a problem that it cannot output a response (such as treatment information) to the patient by a language model based on the respective assessment information of the attending physician and related medical staff.

[0005] In one aspect, there is provided a program or the like capable of outputting a response to the patient by a language model based on the respective assessment information from the attending physician and related medical staff.

Means for Solving the Problems

[0006] The program, relating to one aspect, acquires patient information of a patient, receives assessment information from healthcare professionals and the attending physician selected according to the acquired patient information, acquires second assessment information related to the assessment information from the responsible physician, and provides the acquired assessment information and the second assessment information to a language model, thereby causing the computer to execute a process that outputs a response to the patient. [Effects of the Invention]

[0007] In one respect, it becomes possible to output responses for the patient using a language model, based on assessment information from the attending physician and other relevant healthcare professionals. [Brief explanation of the drawing]

[0008] [Figure 1] This is an explanatory diagram illustrating the overview of a collective intelligence formation system among physicians. [Figure 2] This is a block diagram showing an example server configuration. [Figure 3] This is an explanatory diagram showing an example of the record layout for the patient database and physician database. [Figure 4] This is an explanatory diagram showing an example of the record layout for the response database and knowledge database. [Figure 5] This block shows examples of the configurations for the attending physician's terminal and the specialist's terminal. [Figure 6] This is a block diagram showing an example configuration for the physician's terminal. [Figure 7] This is an explanatory diagram illustrating the operation of a collective intelligence formation system among physicians. [Figure 8] This flowchart shows the processing procedure when a language model outputs a response to a patient. [Figure 9] This flowchart shows the processing procedure when a language model outputs a response to a patient. [Figure 10] This flowchart shows the processing steps involved in retraining a language model. [Figure 11]It is an explanatory diagram for explaining the operation of the doctor's collective knowledge formation system in Embodiment 2. [Figure 12] It is a flowchart showing the processing procedure when outputting the second response to the patient by the language model. [Figure 13] It is a flowchart showing the processing procedure when outputting the second response to the patient by the language model. [Figure 14] It is an explanatory diagram for explaining the process of determining the consideration according to the provision of the assessment information. [Figure 15] It is a flowchart showing the processing procedure when determining the consideration according to the provision of the assessment information.

Mode for Carrying Out the Invention

[0009] Hereinafter, the present invention will be described in detail based on the drawings showing its embodiments.

[0010] (Embodiment 1) Embodiment 1 relates to a form in which a response to the patient is output using a language model based on the assessment information from each of the medical staff selected according to the patient information of the patient and the attending physician of the patient. Note that the assessment information will be described later.

[0011] FIG. 1 is an explanatory diagram showing the outline of the doctor's collective knowledge formation system. The system of this embodiment includes an information processing device 1, information processing terminals 2, 3, and 4, and each device transmits and receives information via a network N such as the Internet.

[0012] The information processing device 1 is an information processing device that performs processing, storage, and transmission / reception of various information. The information processing device 1 is, for example, a server device, a personal computer, or a general-purpose tablet PC (personal computer), etc. In this embodiment, the information processing device 1 is assumed to be a server device, and hereinafter it will be read as server 1 for simplicity.

[0013] The information processing terminal 2 is a terminal device for a primary care physician to receive and transmit a patient's patient information, receive and transmit the primary care physician assessment information of the patient, and receive and display the response to the patient output from the language model. The primary care physician (attending physician) in this embodiment is a physician directly responsible for the treatment of the patient. The primary care physician is responsible for grasping the patient's health condition or medical history, and for diagnosing, determining the treatment plan, implementing treatment, or observing the progress.

[0014] The information processing terminal 2 is an information processing device such as a smartphone, a mobile phone, a wearable device such as a smartwatch, a tablet, or a personal computer terminal. Hereinafter, for the sake of brevity, the information processing terminal 2 will be read as the primary care physician terminal 2.

[0015] The information processing terminal 3 is a terminal device for medical staff to receive and transmit assessment information of medical staff selected according to patient information. The medical staff in this embodiment is a specialist who is involved in the patient's medical treatment but is not the direct primary care physician. The medical staff is a physician in charge of a different medical department from the primary care physician. In addition, the medical staff may include nurses, pharmacists, physical therapists, clinical laboratory technicians, radiological technicians, or psychologists. Hereinafter, for the sake of brevity, the medical staff will be read as the specialist.

[0016] The information processing terminal 3 is an information processing device such as a smartphone, a mobile phone, a wearable device such as a smartwatch, a tablet, or a personal computer terminal. Hereinafter, for the sake of brevity, the information processing terminal 3 will be read as the specialist terminal 3.

[0017] The information processing terminal 4 is a terminal device for a responsible physician to receive the respective assessment information from the primary care physician and the specialist, and to receive and transmit the second assessment information regarding the assessment information. The second assessment information will be described later. The responsible physician in this embodiment is a physician or a clerical staff member who performs overall management, information collection, aggregation, and coordination. The responsible physician may be a primary care physician or a specialist.

[0018] Information processing terminal 4 is an information processing device such as a smartphone, mobile phone, smartwatch or other wearable device, tablet, or personal computer terminal. For simplicity, information processing terminal 4 will be read as responsible physician terminal 4 below.

[0019] Typically, patients with multiple conditions, such as lifestyle-related diseases, cardiovascular diseases, tumors, and mental illnesses, require multiple visits to specialists. However, this requires patients to spend a significant amount of time and incurs considerable costs. Furthermore, communication between specialists is often ineffective, potentially preventing them from providing the patient with the most appropriate treatment plan.

[0020] To solve these problems, this embodiment provides a collective intelligence system for physicians that uses a language model to perform an individualized physician (attending physician or specialist) comprehensive evaluation (diagnosis) for each patient, based on assessment information from specialists and attending physicians selected according to patient information.

[0021] This system integrates evaluations from the attending physician and multiple specialists, enabling the rapid and effective provision of the most suitable treatment plan for each patient. This not only reduces the time, effort, and expense patients would incur from consulting multiple specialists, but also facilitates smooth information sharing among specialists, ensuring the provision of the most appropriate treatment plan for each patient.

[0022] In this embodiment, Server 1 acquires patient information from the attending physician terminal 2. Server 1 accepts a specialist selected according to the acquired patient information via the attending physician terminal 2. Server 1 accepts assessment information from the selected specialist via the specialist terminal 3. Server 1 accepts assessment information from the attending physician via the attending physician terminal 2.

[0023] Server 1 transmits assessment information from the receiving specialist and attending physician to the responsible physician's terminal 4. Server 1 obtains second assessment information related to the assessment information from the responsible physician's terminal 4. Server 1 outputs a response for the patient by providing the obtained assessment information and second assessment information to the language model 151, which will be described later.

[0024] Figure 2 is a block diagram showing an example configuration of Server 1. Server 1 includes a control unit 11, a storage unit 12, a communication unit 13, a reading unit 14, and a large-capacity storage unit 15. Each component is connected by bus B.

[0025] The control unit 11 includes an arithmetic processing unit such as a CPU (Central Processing Unit), MPU (Micro-Processing Unit), GPU (Graphics Processing Unit), FPGA (Field Programmable Gate Array), DSP (Digital Signal Processor), or quantum processor. The control unit 11 reads and executes a control program 1P (program product) stored in the storage unit 12, thereby performing various information processing or control processing related to the server 1.

[0026] Furthermore, the control program 1P can be deployed on a single computer, at a single site, or distributed across multiple sites and run on multiple computers interconnected by a communication network.

[0027] In Figure 2, the control unit 11 is described as a single processor, but it may be a multi-processor system. Furthermore, the control unit 11 may perform various information processing or control processing on the same processor within the server 1, or it may perform these processes on different processors within the server 1.

