Medical interview system

Abstract

A medical interview tool (system, program, method) that automatically transcribes patient speech into text and automatically collects and inputs interview information. This solves the problem of missing information that occurs in conventional medical interview systems, enabling more efficient interviews. [Solution] The medical interview system 2 automatically converts the patient's speech content from the patient terminal 3 into text using speech recognition and extracts symptom and disease information. Based on medical interview templates for each symptom / disease and medical department, it generates additional questions that automatically complete any missing information that cannot be obtained from the speech content and presents them to the patient. Furthermore, by working in cooperation with the medical system 4 to determine whether it is an initial or follow-up visit, it ensures that the medical interview is conducted appropriately. In addition, an automated identity verification process and the generation of additional questions according to priority enable efficient and accurate collection of medical interview information.

Description

Technical Field

[0001] The present invention relates to an interview tool (interview system, program, method) that automatically converts the speech content of a patient into text to automatically input interview information and collects the insufficient interview information by AI.

Background Art

[0002] When receiving medical treatment at a medical facility, it is well known that the patient fills out an interview form in advance with their medical history, reasons for visiting the doctor, etc., and medical staff such as nurses and doctors provide medical treatment based on the information filled out in the interview form. In recent years, there is a technology called an electronic interview form that allows the interview content, which is the patient's answer, to be input via an information processing device such as a tablet, and the input interview content can be imported as the input content of the electronic medical record.

[0003] The technology of Patent Document 1 receives the user's answer data for interview data from the user terminal and notifies the user terminal of interview data for confirming a more detailed medical condition. Assuming a telephone such as a landline phone when the user calls the medical institution directly, it shows the interview content in voice by an automatic response and allows the answer content to be input by a push response. However, the technology of Patent Document 1 mainly depends on the response by the user's push operation or terminal input, and is not a dynamic progress of the dialogue or an automatic data collection / generation process based on the speech content of the patient. Specifically, since it is a mechanism in which an automatic voice response system conducts an interview via a landline phone or the like and the user answers by a push operation, the flow of the conversation is fixed, and a dynamic dialogue according to the speech content of the user and the situation at the time cannot be performed. When the patient's condition is complicated, it becomes difficult to take appropriate measures.

[0004] On the other hand, the system of this application includes a conversion unit that transcribes the patient's speech in real time and extracts symptom and disease information based on that transcription. Furthermore, it is possible to select an appropriate questionnaire template based on that information and dynamically generate and present additional questions. In addition, it can detect any unentered content remaining in the questionnaire template and collect further speech content based on that, enabling two-way interaction with the patient. This realizes a flexible and natural questionnaire process that cannot be provided by the technology of Patent Document 1. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2022-070155

[0006] While electronic questionnaires allow the medical interview to proceed based on the information entered by the patient, they often result in missing fields, and the following factors have been identified as contributing to this problem. <Lack of patient understanding> The questions are difficult to understand, use technical jargon, or contain complex content, making it difficult for patients to answer accurately. <Misunderstanding of the question> The patient misunderstood the intent of the question and decided it was irrelevant to them, thus choosing not to answer. <Operation error> Because I'm unfamiliar with using tablets and similar devices, I often overlook items and proceed before completing the input. <Limited options> If the answer choices do not adequately reflect the patient's condition, the system will withhold the answer and proceed without entering anything. This is especially frequent with questions that offer only a limited number of choices. <Privacy and personal reasons> Patients may hesitate to answer questions about private information or sensitive topics (e.g., mental health, sex life, family history, etc.).

[0007] In other words, while traditional medical interview systems relied on person-to-person communication—where healthcare professionals directly questioned patients and assisted them in filling out questionnaires—electronic questionnaires, while offering increased efficiency, have limitations in terms of the types of answers they can provide. There are no systems that automatically handle unanswered questions, meaning that if there are unanswered questions, the system ultimately relies on human intervention, with nurses or doctors having to review the information, ask questions directly as needed, and supplement the data.

[0008] Furthermore, when conducting a medical interview via the web, patients answer questions based solely on visual information. While this allows patients to carefully read the questions at their own pace, the simultaneous display of all questions can be confusing, making it difficult to know where to focus. This can lead to patients rushing to answer earlier questions before looking at later ones, potentially compromising the accuracy and truthfulness of their responses. In fact, if accuracy is the goal, it would be better to express the questions in simple language and conduct the interview in a question-and-answer format, or even better, to ask questions using short, concise voice phrases.

[0009] Furthermore, if there are many instances of unfilled fields in the medical questionnaire, the medical institution is required to analyze the trend on the system side and improve the templates to clarify the questions and enhance usability. [Overview of the project] [Problems that the invention aims to solve]

[0010] This invention has been made in view of the above problems, and automatically converts the patient's speech content into text using speech recognition and natural language processing, and automatically inputs the content obtained from the medical interview into an electronic medical interview system. Furthermore, it collects any missing medical interview information using AI to create an electronic medical interview form. The invention provides such a medical interview tool (medical interview system, program, method). Furthermore, in conventional medical interview and treatment processes, discrepancies in information between the patient's subjective information (column S) and objective findings (column O), the effort of duplicate input, and oversights were concerns. Therefore, there was a desire to provide a medical interview tool equipped with a mechanism that analyzes the patient's subjective information and automatically generates examination items and physical findings items based on it, enabling healthcare professionals to obtain necessary information accurately and efficiently. Moreover, it was desired to avoid cluttering the user interface by skipping or hiding unnecessary items, and to configure the tool to efficiently transfer the information to column O of the electronic medical record. In view of these issues, the present invention provides a medical interview tool that combines a means for inputting the patient's subjective information, a template generation means for automatically generating necessary findings items from the analysis results, and a user interface for display and input assistance. This allows for the automatic skipping or hiding of items that do not match the patient's condition, enabling efficient selection and input of only the findings and examination items that the physician deems necessary, and ultimately transferring the information to column O of the electronic medical record. [Means for solving the problem]

[0011] (1) A conversion unit that collects the patient's speech content and converts it into text data, A selection unit extracts one or more symptom information or disease information from the aforementioned text data, and selects a questionnaire template for each disease symptom based on the extracted symptom information or disease information. An input unit that inputs the aforementioned text data into the aforementioned disease symptom-specific questionnaire template, A questioning unit that selects or generates additional questions based on the remaining unentered content in the aforementioned disease symptom-specific questionnaire template, The system includes a collection unit that presents the patient with the aforementioned additional questions and collects additional speech content. The input unit is characterized by converting additional speech content into text using the conversion unit to create additional text data, and then inputting data into the disease symptom-specific questionnaire template based on the additional text data.

[0012] This invention efficiently conducts medical interviews by automatically analyzing the content of a patient's speech. Since medical institutions often have limited patient information during initial consultations, the system uses the patient's speech to infer symptom and disease information. It then utilizes disease-specific interview templates to input data while simultaneously calculating any unentered information and asking additional questions to reflect that missing data. The "disease-specific interview templates" define standard interview content designed based on specific diseases and symptoms, specifying a list and format of questions to guide the interview. Because standardized questions are available for each disease and symptom, it enables the rapid and comprehensive collection of necessary information. Furthermore, with interview templates, information can be collected using the same criteria for any patient.

[0013] (2) A conversion unit that collects the patient's speech content and converts it into text data, A selection unit for selecting a medical department-specific questionnaire template from one or more of the following: the aforementioned text data, the information from the reservation system, or the information from the electronic medical record. An input unit that inputs the aforementioned text data into the aforementioned medical department-specific questionnaire template, A questioning unit that selects or generates additional questions based on the remaining unentered content in the aforementioned medical department-specific questionnaire templates, The system includes a collection unit that presents the patient with the aforementioned additional questions and collects additional speech content. The input unit is characterized by converting additional speech content into text using the conversion unit to create additional text data, and then inputting data into the medical department-specific questionnaire template based on the additional text data.

[0014] This invention efficiently conducts medical interviews by automatically analyzing the content of a patient's speech. Since medical institutions have limited patient information during initial consultations, the system uses the patient's speech to infer the medical department, utilizes medical department-specific interview templates to input data, calculates any unentered information, and asks additional questions to reflect that data. Even if a patient has visited a medical institution before, if the symptoms or illness are different from previous visits, it can be treated as an initial consultation. If the patient has previously visited a medical institution for ongoing treatment, it is treated as a follow-up consultation. The medical department is inferred based on patient information registered in the reservation system or electronic medical record. The "medical department-specific interview templates" are lists and formats of standard interview items and questions tailored to the specific medical conditions and symptoms associated with a particular medical department. Each medical department has a different area of ​​expertise, requiring the collection of information specific to that department. In technical terms, these transitions in the medical interview are called open questions, closed questions, and background questions. Open questions are used to broadly gather information from the patient, followed by questions to identify the chief complaint based on the information gathered. Closed questions are then used to delve deeper into the chief complaint, and after that, background questions are asked to confirm the patient's medical history and other background information. Through these screen or voice interaction interfaces, (1) open questions are presented, (2) after the user has finished answering, the chief complaint is identified and a list of closed questions based on the chief complaint is obtained, and (3) only the unanswered closed questions are automatically extracted and displayed sequentially. This (4) eliminates duplication and waste of questions, enabling efficient medical interviews. This process can be applied not only to chief complaints but also to diseases. In the case of diseases, closed questions would include the degree of ADL impairment, presence or absence of complications, severity assessment, past treatments, date of onset, medical institutions visited, and treatments. In this embodiment, the process of the medical interview may be displayed as a UI, and a screen may be displayed showing the chief complaint, the content of the closed question, and the filled-in and unfilled parts of the template.

