Personalized AI agent-based healthcare check system and its implementation method

Abstract

We provide a system that uses AI specifically designed for patients to help maintain and improve their health. [Solution] The present invention is a healthcare check system using a personal AI agent implemented on a computer, comprising: (a) means for outputting text or voice conversations by a chatbot and recording the contents thereof; (b) means for acquiring the user's health management information; and (c) AI processing means in which the AI ​​agent analyzes the user's health condition and provides advice on the health condition based on the information obtained by (a) and (b), wherein the AI ​​agent is a generative AI that has learned specialized knowledge in the medical and pharmaceutical fields using a small language model, and the AI ​​agent periodically connects to an external server to learn external specialized knowledge in the medical and pharmaceutical fields.

Description

TECHNICAL FIELD

[0001] The present invention relates to a personal healthcare check system, and more particularly, to a system and method that use a small-scale language model to check, record, and manage a user's health condition using user-specific artificial intelligence (AI). BACKGROUND ART

[0002] When people feel physical abnormalities such as a cold or bodily pain, they first try to recover on their own by enduring the discomfort or resting to wait for natural healing. If natural healing does not occur, or if the physical abnormality persists or worsens, they go to a pharmacy to find and purchase appropriate medicine. If they cannot find suitable medicine on their own, they consult a pharmacist at the pharmacy to get guidance on appropriate medicine, then purchase and take that medicine. If the symptoms do not improve even after taking the medicine, they visit a local clinic to get a diagnosis, and obtain a prescription for medicine that helps alleviate the symptoms based on the diagnosis. The patient takes the prescription to a nearby pharmacy to receive the medicine specified on the prescription. In addition, in the case of chronic diseases such as diabetes and hypertension, people visit a doctor for a long period of time to get similar medicines issued via prescription, then go to a nearby pharmacy to obtain the medicines listed on the prescription. At such times, the pharmacist explains the efficacy, administration method, and precautions of the medicine to the patient. The patient takes the received medicine and waits for the symptoms to improve.

[0003] In recent years, the systems of "family pharmacies" and "family pharmacists" have emerged. A family pharmacy is a pharmacy that patients can visit regularly. By using the same pharmacy, it becomes easier for pharmacists to manage medications and patients, allowing for appropriate medication guidance and consultations. This fosters a trusting relationship between patients and pharmacists, and regular visits also provide pharmacies with a stable customer base. Pharmacists can better understand patients' health conditions and the effects of their medications, enabling them to provide appropriate advice and support, which greatly benefits patients. Furthermore, family pharmacies manage patients' medication history and allergy information, increasing safety by preventing drug side effects and interactions. Thus, a family pharmacist is a pharmacist with knowledge and experience in drug treatment, health, and care, who can provide consultations tailored to the needs of patients and their communities. They can be described as a patient's personal pharmacist. Because one specific pharmacist is exclusively responsible for a particular patient's medication status, they can continuously monitor for drug duplication, drug interactions, the effectiveness of medications, and the presence of side effects, allowing patients to take their medications with peace of mind. Furthermore, in fields such as healthcare and pharmaceuticals, patient health management and medication guidance are conducted face-to-face or through individual systems. However, adequate solutions remain in place regarding ensuring individual privacy, rapid access, data preservation during disasters, and mechanisms for integrating diverse information (prescriptions, vital signs, complexion, voice information, etc.). [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2018-060526

[0005] [Patent Document 2] Japanese Patent Publication No. 2020-181357 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] When a patient visits a hospital and receives a prescription, they go to a pharmacy where a pharmacist hands them the prescribed medication. At that time, the pharmacist provides a medication information sheet containing information about the effects of the medication and precautions for taking it. However, since pharmacists do not remember the details of many medications, they usually just read the information from the sheet or hand it over with little explanation. Especially when there are many patients visiting the pharmacy, they need to process patients within a limited time, so sometimes they simply hand over the medication along with the information sheet. Furthermore, even if patients want to know more about the contents of their medication and how to take it, they often find it difficult to ask pharmacists because they feel intimidated.

[0007] Furthermore, it takes time for pharmacists to understand patients' health management, and it's difficult to easily ascertain a patient's condition. Moreover, pharmacists rarely try to understand patients' health management because they consider it disrespectful to intrude on patient privacy. Even if a pharmacist does obtain health information from a patient, it takes time and money to store that information, so it's rarely accumulated by the pharmacist or pharmacy. Consequently, for example, when a pharmacist talks to a patient, they often use standard phrases and engage in generic conversations with people taking similar medications. While a primary care pharmacist is supposed to manage the health of specific patients and provide detailed guidance on medications, they face the same time and physical challenges as general pharmacists when dealing with a large number of patients, resulting in underutilization. Furthermore, while there are ideas such as a system (Patent Document 1) in which pharmacists use tablet PCs to provide medication guidance to patients while viewing the screen together, and where the content is automatically recorded as a medication history, and an idea (Patent Document 2) to eliminate patient waiting times and prevent dispensing errors by directly linking digitized prescription data with dispensing preparation robots, no method has yet been proposed for the system itself to collect information and improve patient health. [Means for solving the problem]

[0008] This invention relates to a personalized AI (artificial intelligence) healthcare check system that is well-versed in health management and drug information. It provides a method for dramatically improving the quality of health management and drug information by having a dedicated AI agent for the patient or pharmacist (referred to as the user) provide advice to the user. Specifically, it has the following features: (1) The present invention relates to a healthcare check system using a personal AI agent implemented on a computer, comprising: (a) means for outputting text or voice conversations via a chatbot and recording the contents thereof; (b) means for acquiring user health management information, and (c) AI processing means for an AI agent to analyze the user's health condition and provide advice on the health condition based on the information obtained in (a) and (b), wherein the AI ​​agent is a generative AI that has learned specialized knowledge in the medical and pharmaceutical fields using a small language model, and the AI ​​agent periodically connects to an external server to learn external specialized knowledge in the medical and pharmaceutical fields.

[0009] (2) In addition to (1), the present invention is characterized in that the healthcare check system having the personal AI agent is connected to a healthcare check system having an AI agent dedicated to pharmacists or pharmacies (collectively referred to as pharmacists, etc.), enabling information exchange between the personal and / or personal AI agent and the AI ​​agent dedicated to pharmacists, etc., the computer is a smartphone, personal computer, or small server, the health management information is data such as prescriptions, complexion, voice, and / or vital signs (blood pressure, pulse, body temperature, etc.), and the invention provides a user interface that allows patients, pharmacists, chatbots, and patients' families to participate on the same screen and converse with each other.