[0028] The memory unit 12 includes memory elements such as RAM (Random Access Memory) and ROM (Read Only Memory), and stores control programs 1P or data necessary for the control unit 11 to execute processing. The memory unit 12 also temporarily stores data necessary for the control unit 11 to execute arithmetic processing. The communication unit 13 is a communication module for performing communication-related processing, and transmits and receives information between the attending physician terminal 2, specialist terminal 3, or responsible physician terminal 4, etc., via the network N.

[0029] The reading unit 14 reads a portable storage medium 1a, including a CD (Compact Disc)-ROM or DVD (Digital Versatile Disc)-ROM. The control unit 11 may read the control program 1P from the portable storage medium 1a via the reading unit 14 and store it in the large-capacity storage unit 15. Alternatively, the control unit 11 may download the control program 1P from another computer via a network N or the like and store it in the large-capacity storage unit 15. Furthermore, the control unit 11 may also read the control program 1P from the semiconductor memory 1b.

[0030] The large-capacity storage unit 15 includes a recording medium such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive). The large-capacity storage unit 15 includes a language model 151, a patient database 152, a physician database 153, a response database 154, and a knowledge database 155.

[0031] Language model 151 is a language generation model constructed by pre-training on a large amount of text data (dataset). For language model 151, large-scale language models (LLMs) such as Transformer, ALBERT (A Lite BERT), GPT (Generative Pre-trained Transformer)-2, GPT-3, GPT-4, LLaVA (Large Language and Vision Assistant), MiniGPT-4, or BERT (Bidirectional Encoder Representations from Transformers) can be used.

[0032] The patient database 152 stores information about patients. The physician database 153 stores information about physicians and specialists. The response database 154 stores responses to patients output from the language model 151. The knowledge database 155 stores medical terminology and other medical-related information.

[0033] In this embodiment, the storage unit 12 and the large-capacity storage unit 15 may be configured as a single storage device. Furthermore, the large-capacity storage unit 15 may be composed of multiple storage devices. Moreover, the large-capacity storage unit 15 may be an external storage device connected to the server 1.

[0034] Server 1 may perform various information processing and control processes as a single computer, or it may be performed in a distributed manner across multiple computers. Furthermore, Server 1 may be implemented using multiple virtual machines located within a single server, or it may be implemented using a cloud server.

[0035] Figure 3 is an explanatory diagram showing an example of the record layout for patient DB152 and physician DB153. Patient DB152 includes columns for Patient ID, Name, Age, Gender, Subject / Predicate, Current Illness, Past Medical History, Medication, Physical Examination, Laboratory Data, Attending Physician, and Specialist. The Patient ID column stores a uniquely identifiable patient ID to identify each patient. The Patient ID may also be the patient's medical card number. The Name column stores the patient's name. The Age column stores the patient's age. The Gender column stores the patient's gender.

[0036] The subject-predicate column records the symptoms and problems the patient primarily reports. The present illness column records detailed information about the illness or condition the patient is currently experiencing. The past medical history column records the patient's past illnesses, surgeries, or treatments. The medication column records the names, dosages, or methods of administration of medications the patient is currently taking. The physical examination column records the results of the physical examination, such as the patient's blood pressure, pulse, temperature, breath sounds, or heart sounds.

[0037] The "Test Data" column stores test data for various tests performed on the patient (blood tests, imaging tests, or urine tests, etc.). The "Attending Physician" column stores the name or physician ID of the attending physician responsible for the patient's care. The "Specialist" column stores the name or physician ID of a specialist in a specific field (e.g., internal medicine, surgery, or neurology) selected according to the patient's information.

[0038] Doctor DB153 includes columns for Doctor ID, Medical Specialty, and Name. The Doctor ID column stores the unique ID of each doctor to identify them. The Medical Specialty column stores the doctor's medical specialty. Specialties include internal medicine, surgery, cardiology, nephrology, diabetology, or neurology, etc. The Name column stores the doctor's name.

[0039] Figure 4 is an explanatory diagram showing an example of the record layout for Response DB154 and Knowledge DB155. Response DB154 includes columns for Patient ID, Attending Physician Assessment, Specialist Assessment, Second Assessment, Response, Consultation Date and Time, and Edited Response. The Patient ID column stores the patient ID to identify the patient. The Attending Physician Assessment column stores the assessment information of the patient's attending physician. The Specialist Assessment column stores the assessment information of the specialist corresponding to the patient.

[0040] The second assessment column stores second assessment information regarding the attending physician's assessment information and the specialist's assessment information. The response column stores the responses to the patient output from the language model 151. The responses to the patient include the purpose of treatment, the method of treatment (e.g., drug therapy, surgery, or rehabilitation), the method of progression, or the schedule. The consultation date and time column stores information about the date and time the consultation took place. The edited response column stores the response after it has been edited by the responsible physician.

[0041] Knowledge DB155 includes columns for Knowledge ID, Medical Terminology, Definition, Category, Related Terminology, Description, Treatment, Notes, and Reference Links. The Knowledge ID stores the uniquely identified ID of each knowledge item to distinguish it. The Medical Terminology column stores specialized terms related to medical or therapeutic use (e.g., diabetes, antibiotics, or myocardial infarction).

[0042] The Definition column stores the meaning and interpretation of medical terms (for example, "Diabetes is a disease characterized by chronically high blood sugar levels"). The Category column stores the category to which the medical term belongs. Categories include, but are not limited to, "Disease," "Symptoms," "Treatment," "Medication," "Test," and "Surgery." The Related Terms column stores synonyms or similar terms related to the medical term.

[0043] The explanation column stores detailed explanations of medical terms (e.g., "Treatment methods for diabetes include diet therapy, exercise therapy, and drug therapy..."). The treatment column stores treatments, preventive measures, approaches, drug names (e.g., "Diabetes medication: metformin"), surgeries, or examples of use in a clinical setting (e.g., "The patient was diagnosed with diabetes and started insulin therapy").

[0044] The "Notes" column stores points to note, side effects, contraindications (e.g., "Metformin is contraindicated in patients with renal impairment"), or warnings related to the treatment or management of medical terms. The "Reference Links" column stores reference links to medical guidelines, research papers, or books related to the medical terms or treatments. Links may be, for example, URLs (Uniform Resource Locators).

[0045] The storage configurations described above for each database are merely examples; other storage configurations are also acceptable as long as the relationships between the data are maintained.

[0046] Figure 5 is a block diagram showing an example configuration of the attending physician terminal 2 and the specialist physician terminal 3. The attending physician terminal 2 includes a control unit 21, a storage unit 22, a communication unit 23, an input unit 24, and a display unit 25.

[0047] The control unit 21 includes a processing unit such as a CPU or MPU, and performs various information processing and control processing related to the attending physician terminal 2 by reading and executing the control program 2P (program product) stored in the storage unit 22.

[0048] In Figure 5, the control unit 21 is described as a single processor, but it may be a multiprocessor. Furthermore, the control unit 21 may perform various information processing or control processing using the same processor within the physician terminal 2, or it may perform these processes using different processors within the physician terminal 2.

[0049] The storage unit 22 includes memory elements such as RAM or ROM, and stores the control program 2P or data necessary for the control unit 21 to execute processing. The storage unit 22 also temporarily stores data necessary for the control unit 21 to execute arithmetic processing.

[0050] The communication unit 23 is a communication module for performing communication-related processing and sends and receives information with the server 1, etc., via the network N. The input unit 24 may be a keyboard, mouse, or a touch panel integrated with the display unit 25. The display unit 25 is a liquid crystal display or an organic EL (electroluminescence) display, etc., and displays various information according to the instructions of the control unit 21.