[0015] A "medical interview template" refers to two or more sets of questions optimized for diagnosis and interviews, based on the patient's symptom or disease information. It is a structured template designed to collect appropriate information for a selected disease, selected symptoms, or selected medical department. A template contains two or more sets of questions, some of which correspond to specific diseases or symptoms, or to different phases of the diagnostic process. By consisting of multiple sets of questions corresponding to specific diseases or symptoms and different phases of the diagnostic process, appropriate information can be collected step-by-step according to the progression of the disease and changes in symptoms.

[0016] Furthermore, in one form of the questionnaire template, each question has a set answer format, or value type, which may include binary answers such as "yes" or "no," free text input, or specific value formats such as dates and days. Based on these value type settings, the system can accurately determine whether the patient's answers were sufficiently correct.

[0017] Templates include not only predefined sets of questions but also questions automatically generated according to the patient's symptoms and illness. Even if provided in real time, they can be treated as "medical history templates." Even if the format is simple or the number of questions is small, as long as the value type is set, it functions as a template. In other words, a "medical history template" may be described as a "medical history set," "conversation scenario," "question list," "medical history prompt," or "conversation framework," and is treated as a "medical history template" when two or more medical history questions are selected when certain conditions (chief complaint, department visited, diagnosis, etc.) are determined. Another example is to include prompts in a large-scale language model that, as a framework, support the flow of the conversation and the direction of question list creation, and control the flow of the medical history. In particular, by customizing the template according to the special requirements of each hospital or department, more efficient medical interviews and diagnostic support can be provided. The present invention also includes a form in which interview items are given as prompts to a generative AI model. As a result, the AI can dynamically conduct open-ended and closed-ended questions with the patient based on the interview items, and control the flow of (1) preset interview items, (2) passing them to the model as prompts, and (3) the progress of the conversation with the patient. At that time, as a prompt, the generation of additional interviews regarding the chief complaint may be controlled using the following text. Example of a prompt: - For each reason for seeking medical treatment, conduct 5 to 10 question-and-answer sessions regarding the chief complaint using the following framework. However, in the case of severe conditions with multiple chief complaints, repeat the interaction about 5 to 10 times for each chief complaint. Symptom framework (T-P-Q-R-S-A) - **T (Start time)**: Since when have you noticed that symptom? - **T (Sudden onset)**: Did the symptom occur suddenly? Sudden means within 5 minutes until the time when the symptom was most severe. - **T (Recurrence)**: Have you ever noticed the same symptom before this episode? - **T (Symptom)**: Does this symptom get better or worse? - **Q (Character)**: What kind of pain was it? - **R (Location)**: Please tell me the location where you feel the pain. - **S (Worst severity)**: When it was most painful, how would you rate that pain on a scale of 0 to 10? Please tell me, with no pain being 0 and the most painful pain in life being 10. - **P (Inducing factor / Relieving factor)**: When does the symptom get stronger or better? - **A (Related symptoms)**: When you noticed this symptom, did you notice any other symptoms? Other variations of implementation include the following forms: Variation 1: A form in which open-ended questions are completely omitted and only closed-ended questions are displayed from the start. Modification 2: A format in which all questions are listed at once, and the user's answers are analyzed to highlight unanswered items for the doctor, allowing the doctor to present additional questions. Modification 3: A form in which the generating AI is controlled by template-based prompts or by creating template-based ground truth data and fine-tuning it, and all questions are dynamically generated by the generating AI. Variation 4: A format for providing medical interview results as PDF files or paper printouts for clinics that do not have electronic medical records.

[0018] These mechanisms are particularly necessary when healthcare institutions request that the telephone questionnaire be customized for each hospital or department, and that all items be filled out during the telephone consultation. The template can be adjusted to ensure that specific questions are answered during the telephone consultation. Furthermore, even if the template format is simple and the number of questions is small, it can function as a "consultation template" if the answer format is appropriately set. Dynamically generated questions are also considered part of the template, meaning that real-time question provision is also included in the template.

[0019] One form of social implementation of this telephone medical consultation system involves using generative AI or natural language processing models / programs to collect detailed information about the patient's symptoms and condition. In this case, a consultation template is not necessarily required. Instead, the system uses generative AI that automatically generates the next spoken word based on data collected from the patient, such as electronic medical records and electronic questionnaires, by providing prompts with information about the department the patient is expected to visit and their role (e.g., doctor's assistant, receptionist, doctor). This automatic generation function enables smooth and accurate information collection while maintaining the intent of the consultation, even when a template is not predefined. As a result, the system can flexibly respond regardless of the presence or absence of a template, and is expected to be put into practical use in various medical settings.

[0020] In other words, systems that require a medical interview template, or systems that don't necessarily require a template but use prompts to set information about the department being examined and the role of healthcare professionals (doctor's assistant, receptionist, doctor, etc.), enable automated conversations based on collected data. Even without a predefined template, the generating AI can understand the intent of the interview and smoothly extract information, improving its practicality in medical settings.

[0021] Furthermore, this telephone consultation system utilizes generative AI to facilitate conversations with patients and ask questions based on the intent of the consultation. The AI ​​searches for the type of conversation based on the patient's symptoms and situation, and generates and presents appropriate questions to efficiently extract the information necessary for diagnosis. In addition, the system has set expressions to indicate when to end the conversation in the prompts, and when these expressions are detected, the system automatically stops collecting the spoken content and ends the call. As a result, the consultation process is conducted efficiently and accurately, reducing the burden on healthcare professionals while allowing for an accurate understanding of the patient's condition.

[0022] In other words, AI can search for conversation types that match the patient's symptoms and situation, and generate and present appropriate questions, thereby collecting information in a way that is easy for the patient to understand. By asking precise questions to the patient while considering the purpose of the interview, AI can efficiently extract the information necessary for diagnosis.

[0023] Furthermore, by automatically determining when a conversation should end and setting an expression to indicate the end of the conversation, the AI ​​can automatically stop collecting the utterances and hang up the phone when the patient uses that expression. This allows the medical interview process to be conducted efficiently and accurately, reducing the burden on healthcare professionals while accurately understanding the patient's condition.

[0024] (3) A conversion unit that recognizes the content of the patient's speech and converts it into text data, A determination unit that determines whether a patient is a first-time patient or a follow-up patient based on either the text data or the patient's electronic medical record information, A selection unit extracts one or more symptom information or disease information from the aforementioned text data, and selects a medical interview template based on the extracted symptom information or disease information. An input unit that inputs the aforementioned text data into the aforementioned medical questionnaire template, A questioning unit that selects or generates additional questions based on the remaining unentered content in the aforementioned medical questionnaire template, The system includes a collection unit that presents the patient with the aforementioned additional questions and collects additional speech content. The input unit converts additional speech content into text using the conversion unit to create additional text data, and inputs this information into the medical questionnaire template based on the information obtained by the conversion unit and the determination unit. It is characterized by the following:

[0025] This invention automatically collects interview content by determining whether a patient is having their first or second consultation based on their spoken words and electronic medical record information. The determination is made because the interview process and the questions asked may differ depending on whether it is a first or second consultation. Interview templates may be categorized by medical department or by symptom / disease, for example.

[0026] (4) A registration unit that obtains the patient's date of birth and telephone number in advance, A dialing unit that automatically dials the telephone number based on the telephone number obtained by the registration unit, The authentication unit, which verbally asks the patient connected to the patient terminal by the aforementioned transmitting unit for their date of birth or name, or patient ID or part of patient ID, transcribes the spoken content into text using speech recognition, and verifies the patient's identity by comparing it with pre-registered information, The system is characterized in that it starts the medical interview only after the user's identity has been verified by the aforementioned authentication unit.

[0027] This invention incorporates two-factor authentication, automatically verifying the patient's identity before initiating the medical interview. By confirming personal information such as date of birth and patient ID via voice, it ensures the patient is indeed the correct individual, thereby enhancing security by preventing incorrect information entry and unauthorized access. Furthermore, it automatically redials the call and uses voice recognition for smooth identity verification, reducing the burden on the patient and allowing the medical interview to proceed quickly. In addition to enhanced security, patients do not incur any call charges.

[0028] (5) The input unit determines whether the input of the medical questionnaire template has been completed and confirms the progress of the medical questionnaire. This invention automatically manages the progress of the medical interview and notifies the system if input is incomplete, thereby reducing the burden of verification for healthcare professionals and patients and enabling efficient medical interview management. Furthermore, because it confirms that all necessary information has been entered before proceeding to the next step (medical interview), it ensures that all information necessary for diagnosis is collected without omission.

[0029] (6) The system has an evaluation unit that evaluates priority based on the contents entered in the questionnaire template, The questioning unit is characterized by selecting or generating additional questions based on the priority evaluated by the evaluation unit. This invention evaluates the urgency of the examination and the need for further additional questions based on the importance of the symptoms and information entered, by selecting or generating additional questions based on priority. If there is information that is given high priority (e.g., urgent symptoms), additional questions related to that information are asked. Conversely, if it is judged to be of low priority, the system can adjust the questions according to priority, such as limiting the additional questions to simpler ones. Questions can be asked to address high-priority symptoms immediately. Furthermore, by focusing on the necessary high-priority parts, efficient interviews are achieved, and the time of healthcare professionals and system resources are not wasted. Since symptoms and circumstances differ from patient to patient, flexible additional questions can be asked according to priority, leading to an improvement in the quality of the interview.