[0010] (3) In addition to (1) or (2), the present invention further comprises a facial recognition module that enables access to the healthcare check system with one click by facial recognition of the patient or pharmacist, an image recognition means that uses the camera function of the computer to photograph paper media containing examination data provided by a medical institution and converts the captured image into digital data by image recognition, a data management means that forms storage links with multiple data servers to realize data backup and data preservation in the event of a disaster, and a means that analyzes the patient's gaze, facial expression and voice response, and uses this means to check the patient's medication status and health status and record the results, and further comprises a voice recognition text conversion function (means) that converts voice during medication guidance into text and means that saves the converted data, and a means that reads medication information displayed on a medication check calendar with a camera on the computer and confirms the medication status.

[0011] (4) In addition to (1), (2), or (3), the present invention has at least two display screens, one of which displays the user's own information and images captured by a camera, and the other display screen can display a pharmacist, etc. (meaning a pharmacist and / or a pharmacy), or an AI agent dedicated to the pharmacist, etc., and the healthcare check system having the personal AI agent is equipped with a medicine box for storing the patient's medication, is capable of managing the patient's medication status by capturing images of the patient taking the medication with the camera, has a function to record the date and time the patient takes the medication and notify the patient when that date and time arrives, or has a function to warn the patient if the medication is not taken at that time, and further has a lid on the medicine box, and has a mechanism that automatically opens the lid when the patient takes the medication and automatically removes the stored medication from the outside of the medicine box. (5) In addition to (1), (2), (3), or (4), the present invention is characterized in that when an abnormality occurs in the patient's body, the AI ​​agent observes the affected area (symptom) of the body with the camera function of the computer, advises the patient on whether the abnormal change is due to a side effect of the drug, and displays the method of taking the drug on the display screen of the computer, thereby allowing the patient to take the drug.

[0012] (6) In addition to (1) to (5), the present invention is characterized in that the AI ​​agent reads health checkup result forms and / or physical examination result forms and / or prescriptions by OCR processing, and records digitized health and / or physical and / or medication data in a health management (data) table in chronological order; the AI ​​agent takes in electronic prescription data or obtains medication data via API (My Number Portal and / or pharmacy), and records the medication data in chronological order on the health management (data) table sheet; and the AI ​​agent reads a health management (data) table in which health data (health data) and physical examination results (physical examination results) are listed in chronological order, and if there is an abnormality in the health data or physical examination results, displays a warning on the health management (data) table and / or communicates the patient's abnormality to the patient's family and / or the patient's pharmacist. (7) In addition to the configurations of (1) to (6), the present invention has the following features: The AI ​​agent informs the patient of the timing of medication based on the patient's medication schedule recorded in the health management (data) sheet. When the patient takes their medication, the AI ​​agent takes a picture of them with a camera and records that the medication has been taken in the health management (data) sheet. It is also possible to notify the patient's family or pharmacist of the patient's medication information. Furthermore, if vital data such as blood pressure, pulse rate, and blood glucose level are obtained from a smartwatch or the like, the AI ​​agent obtains this audio or digital data directly or via an API. The obtained data can then be recorded in the health management (data) sheet with the date attached through speech recognition processing or natural language processing. In addition, if a prescription is provided on paper, the AI ​​agent can instruct an external AI capable of OCR processing or image recognition processing to digitize the contents of the prescription and receive that data. Alternatively, the AI ​​agent can be equipped with OCR processing or image recognition processing functions and digitize the prescription itself and enter it into the health management (data) sheet.

[0013] (8) In addition to (1) to (7), the present invention is characterized in that, in the case of anomaly detection of data in the health management (data) table, the AI ​​agent detects specific conditions (e.g., exceeding a standard value) from the health management (data) table, converts them into natural language, and transmits them to the AI ​​agent; and in the case of anomaly detection of data in the health management (data) table, the AI ​​agent performs natural language processing on the data, sends the processing results to an AI specialized for health checkups (AI for health checkups) for analysis, receives the analysis results, enters the data into the health management (data) table along with the date, and transmits it to the patient, family, and pharmacist; and in the case of a diagnosis such as CT, MRI, or ultrasound, the processing results of the image medical data are sent to an AI specialized for medical image diagnosis (AI for medical image diagnosis) for analysis, the AI ​​agent receives the analysis results, enters them into the health management (data) table along with the date, and transmits them to the patient, family, and pharmacist. (9) In addition to (1) to (8), the present invention is characterized in that, when an AI agent receives a diagnosis from a doctor, the AI ​​agent has the doctor's voice converted into data by a speech recognition processing AI, receives that data and records it chronologically on a health management (data) sheet, and when an AI agent with a speech recognition processing function receives a diagnosis from a doctor, the AI ​​agent converts the doctor's voice into data and records it chronologically on a health management (data) sheet, and the AI ​​agent communicates with an AI specialized in that field (for example, a medical AI) regarding the doctor's diagnosis results for the patient, clarifies any unclear points and records them chronologically on a health management (data) sheet. (10) In addition to (1) to (9), the present invention is characterized in that the AI ​​agent has a voice dialogue function, a camera image analysis function, a data storage function, an automatic notification function, and / or a robot cooperation function, and uses the above functions to prevent the patient from forgetting to take their medication, and the AI ​​agent uses the voice dialogue function to prevent the patient from forgetting to take their medication through conversation with the patient, and further uses the camera image analysis function to take a picture of the patient taking their medication and determines from the image whether the patient has taken the medication, and the AI ​​agent saves the date and time of medication taking in the health management (data) table based on the determination result, and further notifies the patient's family and / or pharmacist of the medication record, and when the AI ​​agent receives a confirmation question from the patient about whether or not they have taken their medication, it looks up past medication data in the health management (data) table and answers the patient about whether or not they have taken their medication.

[0014] (11) In addition to (1) to (10), the present invention provides a healthcare check system having an AI agent dedicated to an individual (patient), wherein the system is equipped with a personal mobile information terminal (hereinafter referred to as a smartphone), the AI ​​agent dedicated to the patient is installed on the smartphone, and the system has the function of using the patient's AI agent installed on the smartphone to grasp the patient's health status and notify the patient if there is an abnormality, and further provides a medicine box (medicine box) for storing the patient's medication, the AI ​​agent dedicated to the patient is installed on the medicine box, and the system has the function of using the patient's AI agent installed on the medicine box to grasp the patient's health status and notify the patient if there is an abnormality, and further provides The system is equipped with a personal mobile information terminal (hereinafter referred to as a smartphone), and the patient-specific AI agent is installed on the smartphone. The system has the function of using the patient's AI agent installed on the smartphone to understand the patient's health status and notify the patient if there is an abnormality. The patient's AI agent installed on the smartphone and the patient's AI agent installed on the medicine box can communicate with each other and have the function of sharing the patient's health information held by both AI agents. Furthermore, the patient's AI agent installed on the smartphone and / or the patient's AI agent installed on the medicine box can converse with the patient and have the function of understanding the patient's cognitive function through conversation.