[0051] The specialist physician terminal 3 includes a control unit 31, a storage unit 32, a communication unit 33, an input unit 34, and a display unit 35. The configuration of the specialist physician terminal 3 is the same as that of the attending physician terminal 2, so a detailed explanation is omitted.

[0052] Figure 6 is a block diagram showing an example configuration of the responsible physician's terminal 4. The responsible physician's terminal 4 includes a control unit 41, a storage unit 42, a communication unit 43, an input unit 44, and a display unit 45. Note that the configuration of the responsible physician's terminal 4 is the same as that of the attending physician's terminal 2, so a detailed explanation is omitted.

[0053] Figure 7 is an explanatory diagram illustrating the operation of the collective intelligence formation system for physicians. Server 1 retrieves specialist information for multiple specialists from physician DB 153 based on physician IDs. The specialist information includes the specialist's medical specialty and name. Server 1 transmits the retrieved specialist information to the attending physician terminal 2. The attending physician terminal 2 receives the specialist information transmitted from Server 1 and displays it on its screen.

[0054] The attending physician's terminal 2 acquires patient information for the target patient. Patient information includes, but is not limited to, basic information (name, age, and gender, etc.), subject (symptoms and problems reported, etc.), present illness, past medical history, physical examination data (blood pressure, pulse, temperature, breath sounds or heart sounds, etc.), laboratory data (blood tests, urine tests, imaging tests or electrocardiograms, etc.), and medications, etc.

[0055] Patient information may also include family history, allergy information, lifestyle (smoking, drinking, exercise habits, or diet, etc.), living environment, occupation, or psychological state (e.g., stress or presence of mental illness).

[0056] Specifically, the attending physician terminal 2 obtains basic information, present illness history, past medical history, and the main statement the patient made at the initial or follow-up visit from the patient. If basic information, present illness history, past medical history, family history, and allergy information are already stored (registered) in the patient database 152, the attending physician terminal 2 retrieves the patient's basic information, present illness history, past medical history, family history, and allergy information from the patient database 152 based on the patient ID.

[0057] The attending physician terminal 2 accepts input of medication information, including medications the patient is currently taking or medications prescribed by the attending physician. The attending physician terminal 2 acquires physical examination data (blood pressure, pulse, body temperature, breath sounds or heart sounds, etc.) and laboratory data (blood tests, urine tests, imaging tests or electrocardiograms, etc.) from the relevant testing equipment.

[0058] Furthermore, the process of acquiring patient information is not limited to the methods described above. For example, a medical interview conducted by a nurse or a questionnaire survey using a question-and-answer format may also be used. The questionnaire may include questions such as, "Do you drink alcohol regularly? How much do you consume?", "Have you recently experienced chest pain or shortness of breath?", or "Have you noticed any sudden changes in your weight?".

[0059] For example, the patient information obtained is Name: Yamada XX Age: 72 Gender: Male Chief complaint: Recent fatigue and shortness of breath, decreased urine output. Present illness: A 72-year-old male. Diagnosed with diabetes 15 years ago... Medical history: Pancreatic cancer High blood pressure (20 years) Diabetes (15 years) ××× Medications currently in use: Warfarin 5mg Amlodipine 5mg Physical examination data: Height: 170cm Weight: 80kg (gained 2kg in one month) Blood pressure: 140 / 85 mmHg (slightly high) Heart rate: 75 bpm (at rest) ××× Test data: Creatinine: 1.3 mg / dL (Reference range: 0.6-1.2 mg / dL) Electrocardiogram: Mild ST depression (possibly due to renal failure or electrolyte abnormalities) Abdominal ultrasound: Renal atrophy and thinning of the renal cortex were observed. It could also be "xxx".

[0060] The attending physician terminal 2 allows the attending physician to select multiple specialists based on the acquired patient information. For example, specialists in cardiology, nephrology, and diabetology can be selected. It is also possible to select a single specialist, rather than multiple specialists.

[0061] Furthermore, based on patient information, multiple specialists can be selected using a language model. For example, the attending physician terminal 2 may input patient information and a prompt including a command to select multiple specialists based on that patient information into the language model, and output multiple specialists. The selection command may be, for example, "Based on the patient information, refer to the physician database and output multiple specialists suitable for this patient."

[0062] The attending physician terminal 2 transmits the acquired patient information, the attending physician's ID, and the IDs of the selected specialists to the server 1. The server 1 receives the patient information, the attending physician's ID, and the specialists' IDs transmitted from the attending physician terminal 2.

[0063] Server 1 stores the received patient information, the attending physician's ID, and the specialist physician's ID in the patient database 152. Specifically, Server 1 assigns a patient ID to the received patient information. Server 1 stores the basic information, subject and predicate, present illness history, past medical history, medication information, physical examination data, test data, the attending physician's ID, and the specialist physician's ID as a single record in the patient database 152, associating them with the assigned patient ID.

[0064] Based on the patient's information, the system can obtain the attending physician's assessment information and the specialist assessment information of each specialist through dialogue between the attending physician and selected specialists. The attending physician's assessment information and the specialist assessment information are obtained through a chat room for dialogue between the attending physician and the specialists.

[0065] Specifically, the attending physician terminal 2 or the specialist physician terminal 3 sends a request to create a chat room to server 1. The request to create a chat room includes information about the participating physicians (attending physician or specialist, etc.) (physician ID and IP address, etc.), and settings related to the chat room (e.g., private or public, room name, start time and end time).

[0066] Server 1 receives a chat room creation request sent from the attending physician terminal 2 or the specialist terminal 3. In response to the received chat room creation request, Server 1 connects to, for example, an online meeting distribution service and creates the chat room.

[0067] The online meeting delivery service may be a so-called online meeting tool such as Google Meet®, ZOOM®, or MICROSOFT TEAMS®. Server 1 may also build a chat room using the platform's API (Application Programming Interface) or SDK (Software Development Kit).

[0068] Server 1 sends a notification containing the meeting link for the generated chat room to both the attending physician's terminal 2 and the target specialist's terminal 3. The attending physician's terminal 2 and each specialist's terminal 3 receive the notification sent from Server 1. The attending physician's terminal 2 and each specialist's terminal 3 connect to the chat room according to the received notification.

[0069] Server 1 stores the content of conversations between the attending physician and multiple specialists through the chat room in the memory unit 12 or the large-capacity memory unit 15. The conversation content includes information such as the patient's symptoms, treatment history, current medical history, test results, and additional tests or treatment plans.

[0070] For example, the content of the dialogue is “Attending physician: Here's a case study. This is one of my patients. Please evaluate the situation from your own perspective, suggest areas for improvement, and provide your outlook for the future. Here are some questions the patient has: "If my kidney function deteriorates further, will I need to start dialysis?" "Is my current diabetes treatment sufficient? Do I need to change my medication?" ××× Diabetes Internal Medicine (Specialist): This is XX from the Diabetes Department at XXX Clinic. • On-site evaluation HbA1c, dietary intake xxx • Suggestions for improvement Discontinuation of metformin • Future outlook Insulin is adjusted according to the target HbA1c and blood glucose levels. Nephrologist (specialist): This is XX from the Nephrology Department at XXX Clinic. • On-site evaluation ××× • Suggestions for improvement ××× • Future outlook ××× Cardiology (Specialist): This is XX from the Cardiology Department at XXX Hospital. • On-site evaluation ××× • Suggestions for improvement ××× • Future outlook ××× Attending physician: Thank you so much, Doctor. "XXX" is also acceptable.