[0030] (7) The questionnaire section is characterized by using a questionnaire template that includes questions regarding the onset, alliative and provoke, quality and quantity, region, severity symptoms, and time course when the patient's chief complaint is pain.

[0031] This invention enables accurate and systematic questioning of patients complaining of pain, guiding appropriate diagnosis and treatment direction. When the patient's chief complaint is "pain," a question template containing specific questions is used. The questions in the template are based on the following six main items to collect details about the pain. • Onset: Questions about when and how the pain began. Examples: "Did the pain start suddenly or gradually?" "When did the pain begin?" • Alliative & Provoke: Questions about situations that alleviate or worsen pain. Example: "What helps relieve your pain? Are there any factors that worsen your pain?" • Quality & Quantity: Questions about the nature and intensity of pain. Examples: "Is the pain dull or sharp?" "On a scale of 0 to 10, how would you rate the intensity of the pain?" • Region: Questions about the location of the pain. Example: "Where does it hurt? Is the pain spreading?" • Severity Symptoms: Questions about other symptoms that occur along with the pain. Example: "Do you experience any other symptoms such as nausea or dizziness when you are in pain?" • Time course: Questions about the time course of pain. Example: "Is the pain constant, or does it come and go?" [Effects of the Invention]

[0032] According to the present invention, efficient medical interviews can be achieved by using disease symptom-specific interview templates, medical department-specific interview templates, comprehensive interview templates, and generative AI / LLM that do not require interview templates, and by collecting patient information during initial and follow-up visits, enabling healthcare professionals to quickly collect necessary information. In particular, flexible responses can be made according to the situation of initial and follow-up visits, reducing the burden on patients and improving the quality of medical care. Specifically, (1) the patient's chief complaint can be flexibly grasped using automatic speech recognition and natural language processing, (2) necessary information can be efficiently collected using templates or generative AI, and (3) accurate medical interviews can be achieved while reducing the burden on patients with additional functions such as two-factor authentication and priority evaluation. [Brief explanation of the drawing]

[0033] [Figure 1] A diagram showing the network configuration of the medical interview system. [Figure 2] A block diagram showing an example of the system configuration of a medical interview system. [Figure 3] A block diagram showing an example of the functional configuration of an in-hospital telephone terminal. [Figure 4] A block diagram showing an example of the functional configuration of a patient terminal. [Figure 5]A block diagram showing an example of the functional configuration of the medical interview system terminal 20. [Figure 6] A diagram showing an example of the data structure of a medical interview template. [Figure 7] A flowchart illustrating the processing steps of a medical interview application program. [Figure 8] A flowchart illustrating the processing steps of a medical interview application program. [Figure 9] A flowchart illustrating the processing steps of a medical interview application program. [Best Mode for Carrying Out the Invention]

[0034] Figure 1 shows a network configuration centered around the medical interview system 2. In addition to the computer running the medical interview system 2, the hospital telephone terminal 1 and patient terminal 3 are connected via a telephone / communication network through telecommunication lines, and are also connected to medical systems 4 owned by the medical institution as needed. Medical systems 4 include reservation systems, electronic medical record systems, and electronic medical interview systems. Each of these devices is connected to the internet / intranet / hospital LAN, etc., to form a network. Note that the control lines and information lines shown are those deemed necessary for explanation and do not necessarily represent all control lines and information lines in the actual product, and all components may be interconnected.

[0035] Medical interview system 2 is a computer that automatically transcribes the patient's speech into text and automatically inputs medical interview information, while also asking the patient questions about the interview content. By linking with medical system 4 via a network, medical interview system 2 can be used to build systems that utilize existing medical interview data and two-factor authentication systems that utilize patient information.

[0036] Figure 2 is a block diagram showing an example of the configuration of the medical interview system 2. The medical interview system 2 is linked to the in-hospital telephone terminal 1 and the patient terminal 3, and is connected to each terminal for calling / communication. The medical interview system 2 is a computer connected to a network such as the Internet, and is equipped with a communication IF 22, an input / output IF 23, memory 25, storage 26, and a processor 29.

[0037] Communication IF22 is an interface for inputting and outputting signals so that the medical interview system 2 can communicate with external devices. Input / Output IF23 functions as an interface to an input device for receiving input operations from the user and an output device for presenting information to the user. Memory 25 is for temporarily storing programs and data processed by programs, etc., and is a volatile memory such as DRAM. Storage 26 is a storage device for saving data, such as flash memory or an HDD. Processor 29 is hardware for executing the instruction set written in the program, and is composed of an arithmetic unit, registers, peripheral circuits, etc.

[0038] Note that "processor" refers to one or more processors. At least one processor is typically a microprocessor such as a CPU (Central Processing Unit), but may be other types of processors such as a GPU (Graphics Processing Unit). At least one processor may be single-core or multi-core. Furthermore, at least one processor may be a broader type of processor, such as a hardware circuit that performs some or all of the processing (e.g., an FPGA (Field-Programmable Gate Array) or ASIC (Application Specific Integrated Circuit)).

[0039] Figure 3 is a block diagram showing an example of the functional configuration of the in-hospital telephone terminal 1, comprising a communication IF 12, an input device 13, an output device 14, a memory 15, a storage unit 16, and a processor 19. The communication IF is an interface for inputting and outputting signals so that the in-hospital telephone terminal 1 can communicate with external devices and connect calls with the patient terminal 3. It also functions as an interface to the input device 13 for receiving input operations from the user and the output device 14 for presenting information to the user. The input device 13 is an input device (such as a keyboard, touch panel, touchpad, mouse, or other pointing device) for receiving input operations from the user. The output device 14 is an output device (such as a display or speaker) for presenting information to the user.

[0040] Memory 15 is for temporarily storing programs and data processed by programs, etc., and is a volatile memory such as DRAM. Storage unit 16 is a storage device for saving data, such as flash memory or HDD. Processor 19 is hardware for executing the instruction set written in the program, and is composed of an arithmetic unit, registers, peripheral circuits, etc.

[0041] The in-hospital telephone terminal 1 connects to the network 80 by communicating with communication devices such as a wireless base station 81 compatible with various communication standards including LTE, and a wireless LAN router compatible with IEEE and wireless LAN standards. If the in-hospital telephone terminal 1 is a desktop or laptop PC, it acquires voice signals using broadband or fiber optic lines. If it is a mobile device such as a tablet or smartphone, it can make calls and data communications using a mobile phone line (cellular network), and instead of a fixed telephone line, it connects to a mobile phone network such as 4G or 5G and utilizes mobile communication infrastructure. Furthermore, it is also possible to make calls over the internet using a VoIP application.

[0042] Patient terminal 3 is a tablet or smartphone, and generally has known and general-purpose functions that enable communication with the in-hospital telephone terminal 1. As shown in Figure 3, patient terminal 3 has a camera 31, a PC unit 32, a display unit 33, a touchscreen 34, a speaker 35, and a microphone 36. Patients can answer electronic questionnaires displayed on the display unit 33 by typing on the touchscreen 34 or by selecting from radio buttons, or they can answer questionnaire questions emitted from the speaker 35 by voice input from the microphone 36 and send the corresponding voice signal to the PC unit 32. The voice processing function of the PC unit 32 performs digital-to-analog conversion processing of the voice signal and provides questionnaire data to the questionnaire system 2 via the communication IF 22 / input / output IF 23. The voice processing function is implemented by a voice processing processor, and the speaker 35 converts the voice signal provided by the voice processing function into voice.

[0043] Figure 5 is a block diagram showing an example of the functional configuration of the medical interview system terminal 2. The medical interview system 2 comprises a communication means 220, a storage means 280, and a control means 290, with each block being electrically connected by a bus or the like. The communication means 220 implements the functions of the communication IF 22, the storage means 280 implements the functions of the memory 25 and storage 26, and the control means 290 implements the functions of the processor 29.

[0044] The communication means 220 performs modulation and demodulation processing to enable the medical interview system 2 to acquire voice data from other terminals, such as the medical system 4 and the in-hospital telephone terminal 1 / patient terminal 3, and transmits the signal calculated by the control means 290. The communication means 220 performs reception processing on the signal received from the outside and outputs it to the control means 290. In this way, the communication means 220 interprets commands or input content and provides them to each means, and also functions as an interface that interprets various display commands issued from the storage means 280 and performs output control.

[0045] The storage means 280 is implemented by memory (RAM) 25 and storage 26 such as a disk device (floppy disk, hard disk, or magneto-optical disk, etc.) and stores data, programs, etc. used by the chat system 20. The storage means 280 stores the application programs 282 of this system (medical questionnaire app, two-factor authentication app), as well as data for the work area 281, data storage area 283, and screen definition storage area 284.

[0046] The work area 281 is a region that is allocated when the system starts up and where various data input and output by the system are temporarily stored. The data storage area 283 is a region where data temporarily stored in the work area 281 is semi-permanently stored through write control when a save request is made. The screen definition area 284 is a region where screen definition information for various screens to be output to and displayed on the display unit 33 is stored in advance, and includes format information for the screen settings to be displayed.

[0047] The input device 230 is a device for the user operating the medical interview system 2 to input instructions or information, and may be a keyboard, mouse, reader, or touch-sensitive device. The input device 230 also converts the instructions input by the user into electrical signals and outputs the electrical signals to the control means 290. The input device 230 also includes a receiving port that accepts electrical signals input from external input devices. The output device 240 is a display device 241 such as an LCD or organic EL for presenting information to the user operating the medical interview system 2. The display 241 can display data according to the control content of the control means 290 and can check the communication status between the chat system 20 and other external devices.

