JP-2026075118-A - Medical interview session and interview method

JP2026075118AJP 2026075118 AJP2026075118 AJP 2026075118AJP-2026075118-A

Abstract

[Problem] A medical interview tool (system, program, method) that automatically transcribes patient speech into text and automatically collects and inputs interview information. This will solve the problem of missing information that occurs in conventional medical interview systems and enable more efficient interviews. [Solution] The medical interview system 2 automatically converts the patient's speech 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 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. [Selection Diagram] Figure 1

Inventors

佐藤寿彦

Assignees

株式会社プレシジョン

Dates

Publication Date: 20260507
Application Date: 20260303
Priority Date: 20241011

Claims (13)

On the computer, (1) A process of storing a questionnaire template, which is structured data that associates items with content, (2) A process of collecting the patient's speech content and converting it into text data, (3) The process of analyzing the text data to extract one or more symptom information or disease information, and selecting a medical interview template based on the extracted results, (4) The step of inputting the information extracted from the text data into the corresponding field of the medical questionnaire template, (5) A step of determining which items are not entered based on the input status in the medical questionnaire template, (6) Based on the above determination, the steps include excluding already entered items from the questions and selecting or generating additional questions for the unentered items, (7) The process of presenting the patient with the additional questions and collecting additional speech content, (8) The process of converting the additional spoken content into text and inputting it into the medical questionnaire template, Make it run, A medical interview program characterized by repeating steps (5) to (8) above until there are no more unentered items.
In the medical interview program described in claim 1, The interview program is characterized in that the step of selecting the interview template includes the step of estimating the medical department based on one or more of either the information from the reservation system or the information from the electronic medical record, in addition to the text data, and selecting the corresponding medical department-specific interview template.
In the medical interview program according to claim 1 or 2, Further steps are taken to determine whether it is an initial or follow-up visit. The step of selecting the aforementioned medical interview template is characterized by selecting a corresponding medical interview template according to the judgment result.
In the medical interview program according to claim 1 or 2, The aforementioned input process is characterized by including a step of classifying information obtained based on the patient's free speech according to the value type of each item in the medical interview template and automatically inputting it into the corresponding field.
In the medical interview program according to claim 1 or 2, The aforementioned input step is characterized by including a step of checking the progress of the medical interview based on the input status of the medical interview template.
In the medical interview program according to claim 1 or 2, A medical interview program characterized in that the step of selecting or generating the aforementioned additional questions includes a step of determining the order of questions based on the urgency or priority of the unentered items.
In the medical interview program according to claim 1 or 2, The medical interview program is characterized in that the step of selecting or generating the aforementioned additional questions includes, for any unentered items, a step of generating additional questions using at least one of a generative model, machine learning, or rule-based method, in addition to the question items set in the medical interview template.
In the medical interview program according to claim 1 or 2, A medical interview program characterized in that the step of excluding previously identified items from the questions is performed based on conversation history.
In the medical interview program according to claim 1 or 2, When the patient's chief complaint is pain, the aforementioned interview template is a medical interview program characterized by including questions regarding the mechanism of onset, provocation and remission, nature and intensity, location, accompanying symptoms, and timeline.
In the medical interview program according to claim 1 or 2, The system obtains the patient's phone number and automatically calls back that number. The date of birth or patient identification information of the connected patient is obtained and verified against the registered information. A medical interview program characterized by further executing a step to start a medical interview once the aforementioned identity verification is completed.
In the medical interview program according to claim 1 or 2, By analyzing the patient's subjective information, a list of objective findings and test items is automatically generated. A medical interview program characterized by further performing a step of skipping or hiding findings and test items that do not match the aforementioned subjective information.
In the medical interview program according to claim 1 or 2, A medical interview program characterized by further executing a process of automatically reflecting or transferring the contents entered into the aforementioned medical interview template to a designated SOAP field in the electronic medical record.
A method of conducting a medical interview, which is performed by a computer using the medical interview program described in claim 1 or 2.

Description

This invention relates to a medical interview tool (medical interview system, program, method) that automatically transcribes patient speech into text, automatically inputs medical interview information, and collects any missing medical interview information using AI. It is well known that when visiting a medical facility, patients fill out a questionnaire in advance, detailing their medical history and reasons for visit, and that medical professionals such as nurses and doctors provide medical care based on the information provided. In recent years, there has been a technology called electronic questionnaires, which allows patients to input their questionnaire responses via information processing devices such as tablets, and then import that information into electronic medical records. The technology described in Patent Document 1 receives user responses to medical questionnaire data from the user's terminal and then notifies the user's terminal of further medical questionnaire data to confirm the patient's condition. It also anticipates the use of landlines and other telephones when users call medical institutions directly, using an automated voice response system to present the questionnaire content and allowing users to input their answers via push-button responses. However, the technology in Patent Document 1 relies primarily on user push-button operations and terminal input, and does not involve dynamic dialogue progression or an automated data collection and generation process based on the patient's speech. Specifically, because the automated voice response system conducts the questionnaire via a landline or similar device, and the user responds via push-button operations, the flow of the conversation is fixed, and dynamic dialogue that adapts to the user's speech content and the situation is not possible. This makes it difficult to provide appropriate support when the patient's condition is complex. 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 can select an appropriate questionnaire template based on that information and dynamically generate and present additional questions. In addition, it can detect any unentered information 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. Japanese Patent Publication No. 2022-070155 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 occurs particularly frequently 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.). In other words, while traditional medical interview systems relied on person-to-person communication—healthcare professionals directly questioning patients and assisting them in filling out questionnaires—electronic questionnaires, while offering increased efficiency, have limitations. They lack automated systems to handle incomplete entries. If there are missing entries, nurses or doctors must review the information, ask questions directly as needed, and supplement the information—a system heavily reliant on human intervention. 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 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. Therefore, if accuracy is paramount, it is better to express questions in simple language and conduct the interview in a question-and-answer format. It is preferable to use concise, short audi