[0071] Server 1 acquires the dialogue content obtained in real time as attending physician assessment information and assessment information from multiple specialists. The assessment information includes, but is not limited to, the patient's medical history (chief complaint, present illness, past medical history, family history or allergy history, etc.), physical examination data (blood pressure, pulse, temperature, breath sounds or heart sounds, etc.), laboratory data (blood tests, urine tests, imaging tests or electrocardiograms, etc.), diagnosis (physician's assessment of the current illness or symptoms), risks (risks of surgery, possibility of complications or side effects of treatment, etc.), or treatment plan (e.g., treatment methods or interventions for the patient).

[0072] Furthermore, if the dialogue content is lengthy, assessment information can be extracted from the dialogue. For example, Server 1 filters the acquired dialogue content (patient symptoms, diagnosis, test results, treatment plan, or exchange of opinions between doctors, etc.) and extracts information related to the patient's evaluation and treatment as assessment information from the statements of the attending physician and multiple specialists.

[0073] Alternatively, a language model can be used to extract attending physician assessment information and specialist assessment information. For example, Server 1 may input the acquired dialogue content and prompts containing instructions to extract attending physician assessment information and specialist assessment information from the dialogue content into the language model, and output the attending physician assessment information and each specialist assessment information.

[0074] In this embodiment, the process of obtaining the attending physician's assessment information and the assessment information of multiple specialists through a chat room has been described, but this is not the only method. For example, the attending physician and multiple specialists may obtain assessment information using telephone, SNS (Social Networking Service), or email. SNS includes LINE®, Twitter®, Instagram®, and Facebook®, etc.

[0075] As an example, the attending physician assessment information included in the acquired conversation content is: "Medical history:" A 72-year-old male with a 15-year history of diabetes, a 2-year history of pancreatic cancer, and progressively worsening kidney function. Physical examination data: Blood pressure: 140 / 90 mmHg Pulse: 88 bpm, regular pulse ××× Test data: Blood sugar level: slightly elevated (HbA1c 8.0%) Creatinine level: 2.5 mg / dL (Kidney function is progressively deteriorating) ××× diagnosis: Chronic kidney disease (stages 3 to 4) Diabetes (inadequate HbA1c control) ××× risk: Restrictions on drug selection due to impaired renal function ××× Treatment plan: It could also be a "re-evaluation of diabetes treatment (changes in insulin therapy, review of oral hypoglycemic agents)."

[0076] Furthermore, as an example, the specialist assessment information of the cardiologist included in the acquired conversation content is: "Medical history:" The patient has a 20-year history of hypertension and a 1-year history of TAVI (transcatheter aortic valve implantation). Although the TAVI for aortic stenosis was successful, signs of heart failure (shortness of breath, edema) suggest a problem with cardiac function. Physical examination data: Blood pressure: 140 / 90 mmHg Pulse: 88 bpm, regular pulse Edema: Edema is present in both lower extremities, suggesting it may be related to a heart condition. ××× Test data: Electrocardiogram: No abnormalities found. Echocardiography: LVEF (left ventricular ejection fraction) needs to be checked. ××× diagnosis: The strain on the heart caused by high blood pressure Heart failure (possible, especially increased load due to impaired renal function) risk: Progression of heart failure (worsening of edema and shortness of breath) Impact on the heart due to worsening kidney function (risk of renal cardiopathy) Treatment plan: Evaluation of cardiac function through echocardiography Drug therapy to manage heart failure (use of ACE inhibitors or ARBs) It could also be "management of edema (use of diuretics)."

[0077] Furthermore, as an example, the specialist assessment information of nephrologists included in the acquired conversation content is: "Medical history:" Chronic kidney disease (stages 3-4) caused by diabetes and hypertension ××× Physical examination data: Blood pressure: 140 / 90 mmHg Edema: Edema in both lower extremities, suspected to be caused by impaired renal function. Abdominal examination: No particular abnormalities. ××× Test data: Creatinine: 2.5 mg / dL Urine test: Proteinuria (high urinary protein / creatinine ratio) Blood test: K+ 4.2 mmol / L ××× diagnosis: Chronic kidney disease (stages 3-4) Kidney dysfunction due to diabetes and hypertension risk: Further deterioration of kidney function Increased burden on the kidneys due to poor management of high blood pressure and diabetes Treatment plan: Renal function monitoring (regular checks of creatinine and eGFR) It may also be "renal protective medication (continued use of ACE inhibitors or ARBs)."

[0078] Furthermore, as an example, the specialist assessment information of the diabetologist included in the acquired conversation content is: "Medical history:" 15 years since my diabetes diagnosis... Physical examination data: Blood pressure: 140 / 90 mmHg Edema: Edema in both lower extremities, suspected to be caused by impaired renal function. Abdominal examination: No particular abnormalities. ××× Test data: Blood sugar level: slightly elevated (HbA1c 8.0%) Blood pressure: 140 / 90 mmHg Pulse rate: 88 bpm ××× diagnosis: Diabetes mellitus (HbA1c 8.0%, poorly controlled due to impaired renal function) Chronic kidney disease (stages 3-4) risk: Increased risk of hypoglycemia Increased proteinuria further impairs kidney function. Treatment plan: Discontinuing metformin, and introducing SGLT2 inhibitors or DPP-4 inhibitors. It could also be described as "improving lifestyle habits (diet and exercise therapy)."

[0079] Server 1 transmits the conversation content, including the attending physician's assessment information and assessment information from multiple specialists, to the responsible physician's terminal 4. The responsible physician's terminal 4 receives the conversation content transmitted from Server 1. The responsible physician's terminal 4 displays the received conversation content on its screen.

[0080] The responsible physician's terminal 4 accepts input for the response format and the command to generate a summary. The response format is, for example, Output format (no need for "desu" or "masu") Patient summary (non-bulleted format) Current Situation Assessment Possible interventions Future outlook and treatment plan It could also be "Answers to patients' questions."

[0081] Furthermore, the command to generate a summary is a command to generate summaries of the attending physician's assessment information and the assessment information of multiple specialists. For example, the command to generate a summary could be: "Please summarize in detail the contents of the chat between specialists regarding the following case, focusing on the medically important points. xxx."

[0082] The responsible physician terminal 4 receives input from the responsible physician regarding the attending physician's assessment information and the assessment information of multiple specialists contained in the received conversation. The second assessment information is the assessment information that should be included in the response to the patient from the language model 151.

[0083] As an example, the second assessment information from the responsible physician is: "We also offer this treatment on an outpatient basis. We will likely start with 4 units of rapid-acting insulin and 6 units of slow-acting insulin. The calculation is done as: Fasting C-peptide / Fasting blood glucose × 100. A value of 0.66 < 0.8 is interpreted as insulin-dependent. It is acceptable to say that administering SGLT2 inhibitors to insulin-dependent patients carries a risk of euglycemic ketoacidosis.

[0084] The responsible physician's terminal 4 transmits the format of the received response, the command to generate a summary, and the second assessment information to the server 1. The server 1 receives the format of the response, the command to generate a summary, and the second assessment information transmitted from the responsible physician's terminal 4. Based on the attending physician's assessment information, the assessment information of multiple specialists, and the received format of the response, the command to generate a summary, and the second assessment information, the server 1 generates a prompt to be given to the language model 151.

[0085] The prompts include instructions on how to write responses to patients, attending physician assessment information, assessment information from multiple specialists, instructions for generating summaries, second assessment information, and instructions for including the second assessment information in the response. An instruction to include the second assessment information in the response may be, for example, "Please be sure to include the following important sentences in your response."

[0086] Furthermore, the prompt may include instructions to output links to relevant terms (e.g., technical terms) within the various assessment information. These instructions might be, for example, "Refer to the knowledge database containing the medical information and output links to relevant terms within the assessment information."