[0048] The control means 290 is realized when the processor 29 reads the application program 282 stored in the storage means 280 and executes the instructions contained in the application program 282. The control means 290 also controls the operation of the medical interview system 20 and, by operating according to the application program 282, performs the functions of a patient information acquisition unit 291, a conversion unit 292, a selection / analysis unit 293 (also referred to as the "selection unit" and "analysis unit," corresponding to the selection function and analysis function), an input unit 294, a questioning unit 295, a collection unit 296, a two-factor authentication unit 297, a judgment unit 298, an authentication unit 298, and an evaluation unit 299.

[0049] The patient information acquisition unit 291 acquires patient information (name, age, gender, examination details, test results, current medical history and treatment details, and diagnostic imaging data (acquired from PACS)) from the electronic medical record (EMR) and PACS within the medical system 4. The information acquired is selected selectively according to the settings of the medical interview system.

[0050] The conversion unit 292 is responsible for collecting patient speech content and converting it into text data. It converts speech data collected from the patient terminal 3 via the network into text data using speech recognition technology (Speech-to-Text). The conversion unit can process in real time, and can also process data saved as recordings at a later date. Speech recognition can use a speech recognition engine (e.g., Google Cloud Speech-to-Text, IBM Watson Speech to Text, Amazon Transcribe, etc.). The speech recognition engine decomposes the speech signal into phonemes, matches them with an acoustic model to identify words, and then uses a language model to construct appropriate grammar and word sequences, thereby achieving more accurate text conversion. The conversion unit 292 is equipped with a medical terminology dictionary. When medical terminology or abbreviations are used, general speech recognition engines may not be able to recognize them accurately, so accuracy can be improved by adding dictionaries and models specialized for medical terminology. These speech recognition functions can also be implemented in conjunction with external APIs. Other functions such as voice output and template selection can also be implemented in conjunction with external APIs.

[0051] The selection / analysis unit 293 extracts information related to symptoms and diseases (symptom information, disease information) from text data and implements a selection function to select disease symptom-specific questionnaire templates related to the symptoms and diseases. It analyzes text using natural language processing technology and identifies symptoms such as "cough," "fever," and "headache," and diseases such as "influenza" and "pneumonia," based on the meaning and context of the sentence. Furthermore, it can utilize medical terminology dictionaries and Named Entity Recognition (NER) technology to extract proper medical nouns, symptoms, and disease names. In addition, it accurately extracts symptoms and diseases through keyword extraction and contextual analysis. For example, it extracts that even with different expressions like "I have a slight headache" and "I have a severe headache," the common symptom is "headache." The elements "slight" and "severe" can be used to evaluate priority. The analysis function of the selection / analysis unit 293 analyzes text data and extracts information related to symptoms or diseases.

[0052] The disease symptom-specific questionnaire templates can be highly accurate through a rule-based system that selects a respiratory questionnaire template when "fever" or "cough" is detected, or through a combination of flexible pattern recognition using machine learning. The machine learning model is trained on a large amount of data collected from patients, learning symptom patterns and their association with diseases, enabling it to select the appropriate template even for unknown cases.

[0053] The system uses departmental questionnaire templates to select the appropriate one based on the patient's symptoms and illness, with the selection unit 293 estimating which department is best suited. For example, if the text is "chest pain" and "palpitations," it estimates cardiology; if it's "cough" and "dyspnea," it estimates respiratory medicine; and if it's "abdominal pain" and "diarrhea," it estimates gastroenterology. In addition to the rule-based selection system, machine learning models can also be used to estimate departments based on complex symptom combinations and context. If a patient presents with multiple symptoms, the system can estimate departments according to priority, and in ambiguous cases, the system can present multiple templates or leave the final selection to the physician. Furthermore, for follow-up visits, the system selects the template previously used for that patient based on information from the appointment system and electronic medical record.

[0054] The input unit 294 inputs text data into the medical questionnaire template. The input format is automatically adjusted to match the template specifications, and the data format is adjusted so that the information extracted from the text data accurately fits into the fields of the medical questionnaire template. For example, if the patient speaks detailed information such as the start date of fever or the severity of symptoms, these are entered into separate fields. Also, if the text data is entered into a field with choices such as "yes" or "no," the appropriate choice is automatically selected based on the context.

[0055] The questioning unit 295 selects or generates additional questions based on the remaining unentered content in the medical history template. Alternatively, it generates additional questions to obtain further detailed information. While the additional questions are based on the unentered content of the template, content that does not affect the outcome may be skipped without generating additional questions. Additional questions can also be generated based on the content of referral letters or medical information reports held by the patient. In one example, when selecting or generating additional questions, an expression for determining whether to skip a question is stored in advance as a template, and this expression is evaluated to control the addition or selection of questions. This determination is made using one or more of the following methods: rule-based evaluation, machine learning, or generative AI evaluation.

[0056] The data collection unit 296 presents additional questions to the patient as visual or auditory information and collects additional speech content. The question-answering process by the data collection unit 296 generates additional questions in response to the patient's responses and repeats the process to elicit more detailed information, thereby preventing a lack of information necessary for diagnosis and enabling effective information collection. The data collection unit 296 adjusts the order and content of questions to make it easier for the patient to answer smoothly. To reduce the burden on the patient, it is possible to start with simple questions rather than complex ones and gradually move on to more detailed questions. Also, if the patient gives an incomplete answer to a question or takes a long time to answer, the questions are adjusted based on the patient's response, and the questioning unit 295 and data collection unit 296 generate follow-up questions in real time to obtain more detailed answers.

[0057] The input unit 294 inputs data into the medical questionnaire template from additional text information, which is textualized from the additional utterances. The additional text information is structured, new information obtained from the additional utterances is organized and mapped to the correct fields in the medical questionnaire template, and automatically entered.

[0058] The two-factor authentication unit 297 has the following functions: a registration unit 297A that obtains the patient's date of birth and telephone number in advance; a calling unit 297B that automatically calls the telephone number obtained by the registration unit 297A from the system side; and an authentication unit 297C that, when connected by the calling unit 297B, asks the patient for their date of birth or name and (part of) their patient ID by voice, transcribes the answer into text using voice recognition, and verifies their identity by comparing it with the information registered in advance. If the verification is successful, the system recognizes that identity verification is complete and proceeds with the medical interview process. If there is a mismatch, the system requests verification again or prompts the user to use another means of authentication.

[0059] The determination unit 298 analyzes electronic medical records and text data to determine whether the patient is making their first visit ("initial consultation"), has a record of a previous visit ("return consultation"), or has made a previous visit but is now making a new visit for a different symptom or illness ("initial consultation"). The determination unit 298 uses text data and electronic medical record data containing patient information (name, medical history, symptoms, etc.) to automatically determine whether it is an initial or return consultation using an algorithm or rule-based system. For example, if the text data contains the utterance "This is my first time...", it can be estimated that it is an initial consultation.

[0060] The evaluation unit 299 assesses priority (urgency of treatment and priority of response) based on the patient's medical history. Acute / chronic can also be used as a criterion for evaluating priority. Depending on the priority, additional questions can be created or selected based on the relative importance of the items.

[0061] The learning model for each control function of the control means is obtained by having a machine learning model perform machine learning according to a model learning program based on training data. The training data generally consists of pairs of speech data from many speakers and corresponding, correctly speech-recognized text data, which are commonly used when performing speech recognition by machine learning. There may be multiple learning models and training data. In particular, it is preferable for the storage means to have multiple learning models and training data corresponding to these learning models, as it is preferable to appropriately select the learning model (and thus the training data that underlies it) according to the questionnaire items of the questionnaire template. The learning model in this embodiment is, for example, a parameterized composite function composed of multiple functions. A parameterized composite function is defined by a combination of multiple adjustable functions and parameters. The prediction model in this embodiment may be any parameterized composite function that satisfies the above requirements, but it is assumed to be a multi-layer network model (multilayer network). A prediction model using a multilayer network has an input layer, an output layer, and at least one intermediate or hidden layer provided between the input layer and the output layer. The prediction model is intended to be used as a program module that is part of artificial intelligence software.

[0062] As the multilayer network according to this embodiment, for example, a deep neural network (DNN), which is a multilayer neural network targeted by deep learning, may be used. As the DNN, for example, a convolutional neural network (CNN) that targets images may be used. The above is merely an example of a prediction model, and the prediction model may have other configurations. For example, the prediction model may be a rule-based model in which patient information, medical information, case data, and environmental information are used as variables, and each variable is described by a function to which coefficients derived from past performance are attached.

[0063] Figure 6 shows an example of a medical interview template. The medical interview template uses structured data, associating input fields with input content, and automatically inputting the content that is recognized and converted into text when the patient speaks aloud. The medical interview template is structured data having input fields and input content associated with these input fields. Here, structured data is data that is predefined and formatted to have a certain predetermined structure before being placed in storage. Unstructured data, on the other hand, is data that is stored in its native format and is not processed until it is used.