[0015] (12) In addition to (1) to (11), the present invention provides a healthcare check system having an AI agent dedicated to an individual (patient), which includes a medicine box for storing the patient's medication, the AI ​​agent dedicated to the patient is installed in the medicine box, and the system has the function of using the patient's AI agent installed in the medicine box to grasp the patient's health status and notify the patient if there is an abnormality, the patient's AI agent installed in the medicine box has the function of using a camera function to monitor and analyze the movements of patients around the medicine box to grasp the patient's abnormal state, and the patient's AI agent installed in the medicine box has the function of using a microphone function to monitor and analyze the voice of patients around the medicine box to grasp the patient's abnormal state. Furthermore, in remote treatment or remote clinical trials, the patient's AI agent installed in the medicine box has the function of managing the patient's medication use and health status and transmitting the medication data and health status to the remote treatment or remote clinical trial institution for communication. [Effects of the Invention]

[0016] According to this invention, users can quickly and safely access an AI agent anytime, anywhere, such as at home or while out, and receive personalized healthcare checks. Furthermore, by integrating various technologies such as facial recognition, voice and text output, image recognition, and gaze and facial expression analysis, it becomes possible to accurately grasp medication status and health condition, enabling health management and medication guidance with higher accuracy than before while protecting privacy. The AI ​​agent possesses a wealth of medical and pharmaceutical information, so it can provide patients with appropriate information regarding illnesses and necessary medications. In addition, the AI ​​agent can monitor the patient's daily health condition and medication status, making a significant contribution to maintaining and improving the patient's health, and also helping to prevent missed medication doses. Moreover, this invention aims to address challenges such as the strain on medical care in an aging society and the excessive workload of pharmacists by having the AI ​​agent collect, process, and advise on information, thereby reducing the burden on pharmacists and improving the quantity and quality of patient care. Brief Description of the Drawings

[0017] [Figure 1] Figure 1 is a diagram showing the configuration of a healthcare check system using an AI agent of the present invention. [Figure 2] Figure 2 is a block diagram illustrating the main functions of the present invention. [Figure 3] Figure 3 is a diagram showing a medicine box computer with an AI agent using the healthcare check system of the present invention. [Figure 4] Figure 4 is a diagram showing an example of a distributed AI platform. [Figure 5] Figure 5 is a flowchart of patient medication monitoring by an AI agent for preventing a patient from forgetting to take medicine. Modes for Carrying Out the Invention

[0018] The present invention is a healthcare check system in which a virtual specialized advisor (referred to as an AI agent) specialized in medical information, pharmaceutical information, etc. appearing on a computer provides professional advice such as medical information and pharmaceutical information to users such as patients and pharmacists, so as to maintain and improve the health of the users. Further, the present invention relates to a personal healthcare check system, and more particularly to a system and method that uses a small-scale language model to allow patients, pharmacists, chatbots, family members, etc. to exchange information with each other on the same screen, and utilizes face authentication, voice / text output, image recognition, etc. to check, record and manage the user's health condition. With the recent development of small-scale language models, advanced AI processing has become possible on smartphones, personal computers, tablets and small servers, so that the healthcare check system using a personal-specialized AI agent as in the present invention can be realized. Figure 1 is a diagram showing the configuration of a healthcare check system using the AI agent of the present invention. The AI agent is an AI (Artificial Intelligence) having capabilities specialized in professional fields such as medical care and pharmaceutical affairs, and is in particular a virtual, so-to-speak, advisor that provides professional information advice to users (especially patients and pharmacists) regarding pharmaceutical affairs. The AI agent is generative AI that has learned specialized knowledge in the medical field, pharmaceutical affairs field, welfare field, and / or nursing care field based on a Small Language Model (SLM), and thus can provide highly accurate responses and handling in the specialized fields. Note that the SLM is a relatively small-scale language model composed of hundreds of millions to billions of parameters, is optimized for specific fields (in the present invention, the medical field, pharmaceutical affairs field, welfare field, and / or nursing care field) and their applications, exhibits high performance in these fields and applications, and can provide users with abundant, accurate, safe, and highly reliable information and advice. In Figure 1, the AI agent 10 is implemented in and operates on a computer 11, and before being implemented, it is generative AI that learns accurate information related to specialized fields such as the medical field 23, pharmaceutical affairs field 24, welfare field 25, and / or nursing care field 26 from Big Data (a huge group of data that has existed (or exists) from the past to the present) 22, acquires advanced knowledge of these specialized fields, and then is implemented in the computer 11. Even after being implemented in the computer 11, the AI agent 10 can also learn new information related to the above fields by connecting to an external (data) server 21 as needed or at all times, and further acquires accurate information from academic conferences, industry, and papers that have been supervised and checked by experts in the above fields. Inaccurate information and data in the above fields are not used or are excluded by the AI agent 10 of the present invention. In addition, a large number of AI agents 10 (each AI agent is implemented in its respective computer) also constantly exchange information via the external (data) server 21 to complement each other's lacking knowledge, information, and data, and can improve together.

[0019] Computer 11 is a device that has an environment in which an AI agent can be implemented and operated, such as a smartphone, various personal computers (PCs: notebook PCs, desktop PCs, etc.), tablets, and small servers. The most mobile is a smartphone that can be carried at all times. Since each individual's dedicated AI agent needs to constantly acquire new and accurate information, it connects to an external (data) server 21 when needed (for example, once a day) to obtain and learn fresh and accurate information from a main server that constantly acquires and updates a vast amount of accurate information and data in a specific field (medical, pharmaceutical, welfare, and / or nursing care). Computer 11 has a display screen 19 that can display the AI ​​agent 10. Computer 11 also has a camera (function) 12 that can take pictures of the surroundings. Computer 11 also has a microphone (function) 13 that can pick up sounds in the surroundings. Therefore, the AI ​​agent can hear the voices of patients 16 and pharmacists 17 through the microphone 13 and interpret and understand the content of those voices. Furthermore, the computer 11 has a speaker (function) 14, and the AI ​​agent can output its own interpretations to the outside as words through the speaker (function) 14. In other words, it can communicate the AI's interpretations to the user by voice. Therefore, the AI ​​agent can also converse with the user, such as the patient 16 or pharmacist 17, using the microphone (function) 13 and the speaker (function) 14. The computer 11 also has an authentication device (function) 15. The authentication device is, for example, a fingerprint reader, and reads the user's fingerprint to access the healthcare check system equipped with the AI ​​agent of the present invention and determines whether the user is the user. If the fingerprint is that of the user, the healthcare check system equipped with the AI ​​agent of the present invention is activated, the AI ​​agent 10 appears on the display screen 19 of the computer 11, and then the user can converse with the AI ​​agent 10.If the authentication device is the camera 12, which captures the user's face and performs facial recognition (in this case, the authentication device is the camera), clicking the authentication display (icon, etc.) on the display screen 19 with a mouse or the like will capture the user's face with the camera 12, the computer will perform facial recognition, and the healthcare check system will start. In other words, access to the healthcare check system is possible with a single click. Alternatively, if the authentication is performed by obtaining the user's voice using the microphone (function) 13, the authentication device will be the microphone (function) 13.