[0087] The language model 151 is a language model that uses attending physician assessment information, multiple specialist assessment information, and second assessment information, and is used as a program module that is part of artificial intelligence software. In this embodiment, the language model 151 is a pre-constructed language model (language generation model) that takes prompts containing attending physician assessment information, multiple specialist assessment information, and second assessment information as input and outputs a response to the patient.

[0088] Alternatively, instead of storing the language model 151 in the large-capacity storage unit 15, the server 1 may access and read it from an external language processing server or language processing platform.

[0089] A prompt is an instruction or input sentence created in a format understandable to the language model 151 and given as input to the language model 151. The language model 151 interprets the input prompt and outputs an appropriate response (e.g., a response to the patient).

[0090] For example, the language model 151 divides the prompt into tokens so that it can be converted into a format that it can process. The language model 151 then performs contextual understanding by calculating the relationship between each token in the prompt and other tokens.

[0091] The language model 151 performs response generation processing in response to prompts based on linguistic knowledge obtained through pre-training and fine-tuning. For example, the language model 151 selects the optimal token using generation methods such as greedy decoding, beam search, or sampling. The language model 151 decodes the selected token to convert it back into text format and generates a response to the patient.

[0092] As an example, the generated prompt is: Please summarize the content of the chat between specialists regarding the following case, focusing on the medically important points. Please provide a detailed summary including the opinions of each specialist, specific figures, and conclusions drawn from those opinions. Please also describe the basis for each of your decisions. Attending physician assessment information: ×××××× Cardiologist assessment information: ×××××× Nephrologist assessment information: ×××××× Diabetologist assessment information: ×××××× Output format (no need for "desu" or "masu") Patient summary (non-bulleted format) Current Situation Assessment Possible interventions Future outlook and treatment plan Answers to patients' questions The following important sentence must be included in the text: We also offer this treatment on an outpatient basis. We will likely start with 4 units of rapid-acting insulin and 6 units of slow-acting insulin. The calculation is done as: Fasting C-peptide / Fasting blood glucose × 100. A value of 0.66 < 0.8 is interpreted as insulin-dependent. • Administering SGLT2 inhibitors to insulin-dependent patients carries a risk of euglycemic ketoacidosis. Furthermore, please refer to the knowledge database containing medical information and output links to relevant terms within the attending physician's assessment information or the assessment information of each specialist (cardiologist, nephrologist, and diabetologist).

[0093] Server 1 inputs the generated prompt into language model 151 and outputs (generates) a response to the patient.

[0094] As shown in the diagram, as an example, the response to the patient output is: “Patient Overview The patient has pancreatic cancer with metastases to the lungs, but it is being controlled with chemotherapy. They also have diabetes, and their kidney function is progressively declining. Current Situation Assessment The patient has diabetes and impaired renal function, and also shows signs of heart failure. Renal function has not decreased to below eGFR 10, and symptoms associated with renal failure are... Possible interventions 1.Diabetes treatment: Toformin discontinued. ××× 2. Renal function management: Improvement of edema and shortness of breath with diuretics (torasemide or furosemide intravenous injection). 3. Heart failure management: The use of a diuretic (torasemide) is intended to improve edema and shortness of breath. Future outlook and treatment plan The cardiorenal disease situation persists, and managing both heart failure and renal function is key to treatment. Currently, edema management with diuretics is prioritized, but improvement in cardiac function is... We also offer this treatment on an outpatient basis. We will likely start with 4 units of rapid-acting insulin and 6 units of slow-acting insulin. The calculation is done as: Fasting C-peptide / Fasting blood glucose × 100. A value of 0.66 < 0.8 is interpreted as insulin-dependent. • Administering SGLT2 inhibitors to insulin-dependent patients carries a risk of euglycemic ketoacidosis. Answers to patients' questions 1. "If kidney function deteriorates further, will dialysis be necessary?" When kidney function falls below eGFR 10 and heart failure or potassium control becomes difficult, ××× 2. "Is the current diabetes treatment sufficient? Is a change in medication necessary?" Since the current treatment is not meeting the HbA1c target, it is considered necessary to initiate insulin therapy. For information on technical terms, please visit the following website. "Diabetes": Guidelines of the Japan Diabetes Society It could also be "xxx".

[0095] Furthermore, the provision of links to medical terms is not limited to output processing by language model 151. For example, after the response to the patient is output by language model 151, server 1 can use natural language processing (NLP) techniques to obtain links to medical terms.

[0096] Specifically, Server 1 uses, for example, natural language processing technology to analyze words and phrases in the response to the patient output from Language Model 151, and identifies medically relevant technical terms (such as disease names, symptoms, drug names, treatment methods, or diagnostic results) in consideration of the context. One example of a technical term identified is "diabetes."

[0097] Server 1, based on the identified technical term, refers to Knowledge DB 155, which stores medical information, and obtains a link to that medical term. As an example, the obtained link is the URL to the "Japan Diabetes Society Guidelines."

[0098] Server 1 stores the patient responses output from the language model 151 in the response DB 154. Specifically, Server 1 stores the attending physician assessment information, multiple specialist assessment information, second assessment information, the patient response output from the language model 151, and the consultation date and time as a single record in the response DB 154, associated with the patient ID. The consultation date and time is, for example, the date and time when the patient response was output from the language model 151.

[0099] Server 1 sends a response regarding the patient to the responsible physician's terminal 4. The responsible physician's terminal 4 receives the response regarding the patient sent from Server 1. The responsible physician's terminal 4 displays the received response regarding the patient on its screen. The responsible physician's terminal 4 accepts edits to the response by the responsible physician.

[0100] For example, the responsible physician terminal 4 may accept edits to the patient summary included in the response, changing from "The patient's pancreatic cancer has metastasized to the lungs, but is controlled with chemotherapy." to "This patient is 2 years post-surgery for pancreatic cancer, has lung metastases, but is controlled by both chemotherapy and chemotherapy using TS-1 by their primary care physician, an abdominal surgeon. The patient has pancreatic cancer with lung metastases, but is controlled with chemotherapy."

[0101] The responsible physician's terminal 4 sends the edited response to server 1. Server 1 receives the edited response sent from the responsible physician's terminal 4. Server 1 stores the edited response in the response DB 154, associating it with the patient ID. Server 1 sends the received edited response to the attending physician's terminal 2 and the specialist's terminal 3 of the relevant specialist. The attending physician's terminal 2 and each specialist's terminal 3 each receive the edited response sent from server 1 and display it on their screens.

[0102] Figures 8 and 9 are flowcharts showing the processing procedure when the language model 151 outputs a response to a patient. The control unit 11 of the server 1 retrieves specialist information (medical department and name, etc.) of multiple specialists from the physician DB 153 of the large-capacity storage unit 15 based on the physician ID (step S101). The control unit 11 transmits the retrieved specialist information to the attending physician terminal 2 via the communication unit 13 (step S102).

[0103] The control unit 21 of the attending physician terminal 2 receives multiple specialist information transmitted from the server 1 via the communication unit 23 (step S201). The control unit 21 displays the received multiple specialist information via the display unit 25 (step S202). The control unit 21 acquires patient information of the target patient via the communication unit 23 or the input unit 24 (step S203).

[0104] Patient information includes, but is not limited to, basic information (name, age, and gender, etc.), subject (symptoms and problems reported, etc.), present illness, past medical history, physical examination data (blood pressure, pulse, temperature, breath sounds or heart sounds, etc.), laboratory data (blood tests, urine tests, imaging tests or electrocardiograms, etc.), or medications, etc.