[0064] The electronic medical questionnaire's input fields are defined based on this questionnaire template. The association between the electronic medical record template and the electronic medical questionnaire means that at least the input fields of the electronic medical record template and the questionnaire fields of the electronic medical questionnaire are associated. In other words, assuming that the entered electronic medical questionnaire data will be imported into the electronic medical record, the fields that the patient is asked to input are selected from the input fields of the electronic medical record template as questionnaire fields, and the association between the input fields and questionnaire fields is made with the understanding that the patient's responses to these questionnaire fields will be imported as input into the electronic medical record. Furthermore, the corresponding patient responses are associated as input fields. If the patient's utterances in response to the questionnaire questions are directly entered into the electronic medical record, it is not possible to rule out the possibility that they may not be medically accurate, and it is possible to correct the input fields by referring to the patient's past electronic medical records. Furthermore, the present invention includes various variations for linking to electronic medical records, not only direct copying / automatic transfer, but also file output, API linking, and writing to a cloud database. In addition, when linking to electronic medical records, there are embodiments such as adding to the S portion of the (Subjective, Objective, Assessment, Plan) in SOAP format progress notes, or installing a button to prompt the user to add to the progress notes. This embodiment may acquire the patient's voice during a medical interview, analyze its free speech content to automatically extract descriptions of the chief complaint, present illness, past medical history, and allergy history, and input them into the corresponding items in the S column (Subjective) of the electronic medical record. Voice is received from a terminal application with a microphone, a web browser, or an automated voice response system (IVR) using a telephone line. The received voice is converted into text, and in particular, disease names are classified to automatically determine whether they correspond to the present illness or past medical history. For example, if the patient says, "I've had a headache for three days. I had a stroke before. I think my only allergy is hay fever," the system will analyze the content and extract "headache for three days" as a present illness element, "I had a stroke before" as a past medical history element, and "hay fever" as an allergy history element. If the chief complaint is a headache, the system may, if necessary, read the beginning or emphasized part of the voice-recognized text and automatically place it as an introductory memo in the S column of the electronic medical record. The data classified in this way is written to the S column of the electronic medical record software by the medical record input means (input / output IF23). Specifically, it is linked to the electronic medical record API that healthcare professionals use on a daily basis, and is automatically entered in text format into each item (present illness field, past medical history field, allergy field, etc.). Another embodiment of the present invention illustrates a configuration in which a medical interview system automatically analyzes conversation history or electronic medical record information using a large-scale language model (LLM), a machine learning model, or a rule-based algorithm (analysis unit 293), identifies "information that has not yet been collected," and then has an AI automatically generate additional questions corresponding to this information (question unit 295). First, the content of the patient's speech is converted into text by a conversion unit (speech recognition), and the collection unit stores the textualized conversation history. When extracting symptom and disease information contained therein, the system checks whether some fields of the medical interview template have already been filled in and identifies items that have been determined to be "not entered" or "insufficient." Here, the system uses an LLM (Large-Scale Language Model) as a generative AI, inputting "conversation history to date," "list of unfilled items on the template," and "location of the chief complaint or disease name" into the prompt. The LLM then suggests "which questions should be prioritized from the missing information" and "how to phrase the questions naturally," and the questioning unit presents these suggestions to the patient. For example, a sentence like, "In your previous response, we discussed the severity of the pain, but it seems we haven't yet heard the onset date. Could you tell me when the symptoms started?" is dynamically generated. In a medical setting, rule-based flows and machine learning-based scoring can be used in conjunction, and various strategies can be employed, such as asking more urgent additional questions first if it is determined that there is a high possibility of the symptoms becoming severe. On the other hand, if the patient has already stated the same information in the conversation history, the AI ​​may suggest skipping it as "duplicate information." This avoids the burden of repeatedly asking the same questions to the patient and streamlines the interview process. Furthermore, even if templates are not predefined, the generating AI can utilize the medical knowledge learned within the LLM to spontaneously suggest follow-up questions regarding "medical history," "comorbidities," and "severity." In this embodiment, regardless of the communication method, such as telephone lines (IVR) or web voice input, the system uses the obtained conversation history to determine missing items, and then the generating AI generates additional questions. This extends existing questionnaire template systems, comprehensively covering everything from automatic detection of missing information to dynamic generation of additional questions. As a result, healthcare professionals can monitor the entire interaction, avoid unnecessary duplicate questions, and achieve an efficient UI flow that ensures all necessary information is collected. Another embodiment of the present invention illustrates a mechanism in which, after converting the patient's speech into text, the text is analyzed to detect the difference between the "required information items" in the medical history template and the information already collected, and if there is missing information, the generating AI creates additional questions. First, the patient speaks about their chief complaint and symptoms, the conversion unit converts this into text data, and then the analysis unit extracts context and keywords to organize "what kind of information was obtained." Here, the judgment unit compares this with representative question items included in the medical history template (for example, "presence or absence of fever," "onset time," "degree of pain," "presence of complications," etc.) and recognizes information items that have not yet been acquired or collected as "missing information." For example, if a patient says, "I've had a bad cough since last night, and I think I have a slight fever this morning," the analysis unit will detect symptoms such as "cough" and "fever," while the judgment unit will determine that details such as the exact temperature and when it was measured are "not collected." Then, the question output generation unit of this embodiment generates additional questions using the generation AI. Specifically, it gives instructions to the prompt such as, "The patient says they have a fever, but they haven't mentioned the exact temperature or when it started. Create questions that ask for that information in a way that is easy for the user to answer," and the AI ​​will suggest natural-sounding phrases such as, "When did you first take your temperature, and what was it then?" This system repeatedly presents additional questions using AI to fill in these "missing items" one by one. The data collection unit then transcribes the answers back into text and sends them to the analysis unit. The analysis and judgment units continuously check for differences, excluding items that were filled in by the previous response, and repeat the process until all essential information is obtained. This reduces the need to ask the user about things they have already stated, resulting in a smoother dialogue flow. This embodiment has the following configuration: "analysis unit that extracts necessary information from text data," "determination unit that determines missing information by comparing it with a medical interview template," and "question output generation unit that generates additional questions based on the missing information using AI." Because the AI ​​can automatically generate closed questions to supplement missing items after open questions, flexible medical interviews become possible in medical settings, and input into electronic medical records is less prone to errors and omissions. The algorithm used by the generating AI is not limited to large-scale language models; rule-based or machine learning algorithms may also be used. In any case, the characteristic of this embodiment is that it automatically determines missing information and automatically generates appropriate questions. The above example uses a method that outputs any missing items as additional questions. However, it is also possible to control the output by skipping items that have already been answered.

[0065] In the example shown in Figure 6, the data is divided into sections such as "Symptoms," "Medical History," and "Allergies." The "Symptoms" section includes items such as "Fever," "Cough," "Pain Intensity," and "Duration," the "Medical History" section includes items such as "Pre-existing Diseases" and "Past Treatment History," and the "Allergies" section includes items such as "Medications" and "Foods." The structured data is automatically entered into these subdivided items. For example, if a patient says, "I've had a fever of 38 degrees for three days," the text "Fever" is entered / the checkbox is checked, and "3 days" is entered in the duration field. Any missing information can be filled in by asking additional questions. Additional information obtained from these questions can be color-coded to help check the consistency of the answers. Alternatively, if inconsistencies are observed during conversion by the conversion unit even if the patient's utterance is consistent, improvements can be made to enhance the accuracy of speech recognition and natural language processing, such as addressing slips of the tongue or dialects.

[0066] The medical interview template and the electronic medical record template are linked, and the input content from the medical interview template can be reflected in the electronic medical record template, allowing for the simultaneous creation of the electronic medical record. The contents of the electronic medical record are generated based on the electronic medical record template, and similarly, the electronic medical questionnaire data is generated based on the electronic medical questionnaire based on the medical interview template. In addition, the electronic medical questionnaire is linked to the electronic medical record template (association between items is established), and the electronic medical questionnaire is generated from the electronic medical record template as needed. Once the medical interview is completed, the collected electronic medical questionnaire data can be integrated into the electronic medical record system in real time.

[0067] Electronic medical questionnaires, created based on a questionnaire template, define questionnaire items and their corresponding answers as structured data. This structured data is either pre-organized with questions and answer choices, or it's automatically entered into the questionnaire template based on real-time recognition of the patient's voice during a telephone consultation, using the structured data. A speech recognition engine can be used to record the patient's responses to the questionnaire items. The patient's voice responses are appropriately categorized based on pre-set options (yes / no), numerical values ​​from 1 to 10 (sliders), or free-text responses.

[0068] In electronic medical questionnaires, blank spaces in the structured data indicate items that the patient has not yet answered. For these blank spaces, the system automatically generates questions related to each blank space based on the structured data. This uses an algorithm that creates appropriate questions depending on the type of questionnaire item (e.g., symptoms, medical history, allergies, etc.). For example, if the medical history field is not entered, the system automatically generates a question such as, "What illnesses or injuries have you had in the past?"

[0069] In telephone consultations, a speech recognition system is used to obtain patient responses. Patient responses are transcribed in real time and entered into blank spaces in structured data. In text-based consultations, the content entered by the patient is automatically reflected in the blank spaces. Depending on the patient's answers to questions, new questions can be dynamically generated to gather more detailed information. For example, if the patient answers "yes" to the question "Are you in pain?", follow-up questions such as "How severe is the pain?" are automatically generated.

[0070] Regarding blank fields, the system can automatically fill them in. If a patient provides specific medical history, it is possible to retrieve and supplement related information (past treatments, past medication history, etc.) from the patient's electronic medical record.

[0071] An electronic medical questionnaire is structured data containing questionnaire items, associated questionnaire questions, and questionnaire response content. Each questionnaire item is an index for the electronic questionnaire and is relatively short, using medical terminology, to allow healthcare professionals to identify which item is being referred to. The questionnaire items may be the same as the input items in the electronic medical record template; in fact, it is preferable for the questionnaire items and input items to be the same from the standpoint of identifying the questionnaire response content.

[0072] The questions in the medical history template are questions that would actually be asked of a patient, and are written in a way that patients can easily understand when they read or listen to them. The questions can be in various formats, such as multiple-choice or open-ended. The patient's utterances are the actual responses the patient made to the questions; if the medical history question is multiple-choice, the patient selected one of the options, and if it is open-ended, the patient entered free-form text.