[0020] Computer 11 has a keyboard and a display screen, so it can engage in conversations (mutual communication) via chatbot. It also has a recording mechanism, so it can record the content of conversations. Therefore, people who have difficulty speaking can use this system. Alternatively, it can also have a braille input / output function (for example, input is done via keyboard and a paper written in braille is output), so visually impaired people can also use this system. A smartphone is owned by only one person and used only by that owner, so the AI ​​agent installed on the smartphone is unique to that person and only one (for one person) is needed. However, in the case of a PC, personal server, or tablet, multiple people may use it, so multiple systems or AI agents can be installed. However, since personal health information is personal information, it is not permitted for others to know it in principle, so the above authentication method is important. In other words, when multiple systems (AI agents) are installed on one computer, the healthcare check system that each user can authenticate and access is dedicated to that person, and access to another person's system is impossible. Even with smartphones, there is a risk of loss or theft, so authentication is naturally necessary. Since computer 11 has a camera (function) 12, it can recognize not only the user, patient 16, but also the pharmacist 17. If patient 16 gives permission (for example, by clicking the permission icon on the computer 11's display screen), pharmacist 17 can access the user's (patient's) healthcare check system and obtain the user's (patient's) health management information from the AI ​​agent. Based on this health management information, the pharmacist can provide the user (patient) with more detailed information and health management guidance. Patients, pharmacists, patients' families, and even chatbots can appear (participate) on the computer 11's display screen 19 via the computer 11's camera (function) 12, and these participants can converse with each other using the microphone (function) 13 and speaker (function) 14. The interface here is the camera (function) 12, microphone (function) 13, and / or speaker (function) 14. The AI ​​agent 10 can interpret the content of these participants' conversations and provide appropriate responses and advice. Patients may not understand what pharmacists say, or pharmacists may not understand what patients say. However, each AI agent can translate the content so that it can be understood. Furthermore, a patient-specific AI agent can judge whether the pharmacist's advice is appropriate. In addition, each AI agent can intervene in communication between pharmacists and patients, supporting and facilitating effective communication.

[0021] When a patient goes to the hospital and receives a diagnosis, they are given paper prescriptions and test data (paper media). This system uses the camera function of a computer to photograph the paper media containing the test data provided by the medical institution, and uses the image recognition function (means) of the computer to perform image recognition on the captured image and convert various images into digital data. For example, letters and numbers written on the paper media can be converted into digital data, and an AI agent can analyze this data to provide various health management advice. The system of the present invention can use the camera (function) to observe, read, and analyze the patient's gaze, complexion, and facial expressions, and / or use the microphone (function) and speaker (function) to acquire and analyze words, voice, their volume and intonation, etc., to check, record, and manage the patient's health status. It can also check and record the patient's medication status (for example, reading and checking conversations about medication, medication (check) calendars, or drug information displayed therein using the camera, etc.), so the AI ​​agent can check and manage medication status, and based on the analysis results, can accurately and appropriately suggest methods for maintaining health and improving health status to the patient. Based on the patient's voice and images, or conversations between the patient and pharmacist, or between the patient and the AI ​​agent, if there is an abnormality in the patient's condition, the AI ​​agent can immediately issue an alarm and warn the user (patient, pharmacist). In this case, the AI ​​agent can, of course, be dedicated to the patient, but even if it is dedicated to the pharmacist, it can observe and analyze the patient's condition, or analyze their voice, to inform the pharmacist about the patient's condition.

[0022] Traditionally, pharmacists provide medication guidance to patients through direct conversation. However, this system has a speech-to-text conversion function (means) to convert spoken medication guidance into text, and a function (means) to save the converted data. Therefore, the AI ​​agent can also convey the pharmacist's medication guidance to the patient in text or printed on paper. Consequently, patients can review it later on their own computers or on paper. Pharmacists hand out medications to patients based on prescriptions at pharmacies, but with many patients, time is limited, and explanations of the medications tend to be neglected. Patients also want to receive the medications listed on their prescriptions as quickly as possible, and therefore want to reduce waiting times by simply receiving their medication. This system can solve both these needs of patients and pharmacists. Pharmacists themselves have their own AI agents, which store all the drug information in the pharmacy. The AI ​​agent reads prescription information on its computer and informs the pharmacist where and how much of each drug is available, as well as the dispensing method. It can also provide medication information and record it on paper if necessary. This allows pharmacists to dispense medication to patients in the shortest possible time. The pharmacist's AI agent also organizes the information to be conveyed to the patient verbally, so the pharmacist only needs to explain it directly to the patient, saving time. Furthermore, even with refill prescriptions, the patient-specific AI agent of this invention can handle daily health management, medication management, and medication guidance, allowing doctors and pharmacists to respond with confidence.

[0023] If patients also have an AI agent, it will concisely organize the information they want to know, allowing them to ask the pharmacist accordingly. This system has a method (means) to connect the pharmacist's AI agent with the patient's AI agent, so problems can be resolved in a very short time by communicating between these AI agents, and the patient only needs to receive their medication. Necessary information can be obtained from one's own AI agent when needed (when the patient has time). For example, medication information (effects of the drug, precautions for taking it, etc.) can be directly transmitted from the pharmacist's AI agent to the patient's AI agent (this can be done instantly). This dramatically improves the productivity of pharmacists, and the extra time can be used to acquire their own expertise, thus improving the quality of pharmacists. Patients also benefit from no wasted waiting time and can use their time more effectively. If communication between patients and pharmacists is done via video call instead of face-to-face, the pressure of face-to-face interaction is eliminated, and more meaningful communication can be expected. Necessary medications can be sent by mail or other means.