[0105] The control unit 21 receives the selection of multiple specialists (for example, a cardiologist, a nephrologist, and a diabetologist) by the attending physician based on the acquired patient information via the input unit 24 (step S204). The control unit 21 transmits the acquired patient information, the attending physician's doctor ID, and the doctor IDs of the selected multiple specialists to the server 1 via the communication unit 23 (step S205).

[0106] The control unit 11 of server 1 receives patient information, the attending physician's ID, and the specialist's ID transmitted from the attending physician's terminal 2 via the communication unit 13 (step S103). The control unit 11 stores the received patient information, the attending physician's ID, and the specialist's ID in the patient DB 152 of the large-capacity storage unit 15 (step S104).

[0107] Specifically, the control unit 11 assigns a patient ID to the received patient information. The control unit 11 stores the basic information, subject, present illness, past medical history, medication information, physical examination data, test data, the attending physician's ID, and the specialist physician's ID as a single record in the patient database 152, associating them with the assigned patient ID.

[0108] The control unit 21 of the attending physician terminal 2 receives a chat room creation request via the input unit 24 (step S206). The chat room creation request includes information about the participating physicians (attending physician or specialist, etc.) (physician ID and IP address, etc.), settings related to the chat room (for example, private or public, room name, start time and end time), etc.

[0109] The control unit 21 transmits the received chat room creation request to the server 1 via the communication unit 23 (step S207). The chat room creation request may also be transmitted to the server 1 from the specialist terminal 3 of the selected specialist.

[0110] The control unit 11 of server 1 receives a chat room creation request sent from the attending physician terminal 2 via the communication unit 13 (step S105). In response to the received chat room creation request, the control unit 11 connects to, for example, an online meeting distribution service and creates a chat room (step S106).

[0111] The control unit 11 sends a notification containing the generated chat room meeting link, etc., to both the attending physician terminal 2 and the target specialist terminal 3 via the communication unit 13 (step S107). The control unit 21 of the attending physician terminal 2 receives the notification sent from the server 1 via the communication unit 23 (step S208). The control unit 21 connects to the chat room via the communication unit 23 in response to the received notification (step S209).

[0112] The control unit 31 of each specialist terminal 3 receives notifications sent from the server 1 via the communication unit 33 (step S301). The control unit 31 connects to the chat room via the communication unit 33 in response to the received notification (step S302).

[0113] The control unit 11 of server 1 stores the content of the conversation between the attending physician and multiple specialists (patient's symptoms, treatment history, current medical history, test results, additional tests and treatment plans, etc.) in the storage unit 12 or large-capacity storage unit 15 in real time through the chat room (step S108). The conversation content includes the attending physician's assessment information and the assessment information of multiple specialists. The assessment information includes the patient's medical history, physical examination data, test data, diagnosis, risks, or treatment plan, etc.

[0114] The control unit 11 transmits the acquired conversation content to the responsible physician's terminal 4 via the communication unit 13 (step S109). The control unit 41 of the responsible physician's terminal 4 receives the conversation content transmitted from the server 1 via the communication unit 43 (step S401). The control unit 41 displays the received conversation content via the display unit 45 (step S402). The control unit 41 receives the responsible physician's response format, summary generation command, and input of second assessment information via the input unit 44 (step S403).

[0115] The command to generate a summary is a command to generate summaries for the attending physician's assessment information and the assessment information of multiple specialists. The second assessment information is assessment information related to the attending physician's assessment information and the assessment information of multiple specialists.

[0116] The control unit 41 transmits the format of the received response, the command to generate a summary, and the second assessment information to the server 1 via the communication unit 43 (step S404). The control unit 11 of the server 1 receives the format of the response, the command to generate a summary, and the second assessment information transmitted from the responsible physician terminal 4 via the communication unit 13 (step S110).

[0117] The control unit 11 generates prompts to be given to the language model 151 based on the attending physician assessment information, multiple specialist assessment information, the format of the received response, a summary generation command, and second assessment information (step S111). The prompts include, for example, the format of the response to the patient, the attending physician assessment information, multiple specialist assessment information, a summary generation command, second assessment information, a command that includes the second assessment information in the response, and an output instruction for a link to a relevant term in the various assessment information.

[0118] The control unit 11 inputs the generated prompt to the language model 151 (step S112) and outputs a response to the patient (step S113). The control unit 11 stores the response to the patient output from the language model 151 in the response DB 154 of the large-capacity storage unit 15, associating it with the patient ID (step S114).

[0119] Specifically, the control unit 11 stores the attending physician assessment information, multiple specialist assessment information, second assessment information, the response to the patient output from the language model 151, and the consultation date and time as a single record in the response DB 154, associated with the patient ID.

[0120] The control unit 11 transmits the response to the patient output from the language model 151 to the responsible physician terminal 4 via the communication unit 13 (step S115). The control unit 41 of the responsible physician terminal 4 receives the response to the patient transmitted from the server 1 via the communication unit 43 (step S405).

[0121] The control unit 41 displays the received response to the patient using the display unit 45 (step S406). The control unit 41 receives edits to the response from the responsible physician using the input unit 44 (step S407). The control unit 41 transmits the edited response to the patient to the server 1 using the communication unit 43 (step S408).

[0122] The control unit 11 of server 1 receives the edited response sent from the responsible physician's terminal 4 via the communication unit 13 (step S116). The control unit 11 transmits the received edited response via the communication unit 13 to the attending physician's terminal 2 and the specialist's terminal 3 of the relevant specialist (step S117). The control unit 11 stores the edited response in the response DB 154, associating it with the patient ID (step S118). The control unit 11 terminates processing.

[0123] Next, we will explain the process of retraining the language model 151 based on the assessment information, the second assessment information, and the edited response. Figure 10 is a flowchart showing the processing procedure when retraining the language model 151. Based on each patient ID, the control unit 11 of the server 1 obtains the attending physician assessment information, the assessment information of multiple specialists, the second assessment information by the responsible physician, and the response to the patient after editing by the responsible physician from the patient DB 152 of the large-capacity storage unit 15 (step S121).

[0124] The control unit 11 generates training data for retraining based on the attending physician assessment information, multiple specialist assessment information, second assessment information, and edited responses to the patient acquired for each patient (step S122). The training data is combined data in which the attending physician assessment information, multiple specialist assessment information, and second assessment information for the patient are associated with the edited responses to the patient.

[0125] The control unit 11 performs a retraining process for the language model using the generated training data (step S123). Specifically, the control unit 11 performs a conversion process to convert the training data, which includes the attending physician assessment information, multiple specialist assessment information and second assessment information, and the edited responses to the patient, into a format (tokens) that the language model 151 can process.

[0126] Specifically, the control unit 11 performs training processing for a tokenizer that divides the acquired training data into tokens. The control unit 11 then uses the divided tokens to retrain the language model 151 for a specific task (e.g., question answering). Reinforcement learning may also be performed on the language model 151.

[0127] The control unit 11 evaluates the performance of the retrained language model 151 using evaluation metrics such as accuracy, recall, or precision. The control unit 11 adjusts hyperparameters (learning rate, batch size, or number of epochs, etc.) and repeats the retraining and hyperparameter adjustment process described above until the performance reaches a predetermined threshold.

[0128] The control unit 11 stores the retrained language model 151 in the storage unit 12 or the large-capacity storage unit 15 and deploys it (step S124). The control unit 11 then terminates the process.

[0129] According to this embodiment, it is possible to output a response to the patient using the language model 151 based on the attending physician's assessment information, the assessment information of multiple specialists, and the second assessment information from the responsible physician.

[0130] According to this embodiment, by utilizing the language model 151, it becomes possible to integrate assessments from specialists and multiple specialists to quickly and effectively provide patients with the most suitable treatment plan.