[0073] Figure 7 is a flowchart showing the processing procedure of the medical interview application program executed by the medical interview system 2. The medical interview system 2 virtually connects the in-hospital telephone terminal 1 and the patient terminal 3 and conducts the medical interview through dialogue with the patient.

[0074] First, once the system is started and connected, it collects the speech content from the patient terminal 3 (step S701). The collection is performed by the collection unit 296, which stores the data in the work area 281 of the storage means 280, and also converts the collected speech data into text data (step S702).

[0075] Next, determine whether the patient has finished speaking (step S704). If not, return to step S701 and continue collecting the speech content. If finished, proceed to the next step. The completion can be determined by whether or not the set prompt has been spoken.

[0076] Next, prior to selecting a medical history template, a determination of whether it is an initial or follow-up visit may be made (step S704). The determination of whether it is an initial or follow-up visit is made by the determination unit 298, which determines whether it is an initial or follow-up visit from the text data converted by recognizing the patient's utterances, or from the patient's electronic medical record information. After that, the selection / analysis unit 293 selects a medical history template (step S705). The medical history templates selected are disease symptom-specific medical history templates based on symptom information or disease information, department-specific medical history templates based on text data, reservation system information, or electronic medical record information, and in the case of a follow-up visit, a medical history template that has already been used is selected. Note that the processing in steps S704 and S705 can be skipped. For example, if additional questions have been asked, the medical history template has already been determined and the steps are skipped. Alternatively, if the utterance content is to be analyzed by the generating AI, the process proceeds to step S70X.

[0077] Next, the input unit 295 automatically inputs text data into the selected questionnaire template (step S706). After that, it determines whether there are any unentered / uncollected items remaining in the questionnaire template and whether there is any content for the unentered items (step S707). If there are no unentered / uncollected items, this subroutine terminates.

[0078] If there are any unentered / uncollected entries, the questioning unit 295 selects additional questions for the unentered (uncollected) items or generates additional questions for the unentered (uncollected) items (step S708). The system displays the selected / generated additional questions audibly or visually on the patient terminal 3 (step S709), starts collecting the patient's spoken content (step S701), and then repeatedly executes steps S702 to S706 until it is determined that there are no unentered / uncollected entries (step S707). This allows the system to automatically and efficiently collect necessary information from the patient, proceed with the medical interview, and provide appropriate medical care.

[0079] Figure 8 is a flowchart showing the processing steps of the medical interview application program executed by the medical interview system 2. Before starting the medical interview, two-factor authentication is performed to verify the user's identity. Once authentication is completed by this subroutine, the process shown in Figure 7 is performed.

[0080] First, the two-factor authentication unit 297 of the medical interview system 2 waits for an incoming call from the patient terminal 3 (step S801) and determines whether or not there is an incoming call (step S802). If there is no incoming call, it continues to wait for an incoming call. If it is determined that there is an incoming call, the registration unit 297A of the two-factor authentication unit 297 automatically obtains the incoming phone number or obtains it through input by the patient (step S803), and also obtains patient information such as the date of birth (step S804). In addition to the date of birth, the patient information can also include the name and patient ID (part of it), or the process in step S804 may be skipped and only the incoming phone number may be obtained.

[0081] Next, the call is disconnected (step S805). The calling unit 297B automatically redials the acquired phone number from the calling unit of the medical interview system 2 (step S806). Once the call is complete, it is determined whether a connection has been established (step S807), and if the connection is unsuccessful, it attempts to make the call a predetermined number of times (step S806). If it is determined that a connection has been established, the authentication unit 297C asks questions about patient information by voice (step S808). The questions from the questioning unit 295 are the date of birth, name, and patient ID (part of it), and the collection unit 296 collects the patient's spoken content (step S809), converts the answer into text using speech recognition (step S810), and performs authentication by comparing it with pre-registered information (step S811). For the comparison, patient information is retrieved from the medical system 4 based on the acquired phone number, etc., and it is determined whether authentication is OK or not based on the match / mismatch with the converted text content (step S812). If authentication fails, this subroutine is immediately terminated and the call is disconnected, or the process returns to step S808 and the authentication unit performs the process a predetermined number of times. If authentication is successful, the process moves to the electronic medical interview process, where detailed questions about symptoms and medical history are asked.

[0082] Figure 9 is a flowchart showing the processing steps of the medical interview application program executed by the medical interview system 2. It illustrates the process of analyzing the patient's subjective information (column S), automatically generating objective findings (column O) based on that information, skipping / hiding unnecessary items, and transferring the data to the electronic medical record.

[0083] In the first step (S901) of inputting the patient's subjective information, the medical interview system terminal (control means 290) shown in Figure 5 is activated, and the patient or medical professional inputs the patient's S column information (subjective information) via the user interface. At this time, the input device (input unit 230) and the collection unit (collection unit 296) work together to capture the patient's complaints and symptoms into the system, regardless of whether they are entered by voice or text. The control means 290 temporarily stores this information in a storage means 280 such as memory, and the S column data is prepared. Next, the subjective information entered in column S is analyzed (S902), and the patient information acquisition unit 291 and the selection / analysis unit 293 work together to estimate the suspected pathological condition based on the chief complaint and symptoms, using large-scale language models, machine learning models, or rule-based systems. Specifically, the selection / analysis unit 293 processes the text data using natural language processing to extract candidate examination items and physical findings. Based on the analysis results, the template generation means is activated (S903). Here, the control means 290 functions as the core and manages the main data in template generation. Based on the content obtained by the analysis unit 293, a list of examination items and physical findings items required for objective findings (column O) is automatically generated, and the control means 290 organizes the list of findings items held in the screen definition storage area 284 while referring to the interview template and display screen definition. In the next step (S904), items in the generated item list that do not match the information in column S are skipped or hidden. This is done by the control means 290, which controls the entire system, in cooperation with internal modules such as the judgment unit 298 and the evaluation unit 299 to exclude unnecessary items. Tests and findings that deviate significantly from the patient's chief complaint or medical department are either not displayed on the screen at all, or are controlled to allow the doctor to switch their display on or off, thereby eliminating unnecessary information and reducing the operational burden. Next, the necessary findings items are presented via the display unit (S905). Here, the display device 240 and input / output IF23 work together to create a block that functions as a "display unit," displaying a list of candidate items on a screen for medical professionals. Doctors and nurses can use the touch panel or mouse to select the necessary items while reviewing the presented findings and examination items, or to enter additional comments. Next, the medical professional confirms the selected and entered information (S906). At this point, the input unit 294 and other components function again, reflecting the information entered by the doctor in the checkboxes and text forms on the screen into the data storage area 283. Finally, the content to be reflected in column O (physical findings, test orders, etc.) is determined, and the system also clarifies "which information to transfer to the medical record." In step S907, the medical interview system communicates with the electronic medical record system (medical system 4) to transfer the confirmed information to column O of the electronic medical record. The control means 290 writes the confirmed column O information to the medical record using API communication or file output. In terms of the functional elements in Figure 5, unlike the patient information acquisition unit 291, the input findings information is output to the medical system 4. Once the physician reviews the content and determines that there are no additional questions or items to skip, the processes in the collection unit 296 and input unit 294 are stopped, and the interview results are saved. This completes the flow in which columns S and O are reflected in the electronic medical record in a consistent manner, and findings and test orders are automatically and accurately set in accordance with the patient's subjective information. [Industrial applicability]

[0084] This invention includes a function that converts the patient's speech content into text using speech recognition technology and natural language processing technology, automatically inputs and collects medical interview information based on that text, and generates and presents appropriate additional questions to the patient. As such, it can realize a flexible and natural medical interview process that surpasses the limitations of conventional electronic medical questionnaires, and is useful as a tool that balances efficiency in the medical field with reducing the burden on patients.