[0024] If hospitals are permitted to dispense medication online (without in-person meetings) based on prescriptions, patients can contact the pharmacy online (via video call, etc.) and receive their medication. However, if the patient's AI agent is present (communication can be done through a computer equipped with the AI ​​agent), the AI ​​agent can provide the patient with the necessary information, allowing the patient to ask various questions to the pharmacist at the pharmacy and receive answers. Currently, there are electronic prescription systems, and the patient's AI agent receives this electronic prescription via communication, contacts the pharmacy, and receives the medication via delivery. Thus, the system of the present invention can also be used with electronic prescription systems, allowing patients to receive their desired medication without going to a pharmacy. If pharmacies also have their own dedicated AI agents, patients can receive their desired medication without going to a pharmacy, simply by going to the hospital for a consultation, through communication between the patient's AI agent and the dedicated AI agent at the pharmacy or pharmacist. The contents of the prescription can be checked by the patient's and / or the pharmacy / pharmacist's dedicated AI agent. This saves time and eliminates the need to expend extra effort when feeling unwell. Furthermore, if pharmacies also have their own AI agents dedicated to pharmacists, communication between AI agents becomes possible, saving even more time, effort, and costs. Medications can be obtained via courier service. Moreover, if it becomes standard practice for patients to receive their medication without going to a pharmacy using this system, prescription data can be sent from the hospital to the patient's system and to the pharmacist's (or pharmacy's) system. If the authentication of both systems matches, the AI ​​agents in both systems can communicate with each other, leading to reductions in time, effort, and costs for all three parties: the hospital, the pharmacy (pharmacist), and the patient.

[0025] When a patient receives a diagnosis at a hospital, diagnostic information such as medical records, diagnostic reports, and test results are created. If these are written on paper, the patient's AI agent can scan them with its camera or other means to digitize them and understand the patient's illness information. If these are electronic medical records, the patient's AI agent can directly receive this digital information and understand the patient's illness information. Based on this illness information, the patient's AI agent can observe the patient's daily life and provide advice. It can also contact pharmacists, AI pharmacist agents, and doctors as needed, or obtain general and specialized information to obtain and communicate necessary information for maintaining, improving, and enhancing the patient's health.

[0026] In the healthcare check system of the present invention, by preparing multiple external data servers 21, forming storage links between the multiple data servers, and connecting or coordinating the main computer 11 with the multiple data servers, the contents of the AI ​​agent (including data, etc.) can be constantly backed up, thereby ensuring data preservation even in emergencies (including disasters). Furthermore, by connecting to various external data via a network, the reliability and scalability of the overall system operation can be ensured. (These are referred to as emergency data backup and preservation management means.) This management means allows for early recovery, for example, in the event of a disaster, when the system is infected with a virus, or when data is lost.

[0027] Figure 2 is a block diagram illustrating the main functions of the present invention. Those who use this system are called users, and these users are patients and / or pharmacists. To use this system, a terminal (user-side terminal (computer)) such as a smartphone, tablet, or personal computer (PC) is required. The application software for the system of the present invention is installed on this terminal, and the user can operate it using the terminal. (Figure 2A) The terminal has a display screen and allows for various operations. To access the system of the present invention, the user calls the application software (also called an app) on the terminal and performs the necessary authentication. Authentication methods include physical authentication, voice authentication, and passwords. Physical authentication methods include facial recognition, fingerprint authentication, and biometric authentication, but considering dementia and other factors, facial recognition is simple and highly secure.

[0028] Upon authentication, the AI ​​agent, which forms the core of the system of this invention, appears on the terminal's display screen. The AI ​​agent utilizes a small-scale language model (SLM) to provide core functions such as health checks and advice to users, including patients and pharmacists. On the terminal's display screen, it can appear as a human figure, such as a pharmacist wearing a white coat, to make it more user-friendly. A small-scale language model is a relatively small language model (compared to a large-scale language model), consisting of hundreds of millions to billions of parameters, and is characterized by being optimized for specific fields and applications. This system possesses abundant and accurate knowledge and information on specialized fields such as medicine, pharmaceuticals, welfare, and nursing care, and is capable of specialized learning in these fields. (Figure 2B)

[0029] This system has various subsystems (also called modules) to enhance the functionality of the AI ​​agent. The facial recognition module can perform facial recognition of users using the camera and other devices on the terminal. This facial recognition function ensures that only authorized users can access the system. Furthermore, the system can analyze the user's health status using the camera, including their complexion, enabling health management. It can also grasp the patient's health status in real time from their facial expressions and complexion, and communicate the results to the individual, or, in some cases, to pharmacists or doctors. The data from the facial recognition module is sent to the terminal and the AI ​​agent. The AI ​​agent uses this data to communicate analysis data and results to the user, and if there are problems with the user's health, it provides improvement methods based on the AI ​​agent's knowledge and information. It can also read text using the camera and understand the content of documents, allowing it to read patient prescriptions and medical records and provide necessary information to the user. (Figure 2C)

[0030] The voice input / output module allows the AI ​​agent to input and output voice using the microphone and speaker built into the terminal. The AI ​​agent can understand what the user is saying and converse with them. It can also respond to text input with voice output or text output. Furthermore, it can output and input conversations in both text and voice, record voice data, and convert medication guidance into text. During medication guidance, the voice is converted into text and saved as simplified data on the patient's terminal and / or the pharmacist's or pharmacy's terminal or server. By checking the user's (patient's) cognitive function through their voice, the AI ​​agent can advise the user (patient, pharmacist, doctor, etc.) on ways to improve. (Figure 2D)

[0031] The image recognition / OCR (Optical Character Reader) module has the function of taking pictures of paper-based test data, prescriptions, etc., provided by medical institutions with a smartphone camera and converting them into digital data through image recognition. The AI ​​agent analyzes these results and suggests improvement methods to the user. For example, if there is data on hypertension, it can suggest methods to improve hypertension, such as guidance on reducing salt intake or antihypertensive medication. The AI ​​agent can also monitor the patient's lifestyle, predict the health risks of that lifestyle, and provide lifestyle guidance to the patient. (Figure 2E)

[0032] The conversation recording and management function records and saves the content of conversations that take place between chatbots, patients, pharmacists, family members, etc. The AI ​​agent can analyze this recorded and saved data to understand changes in the patient's health condition. Based on the analysis results, the AI ​​agent can also suggest ways to improve the patient's health condition to the user. Furthermore, it can participate in conversations between users, discuss them, and answer users' questions (Figure 2F). Records can also be made tamper-proof using blockchain. For example, multiple users who have a dedicated AI agent in this system can each create a blockchain and store their respective data, making tampering difficult. Since each user's own data can be protected by facial recognition, etc., there are no security issues with the data of others.