[0131] According to this embodiment, by performing a retraining process on the language model 151, it is possible to improve the accuracy of responses to patients. (Embodiment 2) Embodiment 2 relates to a configuration in which, after outputting a response to a patient from the language model 151, a second response is output using the language model 151 with the test data obtained by performing a test on the patient. Note that explanations of content that overlaps with Embodiment 1 will be omitted.

[0132] Figure 11 is an explanatory diagram illustrating the operation of the collective intelligence formation system for physicians in Embodiment 2. Note that explanations of content that overlaps with Figure 7 are omitted.

[0133] After the language model 151 outputs a response to the patient, the attending physician, responsible physician, or nurse, etc., will perform tests (blood tests, urine tests, imaging tests, or electrocardiograms, etc.) on the patient according to that response.

[0134] Server 1 acquires the test data obtained by performing tests on the patient from the relevant testing equipment. Server 1 outputs the acquired test data to a chat room in Embodiment 1 for dialogue between the attending physician and multiple specialists. After the test data is output to the chat room, the attending physician and multiple specialists engage in further dialogue in the chat room based on the test data.

[0135] Furthermore, based on the patient's test data, it is possible to add specialists other than those previously selected (for example, specialists in cardiology, nephrology, and diabetology).

[0136] For example, if a new disease (e.g., vasculitis) is discovered based on the patient's test data, the attending physician terminal 2 accepts the attending physician's selection of a specialist (e.g., a vascular surgeon) appropriate for that disease. The attending physician terminal 2 sends the selected new specialist to server 1. Server 1 invites the new specialist sent from the attending physician terminal 2 to the chat room.

[0137] Furthermore, based on the patient's examination data, the language model can be used to add specialists different from those previously selected. For example, the attending physician terminal 2 may input a prompt to the language model containing patient information, the patient's examination data, information about the previously selected specialists (physician ID, medical specialty, and name, etc.), and a command to select a specialist different from the previously selected specialists based on the patient's examination data, and output a different specialist. The selection command may be, for example, "Based on the patient's examination data, refer to the physician database and output a specialist suitable for the patient, different from the previously selected specialists."

[0138] Server 1 retrieves conversation content from the chat room through dialogue between the attending physician and multiple specialists. The conversation content includes new attending physician assessment information and multiple new specialist assessment information. The conversation content may also include past conversation content obtained through calls in the chat room (attending physician assessment information and multiple specialist assessment information, etc.). Past conversation content may include, for example, the previous conversation or the content of several past conversations.

[0139] Server 1 transmits the acquired conversation content to the responsible physician's terminal 4. The responsible physician's terminal 4 receives the conversation content transmitted from Server 1. The responsible physician's terminal 4 receives new second assessment information from the responsible physician regarding new attending physician assessment information and multiple new specialist assessment information included in the conversation content. The responsible physician's terminal 4 transmits the received new second assessment information to Server 1.

[0140] Server 1 receives new second assessment information transmitted from the responsible physician's terminal 4. Server 1 outputs a second response to the patient by providing the language model 151 with the new attending physician assessment information, multiple new specialist assessment information, and the new second assessment information. Server 1 transmits the second response output from the language model 151 to the responsible physician's terminal 4.

[0141] The responsible physician's terminal 4 receives the second response sent from the server 1. The responsible physician's terminal 4 accepts edits to the second response by the responsible physician, similar to the processing in Embodiment 1. The responsible physician's terminal 4 sends the edited second response to the server 1. The server 1 sends the edited second response sent from the responsible physician's terminal 4 to the attending physician's terminal 2 and the specialist's terminal 3 of the relevant specialist.

[0142] Figures 12 and 13 are flowcharts showing the processing procedure when the language model 151 outputs a second response to the patient. Note that the same reference numerals are used for elements that overlap with Figures 8 and 9, and their explanations are omitted.

[0143] The control unit 21 of the attending physician terminal 2 executes the processes in steps S206 to S209. The control unit 11 of the server 1 executes the processes in steps S105 to S107. The control unit 31 of the specialist physician terminal 3 executes the processes in steps S301 to S302.

[0144] After outputting a response to the patient from the language model 151, the control unit 11 of the server 1 acquires the examination data obtained by performing the examination on the patient via the communication unit 13 (step S131). For example, the control unit 11 may acquire the patient's examination data from the target examination device (e.g., an electrocardiogram device) via the communication unit 13. Alternatively, the control unit 11 may acquire the patient's examination data from the attending physician's terminal 2 via the communication unit 13.

[0145] The control unit 11 outputs the acquired test data to the chat room (step S132). The control unit 11 executes the processes in steps S108 to S109. The dialogue content also includes new attending physician assessment information and new specialist assessment information from multiple specialists.

[0146] The control unit 41 of the responsible physician's terminal 4 executes the processes in steps S401 to S402. The control unit 41 receives the response format, summary generation command, and input of new second assessment information from the responsible physician via the input unit 44 (step S431).

[0147] The generation command is a command to generate summaries of new attending physician assessment information and multiple new specialist assessment information. The new second assessment information is assessment information relating to the new attending physician assessment information and multiple new specialist assessment information.

[0148] The control unit 41 transmits the format of the received response, the command to generate a summary, and new second assessment information to the server 1 via the communication unit 43 (step S432). The control unit 11 of the server 1 receives the format of the response, the command to generate a summary, and new second assessment information transmitted from the responsible physician terminal 4 via the communication unit 13 (step S133).

[0149] The control unit 11 generates a prompt to give to the language model 151 based on the new attending physician assessment information, multiple new specialist assessment information, the format of the received response, the command to generate a summary, and the new second assessment information (step S134).

[0150] Prompts may include, for example, instructions on how to write responses to patients, new attending physician assessment information, multiple new specialist assessment information, instructions for generating summaries, new second assessment information, instructions for including such new second assessment information in the response, and instructions for outputting links to relevant terms in various new assessment information.

[0151] The control unit 11 executes the process in step S112 and outputs a second response to the patient (step S135). The control unit 11 stores the second response to the patient output from the language model 151 in the response DB 154 of the large-capacity storage unit 15, associating it with the patient ID (step S136).

[0152] Specifically, the control unit 11 stores the new attending physician assessment information, multiple new specialist assessment information, new second assessment information, the second response to the patient output from the language model 151, and the consultation date and time as a single record in the response DB 154, associated with the patient ID.

[0153] The control unit 11 transmits the second response output from the language model 151 to the responsible physician terminal 4 via the communication unit 13 (step S137). The control unit 41 of the responsible physician terminal 4 receives the second response transmitted from the server 1 via the communication unit 43 (step S433).

[0154] The control unit 41 displays the received second response using the display unit 45 (step S434). The control unit 41 receives edits to the second response from the responsible physician using the input unit 44 (step S435). The control unit 41 transmits the edited second response to the server 1 using the communication unit 43 (step S436).

[0155] The control unit 11 of server 1 receives the edited second response sent from the responsible physician's terminal 4 via the communication unit 13 (step S138). The control unit 11 transmits the received edited second response to the attending physician's terminal 2 and each specialist's terminal 3 via the communication unit 13 (step S139). The control unit 11 stores the edited second response in the response DB 154, associating it with the patient ID (step S140). The control unit 11 terminates processing.

[0156] According to this embodiment, after outputting a response to the patient from the language model 151, it becomes possible to output a second response using the language model 151 with the test data obtained by performing a test on the patient.

[0157] According to this embodiment, by outputting a second response from the language model 151 based on the examination data after the response to the patient, it becomes possible to improve the accuracy of diagnosis and treatment and provide the patient with the optimal treatment plan. (Embodiment 3) Embodiment 3 relates to a method for determining compensation for the provision of assessment information by the attending physician and specialists. Note that explanations of content that overlaps with Embodiments 1 and 2 will be omitted.