[0085] [Note B1] A program for operating a computer that includes a processor and memory, The processor, in the program, A conversion function that collects patient speech content and converts it into text data, A selection function that extracts one or more symptom information or disease information from the aforementioned text data, and selects a questionnaire template for each disease symptom based on the extracted symptom information or disease information. An input function that inputs the aforementioned text data into the aforementioned disease symptom-specific questionnaire template, A question function that selects or generates additional questions based on the remaining unentered content in the aforementioned disease symptom-specific questionnaire templates, The system includes a function to present the patient with the aforementioned additional questions and to collect additional speech content. The program is characterized by having the input function convert additional spoken content into text using the conversion unit to create additional text data, and then inputting data into the disease symptom-specific questionnaire template based on the additional text data. [Note B2] A program for operating a computer that includes a processor and memory, The processor, in the program, A conversion function that collects patient speech content and converts it into text data, A selection function to select a medical department-specific questionnaire template from one or more of the following: the aforementioned text data, the information from the reservation system, or the information from the electronic medical record. An input function that inputs the aforementioned text data into the aforementioned medical department-specific questionnaire template, A question function that selects or generates additional questions based on the remaining unentered content in the aforementioned medical department-specific questionnaire templates, The system includes a function to present the patient with the aforementioned additional questions and to collect additional speech content. The program is characterized in that the input unit converts additional spoken content into text using the conversion unit to create additional text data, and then performs data input into the medical department-specific questionnaire template based on the additional text data. [Note B3] A program for operating a computer that includes a processor and memory, The processor, in the program, A conversion function that recognizes the patient's speech and converts it into text data, A determination function that determines whether a patient is a first-time patient or a follow-up patient based on either the text data or the patient's electronic medical record information, A selection function that extracts one or more symptom information or disease information from the aforementioned text data, and selects a medical questionnaire template based on the extracted symptom information or disease information. An input function that inputs the aforementioned text data into the aforementioned medical questionnaire template, A question function that selects or generates additional questions based on the remaining unentered information in the aforementioned medical questionnaire template, The system includes a function to present the patient with the aforementioned additional questions and to collect additional speech content. The input function is a program characterized by creating additional text data by converting additional speech content into text using the conversion function, and inputting this information into a medical questionnaire template based on the information obtained by the conversion function and the judgment function. [Note B4] A registration function to obtain the patient's date of birth and phone number in advance, A calling function that automatically dials the telephone number obtained by the aforementioned registration function, The aforementioned transmission function provides an authentication function that, when a patient's terminal is connected, verbally asks the patient for their date of birth or name, or their patient ID or part of their patient ID, transcribes the spoken content into text using speech recognition, and verifies their identity by comparing it with pre-registered information. A program with two-factor authentication, as described in any of the [Appendix B1] to [Appendix B3], characterized by enabling a function to start a medical interview when identity verification is completed by the aforementioned authentication function. [Note B5] The input function is a program as described in any one of [Appendix B1] to [Appendix B3], characterized in that it determines whether the input of the medical questionnaire template has been completed and confirms the progress of the medical questionnaire. [Note B6] It has an evaluation function that evaluates priority based on the content entered in the aforementioned medical questionnaire template, The program according to any one of [Appendix B1] to [Appendix B3], characterized in that the question function selects or generates additional questions based on the priority evaluated by the evaluation function. [Note B7] The aforementioned questioning function is characterized by using a questionnaire template that includes questions regarding the onset, alliative and provoke, quality and quantity, region, severity symptoms, and time course when the patient's chief complaint is pain, as described in any of the programs in [Appendix B1] to [Appendix B3].

[0086] [Note C1] A method performed by a computer comprising a processor and memory, The program stored in the aforementioned memory, A conversion step that collects the patient's speech content and converts it into text data, A selection step involves extracting one or more symptom information or disease information from the aforementioned text data, and selecting a questionnaire template for each disease symptom based on the extracted symptom information or disease information. An input step is to input the aforementioned text data into the aforementioned disease symptom-specific questionnaire template, A question step to select or generate additional questions based on the remaining unentered content in the aforementioned disease symptom-specific questionnaire template, The process includes a collection step in which the patient is presented with the aforementioned additional questions and additional utterances are collected. The method is characterized in that the input step includes the step of creating additional text data by converting additional utterances into text, and inputting data into the disease symptom-specific questionnaire template based on the additional text data. [Note C2] A method performed by a computer comprising a processor and memory, The program stored in the aforementioned memory, A conversion step that collects the patient's speech content and converts it into text data, A selection step to select a medical department-specific questionnaire template from one or more of the following: the aforementioned text data, the information from the reservation system, or the information from the electronic medical record. The input step involves inputting the aforementioned text data into the aforementioned medical department-specific questionnaire template, A question step to select or generate additional questions based on the remaining unentered content in the aforementioned medical department-specific questionnaire template, The system includes a collection step in which the patient is presented with the aforementioned additional questions and additional utterances are collected. The input step is characterized by including the step of converting additional utterances into text using the conversion unit to create additional text data, and inputting data into the medical department-specific questionnaire template based on the additional text data. [Note C3] A method performed by a computer comprising a processor and memory, The program stored in the aforementioned memory, A conversion step that recognizes the patient's speech and converts it into text data, A determination step of determining whether the patient is a first-time patient or a follow-up patient based on either the text data or the patient's electronic medical record information, A selection step involves extracting one or more symptom information or disease information from the aforementioned text data, and selecting a medical questionnaire template based on the extracted symptom information or disease information. An input step of inputting the aforementioned text data into the aforementioned medical questionnaire template, A question step to select or generate additional questions based on the remaining unentered content in the aforementioned medical questionnaire template, The process includes a collection step in which the patient is presented with the aforementioned additional questions and additional utterances are collected. The input step is characterized by including the step of converting additional speech content into text using the conversion function to create additional text data, and inputting the information into the medical interview template based on the information obtained by the conversion function and the judgment function. [Note C4] A registration step to obtain the patient's date of birth and telephone number in advance, A calling step which involves automatically calling the telephone number obtained by the registration function, The authentication step involves verbally asking the patient connected via the aforementioned transmission step for their date of birth or name, or their patient ID or part of their patient ID, transcribing the spoken content into text using speech recognition, and verifying their identity by comparing it with pre-registered information. A method for two-factor authentication described in any of the [Appendix C1] to [Appendix C3], characterized by including a step of initiating a medical interview when identity verification is completed in the aforementioned authentication step. [Note C5] The input step is characterized by determining whether the input of the medical questionnaire template has been completed and confirming the progress of the medical questionnaire, as described in any one of the methods described in [Appendix C1] to [Appendix C3]. [Appendix C6] The system includes an evaluation step that evaluates priority based on the content entered in the aforementioned medical questionnaire template. The method according to any one of [Appendix C1] to [Appendix C3], characterized in that the questioning step selects or generates additional questions based on the priority evaluated in the evaluation step. [Note C7] The aforementioned questioning step is characterized by using a medical history template that includes questions regarding the onset, alliative and provoke, quality and quantity, region, severity symptoms, and time course, when the patient's chief complaint is pain. This method is described in any of the [Appendix C1] to [Appendix C3].

[0087] [Note D1] The patient's speech is collected and converted into text data. From the aforementioned text data, extract one or more symptom information or disease information, and based on the extracted symptom information or disease information, select a questionnaire template for each disease symptom. The text data is entered into the disease symptom-specific questionnaire template, Based on the remaining unentered information in the aforementioned disease symptom-specific questionnaire templates, additional questions are selected or generated. The patient is presented with the aforementioned additional questions, and additional verbal responses are collected. A method characterized by converting additional speech content into text to create additional text data, and then inputting data into the disease symptom-specific questionnaire template based on the additional text data. [Note D2] The patient's speech is collected and converted into text data. Select a medical department-specific questionnaire template from one or more of the following: the aforementioned text data, the information from the reservation system, or the information from the electronic medical record. The aforementioned text data is entered into the medical department-specific questionnaire template, Based on the remaining unentered information in the aforementioned medical department-specific questionnaire templates, additional questions are selected or generated. The aforementioned additional questions are presented to the patient, and additional verbal content is collected. A method characterized by converting additional spoken content into text to create additional text data, and then inputting data into the medical department-specific questionnaire template based on the additional text data. [Note D3] The system recognizes the patient's speech and converts it into text data. The system determines whether the patient is a first-time patient or a returning patient based on either the text data or the patient's electronic medical record information. From the aforementioned text data, extract one or more symptom information or disease information, and based on the extracted symptom information or disease information, select a medical questionnaire template. The text data is entered into the questionnaire template. Based on the remaining unentered information in the aforementioned questionnaire template, additional questions are selected or generated. The aforementioned additional questions are presented to the patient, and additional verbal content is collected. A method characterized by inputting the information obtained by the conversion and determination into a medical questionnaire template.

[0088] [Note A1-1] A conversion unit that collects the patient's speech content and converts it into text data, An analysis unit that analyzes the aforementioned text data to extract symptom information or disease information, The information extracted by the analysis unit (symptom information / disease information) is compared with the medical questionnaire template, and among the questions included in the medical questionnaire template, (1) A method that skips and does not output items that have already been answered, (2) A method that outputs any missing items as new additional questions, A determination unit that controls the questions asked to the patient using one of the following: A question unit outputs a skip or additional question based on the control result of the determination unit, The system includes a collection unit that collects additional spoken content in response to questions presented by the aforementioned questioning unit and converts it into text data. A medical interview system characterized in that the additional speech content obtained by the collection unit is again supplied to the analysis unit, and the determination unit repeatedly performs control using either the skip method (1) or the missing item addition method (2). [Appendix A2-1] A medical interview system comprising: a conversion unit that collects the content of a patient's speech and converts it into text data; an analysis unit that analyzes the text data and extracts information related to symptoms or diseases; a determination unit that compares the information extracted by the analysis unit with a pre-set medical knowledge base or diagnostic guidelines and determines which information items necessary for diagnosis or evaluation have not been collected; a question generation unit that generates additional questions using a generating AI based on the information items determined to be uncollected by the determination unit; and a collection unit that presents the additional questions to the patient and collects additional speech content, wherein the collection unit converts the additional speech content into text data using the conversion unit and supplies it again to the analysis unit and determination unit, and repeats this process. [Appendix A1-2] A method performed by a computer comprising a processor and memory, The program stored in the aforementioned memory, A conversion step that collects the patient's speech content and converts it into text data, The analysis step involves analyzing the aforementioned text data to extract symptom information or disease information, The information extracted by the analysis unit (symptom information / disease information) is compared with the medical questionnaire template, and among the questions included in the medical questionnaire template, (1) A method that skips and does not output items that have already been answered, (2) A method that outputs any missing items as new additional questions, A determination step that controls the questions asked to the patient using one of the following: A question step that outputs a skip or additional question based on the control result of the judgment step, The system includes a collection step which collects additional spoken content in response to the questions presented in the questioning step and converts it into text data. A method characterized by supplying the additional speech content obtained in the collection step back to the analysis step, and repeatedly performing control in the determination step using either the skip method (1) or the missing item addition method (2). [Appendix A2-2] A method performed by a computer comprising a processor and memory, The program stored in the aforementioned memory, A method comprising: a conversion step of collecting the patient's utterances and converting them into text data; an analysis step of analyzing the text data to extract information related to symptoms or diseases; a determination step of comparing the information extracted by the analysis step with a pre-configured medical knowledge base or diagnostic guidelines to determine which information items necessary for diagnosis or evaluation have not been collected; a question generation step in which a generating AI generates additional questions based on the information items determined to be uncollected by the determination step; and a collection step of presenting the additional questions to the patient to collect additional utterances, wherein the collection step converts the additional utterances into text data using the conversion step and supplies it again to the analysis step and the determination step, and repeats this process. [Appendix A1-3] A program for operating a computer that includes a processor and memory, The processor, in the program, A conversion function that collects patient speech content and converts it into text data, An analysis function that analyzes the aforementioned text data to extract symptom information or disease information, The information extracted by the aforementioned analysis function (symptom information / disease information) is compared with the medical questionnaire template, and among the questions included in the medical questionnaire template, (1) A method that skips and does not output items that have already been answered, (2) A method that outputs any missing items as new additional questions, A judgment function that controls questions to the patient using one of the following, A question function that outputs a skip or additional question based on the control result of the aforementioned judgment function, The system includes a collection function that collects additional spoken content in response to questions presented by the aforementioned questioning function and converts it into text data. A program characterized by supplying the additional speech content obtained by the collection function back to the analysis function, and implementing a function that repeatedly controls the judgment function using either the skip method (1) or the missing item addition method (2). [Appendix A2-3] A program for operating a computer that includes a processor and memory, The processor, in the program, A program comprising: a conversion function that collects the content of a patient's speech and converts it into text data; an analysis function that analyzes the text data and extracts information related to symptoms or diseases; a determination function that matches the information extracted by the analysis function using LLM and determines which information items necessary for diagnosis or evaluation have not been collected; a question generation function in which a generating AI generates additional questions based on the information items determined to be uncollected by the determination function; and a collection function that presents the additional questions to the patient and collects additional speech content, wherein the collection function converts the additional speech content into text data using the conversion function and supplies it again to the analysis function and determination function, thereby enabling repeated control of this process.