[0033] The user interface function allows multiple participants to connect on the same screen and communicate in real time. It uses cameras and other devices to acquire individual and / or group images of each participant, displaying multiple images on the terminal (computer) screen for image processing. It also uses microphones and / or speakers to send data to the terminal (computer) individually or group-wide, outputting and analyzing each person's voice. The AI ​​agent can use this function to converse with stakeholders (patients, pharmacists, doctors, family members, etc.), providing them with information on medical care, pharmaceuticals, and welfare, and gathering their opinions to incorporate into future health management. (Figure 2G)

[0034] The data management function enables the storage and backup of various recorded data, as well as data preservation through linkage with multiple external data servers. In particular, it accumulates long-term health information of users to monitor their health status. The AI ​​agent analyzes this data and, if there is a change in the user's health status, issues a warning to draw the user's attention. It can also suggest ways to improve the situation. Furthermore, it can acquire and manage vital signs, prescription data, test results, etc., in conjunction with various sensors (Figure 2, H). In this way, by combining subjective information from the patient with sensor information, etc., it is possible to visualize changes in the body due to forgetting to take medication, etc., and communicate this to the patient, their family, or pharmacist, enabling early detection of changes in health and appropriate action.

[0035] The external data server / healthcare institution data linkage function is an external resource used for system-wide data backup, linking to the latest medical and pharmaceutical data, and linking with test data provided by healthcare institutions. In the event of a malfunction in the user's terminal, the aforementioned functions, or the AI ​​agent, all activity records and data of the user's terminal, the aforementioned functions, and the AI ​​agent are stored on the external server. Therefore, if a malfunction occurs in the user's terminal, the aforementioned functions, or the AI ​​agent, the AI ​​agent can be restored by replacing the faulty terminal and connecting to the external server. Furthermore, the latest data related to medical care, pharmaceuticals, and welfare is constantly entered into the AI ​​agent through this function. The AI ​​agent also compares the latest data with older data and can select and communicate the best information to the user. (Figure 2, I)

[0036] As described above, the AI ​​agent is an AI (artificial intelligence) created in a virtual space such as a computer, possessing specialized knowledge in fields such as medicine, pharmaceuticals, welfare, and nursing care, and utilizing that expertise to provide advice to patients and pharmacists. Specifically, it collects, processes, and adjusts various information (medical history, information from the patient themselves, doctor's diagnosis, prescriptions, biometric information, health checkup results, information from the internet, third-party observation results, etc.) and provides advice to patients, caregivers, pharmacists, etc. This AI agent can also be dedicated to individual patients, pharmacists, and caregivers, enabling them to provide detailed advice. It can also provide information such as the timing of health checkups, the timing of examinations, frequency, and importance. The AI ​​agent not only presents information but also has the function of monitoring changes in the patient's condition (vital data (pulse, respiration, temperature, blood pressure, level of consciousness, etc.), voice biomarkers, etc.) and issuing alerts to pharmacists and family members. The final decision can be made by the pharmacist or patient themselves, but the AI ​​agent's support in providing information enables pharmacists and patients to make higher quality decisions.

[0037] AI agents can use their information and knowledge to directly communicate the degree of change in a patient's complexion, facial expressions, and voice, and can also suggest possible symptoms, problems, and ways to improve them while interacting with the patient. AI agents dedicated to pharmacists can also take a step-by-step approach, transmitting patient information to the pharmacist first, who then reviews and adjusts the information before relaying it to the patient. AI agents dedicated to patients can monitor the patient's daily condition using cameras and other devices, and can also interact directly with the patient, allowing them to prompt patients to take their medication as a measure to prevent missed doses. For example, the AI ​​agent can read a schedule sheet listing medication dates and times using a camera, and when the patient's medication time approaches, it can use its speaker, camera, and microphone (loopphone) functions to notify the patient. Once the AI ​​agent confirms that the patient has taken their medication using the camera, it can check off the item on the schedule sheet. Furthermore, by linking with a pharmacist AI agent, the patient's AI agent can notify the pharmacist's AI agent of their medication adherence, allowing the pharmacist to know if the patient is following their medication schedule and to incorporate this information into medication guidance during the next visit. Additionally, the AI ​​agent can be present during regular meetings between the patient, pharmacist, and patient's family, enabling a combination of human and AI agent support.

[0038] As a patient's illness progresses or worsens, they may find it difficult to move on their own. In such situations, the patient's AI agent can communicate with family members and caregivers. For example, the AI ​​agent can inform family members and caregivers of medication schedules, dosage instructions, and precautions, ensuring the patient takes their medication. By registering family members and caregivers with the AI ​​agent using facial recognition, the AI ​​agent can also transmit and share the patient's personal information with them. The AI ​​agent can also take on a human form (visually), demonstrating how to take medication to those unfamiliar with the process by showing them images of a human mouth. Patients can then mimic these images to successfully take their medication. Similarly, when using topical patches, the AI ​​agent can show images of a human body demonstrating how to apply them correctly to the affected area. Furthermore, when an injury occurs, the body swells, the skin discolors, or any other abnormality occurs, the AI ​​agent can examine the affected area (by taking a picture with the AI ​​agent's (computer's) camera), and then instruct the patient to use appropriate ointments, patches, or compresses, which they can then obtain at a pharmacy. In some cases, the AI ​​agent can even communicate with the pharmacy to arrange for the medication. Depending on the symptoms, it can also instruct the patient to see a doctor immediately and can refer them to nearby hospitals. The AI ​​agent is particularly useful for infants, the elderly, and those requiring care. Because the AI ​​agent is dedicated to the patient, it has a good understanding of the patient's physical condition and can warn against incorrect medication use, thus resolving the problem of medication overdose. If a patient has a fever, the AI ​​agent can measure their temperature with a camera (with thermal imaging capabilities) and instruct them on appropriate medication. Moreover, AI agents equipped with infrared and X-ray cameras can also perform bone image analysis and provide medication and hospital referrals.

[0039] Figure 3 shows a medicine box computer with an AI agent using the healthcare check system of the present invention. This device is the system shown in Figure 1, but with an additional medicine box. In the system shown in Figure 3, the entire system can be controlled by a computer 31. The computer 31 is equipped with a medicine box 33, which can store medicines 47 and 48 inside. This medicine box has a lid (cover) or door 32. The computer 31 is equipped with a camera (function) 42, a microphone (function) 43, a speaker (function) 44, and an authentication (function) 45, similar to the system shown in Figure 1. It is equipped with image monitors (display screens) 34 and 35. Image monitor 34 can display the surrounding situation, including itself, using images captured by the camera (function) 42. Image monitor 35 can display the pharmacist and / or the pharmacist's AI agent by linking with the system held by the pharmacist. Note that in Figures 1 and 3, the cameras (functions) 12 and 42, microphones (functions) 13 and 43, speakers (functions) 14 and 44, and authentication (functions) 15 and 45 are described as external functions (devices) to computers 11 and 31, but they may also be functions built into computers 11 and 31.