[0158] Figure 14 is an explanatory diagram illustrating the process of determining compensation for providing assessment information. By using this system, attending physicians and specialists can collaborate in providing medical care to patients, thereby enabling the provision of more optimal treatment plans. Furthermore, appropriate compensation can be determined based on the assessment information provided by both the attending physician and the specialist.

[0159] Compensation is an incentive (remuneration) paid to the attending physician or specialist based on the attending physician's assessment information or the specialist's assessment information. In addition to monetary compensation, compensation includes various forms of rewards such as virtual coins or points.

[0160] The compensation for specialists is calculated, for example, based on the specialist's "assigned price." The specialist's "assigned price" refers to the amount set by the specialist themselves for the assessment information they provide, based on their experience, evaluation, or the patient's needs.

[0161] For example, in the chat room of Embodiment 1, each time an assessment is performed on a patient, the specialist may be given a set "asking price" for the assessment (e.g., "10,000 yen"). Alternatively, in addition to a predetermined fixed fee for a predetermined period (e.g., one week, one month, or three months) (e.g., "30,000 yen"), the specialist may be given an amount calculated based on the "asking price" for one assessment and the number of assessments ("asking price" x number of assessments).

[0162] Furthermore, the calculation of compensation for specialists is not limited to a "negotiated price." For example, a predetermined fixed fee (e.g., "15,000 yen") could be paid to the specialist each time they perform an assessment on a patient.

[0163] The recipient of compensation for specialists may be the patient, the operator running the system, or the medical facility where the specialist works.

[0164] Furthermore, when a primary physician and a specialist collaborate to provide medical care to a patient, the primary physician provides the specialist with medical knowledge (e.g., primary physician assessment information), and therefore the specialist can provide the primary physician with compensation (honorarium). The compensation to the primary physician may be a predetermined amount (e.g., "5,000 yen"), or it may be calculated as a predetermined percentage (e.g., 10%) of the compensation paid to the specialist.

[0165] Furthermore, the recipient of compensation to the attending physician is not limited to specialists; it could also be, for example, the operator running the system, or the medical facility where the attending physician works.

[0166] Figure 15 is a flowchart showing the processing procedure for determining compensation for providing assessment information. The control unit 11 of server 1 retrieves the attending physician and multiple specialists corresponding to the patient from the patient DB 152 of the large-capacity storage unit 15 based on the patient ID (step S151).

[0167] The control unit 11 determines the compensation for each specialist (step S152). For example, the control unit 11 retrieves a predetermined fixed fee for a predetermined period (e.g., "30,000 yen") and the "asking price" for each specialist's assessment, which are stored in advance in the memory unit 12 or the large-capacity memory unit 15. The control unit 11 then aggregates the number of times each specialist performed an assessment for the patient during the predetermined period from the response DB 154 in the large-capacity memory unit 15.

[0168] The control unit 11 calculates the compensation for each assessment based on the "asking price" for one assessment and the number of assessments conducted for each specialist. The control unit 11 adds the calculated compensation for each assessment to a predetermined fixed fee for each specialist, and determines this as the compensation for the specialist over a predetermined period.

[0169] The control unit 11, for example, grants the appropriate specialist compensation to the specialist via the communication unit 13 through the compensation system (step S153). The control unit 11 determines the compensation for the attending physician (step S154). For example, the control unit 11 calculates the compensation that should be received from each specialist by taking a predetermined percentage (for example, 10%) from the compensation for each specialist. The control unit 11 determines the sum of the compensation that should be received from each specialist as the compensation for the attending physician.

[0170] The control unit 11, for example, grants the designated physician the compensation via the communication unit 13 through the compensation system (step S155). The control unit 11 then terminates the process.

[0171] According to this embodiment, it becomes possible to determine compensation for the provision of assessment information by the attending physician and specialists.

[0172] According to this embodiment, by providing compensation to the attending physician and specialist, they will strive to provide more accurate and reliable assessment information, thereby improving the quality of medical care and diagnostic accuracy.

[0173] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The scope of the invention is indicated by the claims, not in the sense described above, and all modifications are intended to be in the sense and scope equivalent to the claims.

[0174] The matters described in each embodiment can be combined with each other. Furthermore, the independent and dependent claims described in the claims can be combined with each other in any combination, regardless of the form of reference. In addition, the claims use a form in which claims referencing two or more other claims (multi-claim form), but are not limited to this. A form in which multi-claims referencing at least one multi-claim (multi-multi-claim) may also be used. [Explanation of symbols]

[0175] 1. Information processing device (server) 11 Control Unit 12 Storage section 13 Communications Department 14 Reading section 15 Mass storage 151 Language Models 152 Patient DB 153 Doctor Database 154 Response DB 155 Knowledge DB 1a Portable storage medium 1b Semiconductor memory 1P Control Program 2. Information processing terminal (attending physician's terminal) 21 Control Unit 22 Memory section 23 Communications Department 24 Input section 25 Display section 2P control program 3. Information processing terminal (specialist's terminal) 31 Control Unit 32 Storage section 33 Communications Department 34 Input section 35 Display section 4. Information processing terminal (responsible physician's terminal) 41 Control Unit 42 Storage section 43 Communications Department 44 Input section 45 Display section B Bus N Network

Claims

1. Obtain patient information from the patient, Based on the patient information obtained, assessment information is received from the selected healthcare professionals and the attending physician. Second assessment information regarding the aforementioned assessment information was obtained from the responsible physician. By providing the acquired assessment information and the second assessment information to the language model, a response to the patient is output. A program that instructs a computer to perform a process.

2. The aforementioned medical professionals are multiple, and the assessment information is obtained through dialogue between the attending physician and the multiple aforementioned medical professionals. The program according to claim 1.

3. The assessment information is output to the responsible physician. Obtain the second assessment information to be included in the response from the responsible physician. The program according to claim 1 or 2.

4. The language model is given a prompt having the response format, the assessment information, a command to generate a summary for the assessment information, the second assessment information, and a command that includes the second assessment information in the response. The program according to claim 1 or 2.

5. The responsible physician has requested edits to the above response. Output the edited response. The program according to claim 1 or 2.

6. Based on the aforementioned assessment information, the second assessment information, and the edited response, the language model is retrained. The program according to claim 5.

7. After the above response, the examination data obtained by performing the examination on the patient is acquired. The acquired test data is output to a chat room for dialogue between the attending physician and multiple healthcare professionals. The program according to claim 1 or 2.

8. After outputting the aforementioned test data to the chat room, new assessment information is obtained from the attending physician and multiple healthcare professionals. We obtained new second assessment information from the aforementioned responsible physician. By providing the newly acquired assessment information and the new second assessment information to the language model, a second response is output. The program according to claim 7.

9. The compensation for providing assessment information by the aforementioned medical professionals will be determined. The program according to claim 1 or 2.

10. Referencing a database containing medical information, the system outputs links to relevant terms in the assessment information or second assessment information during the response. The program according to claim 1 or 2.

11. Obtain patient information from the patient, Based on the patient information obtained, assessment information is received from the selected healthcare professionals and the attending physician. Second assessment information regarding the aforementioned assessment information was obtained from the responsible physician. By providing the acquired assessment information and the second assessment information to the language model, a response to the patient is output. Information processing methods.

12. An information processing device comprising a control unit, The control unit, Obtain patient information from the patient, Based on the patient information obtained, assessment information is received from the selected healthcare professionals and the attending physician. Second assessment information regarding the aforementioned assessment information was obtained from the responsible physician. By providing the acquired assessment information and the second assessment information to the language model, a response to the patient is output. Information processing device.

Images