[0089] This invention provides a means to streamline the entire medical treatment process by automatically associating the patient's subjective information (Column S) with the necessary examination and physical findings (Column O). Specifically, the system of this invention includes a means to analyze the content entered by the patient in the subjective information (Column S) and automatically generate corresponding objective findings (Column O) as a template. In particular, if the patient's utterance or input item is "abdominal pain," the system automatically lists the abdominal findings and examination items, thereby mechanically associating Column S and Column O to ensure that medical professionals do not overlook any examinations or findings. Furthermore, it has a function to skip or hide Column O items that do not match the content of Column S and dynamically reset them according to the requirements of each medical department and disease. This avoids the display of unnecessary examination items and physical findings as in conventional systems, and reduces the operational burden on medical professionals. [Note S1] A means for inputting subjective patient information (S column), A template generation means that analyzes the subjective information of the patient obtained by the S input means and automatically generates a list of examination items or findings items corresponding to objective findings (column O) based on the analysis results, A display unit that skips or hides items from the observation items generated by the template generation means that do not match the subjective information, and displays only the items that are determined to match, The system includes a user interface that allows a physician to select or input necessary findings and test items via the aforementioned display unit. A medical interview system characterized by transferring the input content to column O of the electronic medical record. [Note S2] In the medical interview system described in [Appendix S1], The template generation means is configured to dynamically add multiple examination items and physical findings items according to the pathological condition estimated from the subjective information (column S) of the patient, using one or more of a large-scale language model or machine learning model, a rule-based system, or a search function, and to allow the physician to select or input these additional items to supplement the objective findings (column O). Furthermore, the medical interview system is characterized by automatically skipping input if it is determined that the additional items are unnecessary. [Appendix S1-2] A method performed by a computer comprising a processor and memory, The program stored in the aforementioned memory, The S input step involves entering the patient's subjective information (S column), A template generation step that analyzes the subjective information of the patient obtained by the S input step and automatically generates a list of examination items or findings items corresponding to objective findings (column O) based on the analysis results, A display step in which, among the observation items generated in the template generation step, items that do not match the subjective information are skipped or hidden, and only items that are determined to match are displayed; The system includes a user interface step that allows a physician to select or input necessary findings and test items via the display unit shown in the aforementioned display step. A method characterized by having the step of transferring the input content to column O of the electronic medical record. [Note S2-2] In the method described in [Appendix S1-2], The template generation step is configured to dynamically add multiple examination items and physical findings items according to the pathological condition estimated from the subjective information (column S) of the patient, using one or more of the following: a large-scale language model or machine learning model, a rule-based system, or a search engine. The objective findings (column O) can then be supplemented by the physician selecting or inputting these additional items. A method characterized by having a step of automatically skipping input if it is determined that the additional item is unnecessary. [Appendix S1-3] A program for operating a computer that includes a processor and memory, The processor, in the program, The S input function allows you to enter subjective information about the patient (S column), A template generation function that analyzes the subjective information of the patient obtained by the aforementioned S input function and automatically generates a list of examination items or findings items corresponding to objective findings (column O) based on the analysis results, A display function that skips or hides items from the observation items generated by the template generation function that do not match the subjective information, and displays only the items that are determined to match. The system includes a user interface function that allows a physician to select or input necessary findings and test items via the display unit shown by the aforementioned display function. A program characterized by implementing a function to transfer the aforementioned input content to column O of the electronic medical record. In the program described in [Appendix S1-3], The template generation function is configured to dynamically add multiple examination items and physical findings items according to the pathological condition estimated from the subjective information (column S) of the patient, using one or more of the following: a large-scale language model or machine learning model, a rule-based system, or search. The physician can then select or input these additional items to supplement the objective findings (column O). Furthermore, the program is characterized by its ability to automatically skip input if it is determined that the additional item is unnecessary. [Explanation of Symbols]

[0090] 1. In-hospital telephone terminal (communication interface 12, input device 13, output device 14, memory 15, storage unit 16, processor 19) 2. Medical interview system (communication IF22, input / output IF23, memory 25, storage 26, processor 29) 3. Patient terminal 4. Medical systems (appointment systems, electronic medical record systems, electronic medical questionnaire systems, etc.)

Claims

1. A conversion unit that collects the patient's speech content and converts it into text data, A selection unit extracts one or more symptom information or disease information from the aforementioned text data, and selects a questionnaire template for each disease symptom based on the extracted symptom information or disease information. An input unit that inputs the aforementioned text data into the aforementioned disease symptom-specific questionnaire template, A questioning unit that selects or generates additional questions based on the remaining unentered content of the aforementioned disease symptom-specific questionnaire templates, The system includes a collection unit that presents the patient with the aforementioned additional questions and collects additional speech content. The input unit converts additional speech content into text using the conversion unit to create additional text data, and then inputs data into the disease symptom-specific questionnaire template based on the additional text data. A medical interview system characterized by the following features.

2. A conversion unit that collects the patient's speech content and converts it into text data, A selection unit for selecting a medical department-specific questionnaire template from one or more of the following: the aforementioned text data, the information from the reservation system, or the information from the electronic medical record. An input unit that inputs the aforementioned text data into the aforementioned medical department-specific questionnaire template, A questioning unit that selects or generates additional questions based on the remaining unentered content in the aforementioned medical department-specific questionnaire templates, The system includes a collection unit that presents the patient with the aforementioned additional questions and collects additional speech content. The input unit converts additional speech content into text using the conversion unit to create additional text data, and then inputs data into the medical department-specific questionnaire template based on the additional text data. A medical interview system characterized by the following features.

3. A conversion unit that recognizes the patient's speech and converts it into text data, A determination unit that determines whether a patient is a first-time patient or a follow-up patient based on either the text data or the patient's electronic medical record information, A selection unit extracts one or more symptom information or disease information from the aforementioned text data, and selects a medical interview template based on the extracted symptom information or disease information. An input unit that inputs the aforementioned text data into the aforementioned medical questionnaire template, A questioning unit that selects or generates additional questions based on the remaining unentered content in the aforementioned medical questionnaire template, The system includes a collection unit that presents the patient with the aforementioned additional questions and collects additional speech content. The input unit converts additional speech content into text using the conversion unit to create additional text data, and inputs this information into the medical questionnaire template based on the information obtained by the conversion unit and the determination unit. A medical interview system characterized by the following features.

4. A registration unit that obtains the patient's date of birth and telephone number in advance, A dialing unit that automatically dials the telephone number based on the telephone number obtained by the registration unit, The authentication unit, which verbally asks the patient connected to the patient terminal by the aforementioned transmitting unit for their date of birth or name, or patient ID or part of patient ID, transcribes the spoken content into text using speech recognition, and verifies the patient's identity by comparing it with pre-registered information, A medical interview system with two-factor authentication according to any one of claims 1 to 3, characterized in that the medical interview is initiated when the identity of the person is verified by the authentication unit.

5. The medical interview system according to any one of claims 1 to 3, characterized in that the input unit determines whether the input of the medical interview template has been completed and confirms the progress of the medical interview.

6. It has an evaluation unit that evaluates priority based on the content entered in the aforementioned medical questionnaire template, The medical interview system according to any one of claims 1 to 3, characterized in that the questioning unit selects or generates additional questions based on the priority evaluated by the evaluation unit.

7. The aforementioned questionnaire section is characterized by using a questionnaire template that includes questions regarding the onset, alliative and provoke, quality and quantity, region, severity symptoms, and time course when the patient's chief complaint is pain.

Images