[0040] The healthcare check system of the present invention can be used by installing the application software for the healthcare check system of the present invention on computer 31. When a user (patient, pharmacist) is authorized by authentication (function) 45, the healthcare check system is activated (the system icon may appear on the display screen 34, in which case the user should click the icon), and a dedicated AI agent appears on the image monitor 34. After the patient receives the medicine from the pharmacy, they store the medicines 47 and 48 in the medicine box 33. The medicine box 33 has a cover (lid) 32 that can close the opening of the medicine box 33. The AI ​​agent reads the labels and instructions for the medicines 47 and 48 with the camera (function) 42 (or a built-in camera inside the medicine box) and records the type of medicine, the date and time of administration, etc. Based on the recorded information, the AI ​​agent notifies the patient with an alarm or the like when it is time to take the medicine. When the patient approaches (which can be detected by the AI ​​agent or computer using the camera function), they open the cover 32 and take out the medicine to be taken. (By attaching a handling mechanism to the medicine box, the AI ​​agent will automatically dispense the medication. The AI ​​agent also monitors and records whether the medication has been taken using the camera function 42. Once the patient has finished taking the medication, the medications 47 and 48 are returned (automatically) to the inside of the medicine box 33 and the cover 32 is closed to prepare for the next dose. In this way, by using the medicine box computer with an AI agent using the healthcare check system of the present invention, it is possible to reliably prevent patients from forgetting to take their medication.

[0041] The AI ​​agent-equipped medicine box computer using the healthcare check system of the present invention is also equipped with an image monitor 35 for pharmacists. If the pharmacist also has the healthcare check system of the present invention, the patient's AI agent information is transmitted to the pharmacist and the pharmacist's dedicated AI agent, allowing the pharmacist and the pharmacist's dedicated AI agent to know the patient's medication status in real time and to converse with the patient, such as recommending medication. Communication with the pharmacist can be done wirelessly using the antenna 36 provided on this computer, or via a wired connection to the pharmacist's computer. In this system, the pharmacist becomes like a primary care pharmacist for the patient and can constantly know the patient's condition, so they have more patient information than conventional face-to-face primary care pharmacists and can provide detailed patient guidance and follow-up. Patients can also contact the pharmacist and the pharmacist's dedicated AI agent whenever necessary through the AI ​​agent. Furthermore, the AI ​​agent has accumulated a wealth of accurate pharmaceutical and drug information, so patients can take their medication with peace of mind. Furthermore, if medications are individually classified into separate pill cases and stored in the medicine box 33, the (automatic) handling mechanism will retrieve the medication from the medicine box and present it to the patient when it is time to take their medication. The patient simply needs to take the medication contained in the pill case. This system is extremely safe, secure, and reliable, especially for people who tend to forget their medication, the elderly, patients with dementia, and patients who require care. The AI ​​agent of this computer can communicate with an external server wirelessly (or via wired connection) through the antenna 36, ​​enabling it to continuously learn by obtaining updates on specialized information and necessary pharmaceutical and drug information. The operations and monitoring described in this specification, as well as similar operations and monitoring, can be performed autonomously and independently by the medicine box computer and the AI ​​agent.

[0042] Because AI agents are autonomous (or self-sustaining) computers that can accumulate, process, and apply data on their own, they can compile patients' life logs (medical history, medication information, biometric information, health checkup results, etc.) and utilize them to provide more personalized advice. In this way, by using AI agents, information can be collected and recorded with the patient's consent, and access control can be appropriately implemented, allowing for maximum utilization of data while protecting privacy.

[0043] Currently, pharmacists primarily perform dispensing and auditing duties, but in the future, they are expected to support patients' health management, maintaining their health and preventing illness. The AI ​​agent can constantly monitor and record patients' conditions and rapidly transmit rich and accurate information, thereby improving the quality of pharmacists' work. Voice biomarker technology is developing as a new method for evaluating health status, disease diagnosis, and treatment effectiveness using the characteristics of voice, and the AI ​​agent of this invention can also utilize this voice biomarker technology. Specifically, by routinely accumulating and analyzing conversations with patients and patient voices using the AI ​​agent, the patient's physical and mental condition can be visualized, aiding in the early detection and improvement of changes in the patient's condition, its precursors, and mental health issues, particularly dementia and depressive symptoms. This invention aims to address challenges such as the strain on healthcare in an aging society and the excessive workload of pharmacists by having the AI ​​agent collect, process, and advise on information, thereby reducing the burden on pharmacists and improving the quantity and quality of patient care. If the current situation continues, the healthcare system will become unsustainable due to the strain on medical resources and personnel shortages caused by the aging population. However, applying the healthcare check system with the AI ​​agent of the present invention can help solve these problems. Furthermore, the AI ​​agent can also be used as a preventative measure against illness in nursing care facilities.

[0044] This presents an example of how an AI agent equipped with a medicine box according to the present invention can prevent patients from forgetting to take their medication. The AI ​​agent observes the patient who has woken up in the morning using camera 42 and speaks to the patient through speaker 44, saying, "Good morning," "It's time to take your medicine," and "Look at the medicine box." The patient replies, "Yes, it's time to take my medicine," and "I've looked at the medicine box." The AI ​​agent listens to and interprets this voice using microphone 43 and confirms that the patient is looking at the medicine box 33 by checking their gaze with camera 42. Next, the AI ​​agent opens the lid 32 of the medicine box 33 and places the medicine 47 outside. The patient takes the medicine 47 out of the bag. The AI ​​agent observes with camera 42 that the patient has taken out the medicine 47 and speaks to the patient through speaker 44, saying, "Take your medicine." The AI ​​agent further says, "If you don't know how to take the medicine, I can show you," and if the patient says, "Show me," it demonstrates how to take the medicine on screen 34. The patient says, "Okay, I'll take it," and takes the medicine. The AI ​​agent observes the patient's actions and behavior through camera 42 and says to the patient, "What did it taste like? I'm glad you were able to take it today, let's take it again tomorrow," then returns the medication 47 to the medicine box 33 and closes the medicine box lid 32. In this way, the AI ​​agent of this system provides medication reminders (verbal cues) to support the patient's medication adherence, for example, medicine This can help prevent you from forgetting to take your medication.

[0045] By embedding a biomarker in the medication and receiving signals from the biomarker with sensors on an AI agent, it's possible to confirm that the medication has been taken. Furthermore, pharmacists and their AI agents can also learn about the patient's medication status and engage in conversation using the pharmacist's screen 35. In response to prompts from the patient's individual AI agent, the AI ​​agent can also check how the patient's five senses (sight, hearing, taste, touch, and smell) are functioning. For touch and smell, tactile and odor sensors can be used, or the AI ​​agent can ask questions like "What does it feel like?" or "What does it smell like?" and observe the response via audio or image. The AI ​​agent can comprehensively and chronologically check an individual's health status (e.g., dementia or Parkinson's disease), and track the history of symptom changes in detail. The data is saved daily, and if problematic data is detected in the health maintenance and management system, the AI ​​agent can contact the pharmacist, family, or their AI agent to address the patient's needs. Additionally, the AI ​​agent can display text on the screen for deaf individuals, communicating with them via text instead of voice, providing a chatbot-like interaction. Patients can also communicate using a keyboard (for Braille). In this way, the system and AI agent of the present invention can continuously monitor each patient's health status and medication adherence based on data such as the patient's voice, facial expression, response behavior, and vital signs. Through interactive support via the AI ​​agent, it is possible to realize a so-called "digital family pharmacist" that monitors patients 24 hours a day, 365 days a year. It is also possible to continuously monitor changes and side effects caused by medication and share the results of follow-up with the patient, family, and pharmacist.

[0046] In this system, each user's (pharmacist, doctor, caregiver, patient, etc.) personal AI can be linked on a decentralized (AI) platform. Figure 4 shows an example of a decentralized AI platform. On a decentralized AI platform, each user's own AI operates on their own device, and data is shared between users when needed (P2P data sharing). So-called federated learning takes place, where the models learned by each AI are aggregated on an external server (central server). Furthermore, by combining it with blockchain technology, secure medical and pharmaceutical support can also be realized. Each individual AI agent for users A, B, and C has, for example, a medical history layer 61, a prescription layer 62, a vital signs / health check layer 63, a medical record layer 64, and a general data (related to medical and pharmaceutical matters, etc.) layer 65, and the user's information is written to these layers. For example, in case A, medical history 71 is recorded in the medical history layer 61, vital signs and health checkup information 73 is recorded in the vital signs and health checkup layer 63, medical record information 74 is recorded in the medical record layer 64, and other general information 75 is recorded in the general data layer 65. There is no prescription (presumably), so it is not recorded in the prescription layer 62.

[0047] For user B, medical history 81 is recorded in the medical history layer 61, prescription information 82 in the prescription layer 62, medical record information 84 in the medical record layer 64, and other general information 85 in the general data layer 65. For user C, medical history 91 is recorded in the medical history layer 61, prescription information 92 in the prescription layer 62, and other general information 95 in the general data layer 65. In this way, each user's information is managed and recorded in layers, so information can be easily retrieved and the capacity of each AI agent can be minimized. Therefore, it is easier to operate even on smartphones and tablets that do not have a large capacity, in addition to using SLM. Information exchange through communication between each AI agent and an external server can also be performed quickly. Here, each user's AI agent learns locally on the device, and learning data containing personal information is not sent externally, thus ensuring privacy. General data, specialized information, personal information authorized by each individual, and personal information that does not identify an individual are sent to an external server for management. Thus, the system of the present invention utilizes a distributed small-scale language model (SLM), enabling a mechanism in which individual AI for each patient operates on a device while securely sharing and learning data as needed.

[0048] As described in detail above, the present invention is a system in which AI using SLM technology quickly provides users (patients, pharmacists, etc.) with expert information and abundant advice specific to the medical, pharmaceutical, welfare, and / or nursing care fields, and performs (daily) health management on behalf of or together with users, thereby contributing to the maintenance and improvement of users' health. In addition to the many advantages and effects described above, by using the present invention, the AI ​​agent of this system can collect, process, and advise on information to address issues such as the strain on medical care and the stagnation and decline of medical care due to the shortage of personnel in an aging society, the excessive workload of pharmacists, and preventative measures against illness in nursing care facilities, thereby reducing the burden on pharmacists, supplementing medical personnel, and improving the quantity and quality of patient care. Continuous monitoring and management of patients (for example, 24 / 7 support) is limited and unrealistic with conventional human-centered responses, but by using the AI ​​agent of this system, it is possible to improve the quality of daily health management and medication management for pharmacists and patients.

[0049] Currently, personal health data is provided to individuals through health checkups, test results from hospital visits, and vital signs from smartwatches. Prescription results are also stored within the My Number Portal. However, because the data is provided separately and not integrated, it is difficult to say that comprehensive, chronological assessments are being made regarding whether medication is appropriate or whether health is being maintained. In this invention, a dedicated AI agent for each patient grasps almost all of the patient's health data, and if any abnormality is detected in that health data, the AI ​​agent notifies or warns the patient or relevant parties (e.g., family members, pharmacist) to prompt action. For example, if health data and test results obtained through various media are entered into a health management (data) table (a table that lists health data and test results chronologically, also called a health management table or health management data table, which shows when (who, where) and what kind of health data or test results were obtained, for example, an Excel sheet (registered trademark) or Google Sheets, and including similar spreadsheet software), the AI ​​agent can process the data in that health management (data) table using natural language processing, and communicate and share the results with patients, pharmacists, and family members to help maintain health and manage medication. In particular, the AI ​​agent can look at the test data and vital signs listed in the health management (data) table and issue warnings for anything exceeding a certain level that may indicate illness, prompting patients and related parties to take action.

[0050] When a patient undergoes an examination at a hospital, the results are recorded on an examination results sheet. Similarly, when a patient undergoes a health checkup, they receive a health checkup results sheet. The AI ​​agent performs OCR (Optical Character Recognition) processing on these examination results sheets and health checkup results sheets, and based on the processing results, records the date and data on a health management (data) sheet. Alternatively, the AI ​​agent itself may be equipped with OCR processing software to perform the OCR processing. When digital data of examination results or health checkups is provided by the hospital, the AI ​​agent directly records the date and data on the health management (data) sheet. Examples of OCR tools include Google Cloud Vision API, Azure Computer Vision, and Tegaki OCR.

[0051] When vital data such as blood pressure, pulse, and blood sugar levels are obtained from smartwatches or other devices, the AI ​​agent processes the voice data or digital data directly or via API using voice recognition and / or natural language processing, and records the date and data on the health management (data) sheet. When a prescription is issued by a doctor, the AI ​​agent automatically enters the prescription data into the health management (data) sheet for each date (automatically entering prescriptions from My Number Portal or pharmacy prescription history). If the prescription is issued on paper, it is digitized using OCR processing or image recognition processing and entered into the health management (data) sheet. The AI ​​agent can instruct an external AI capable of OCR processing or image recognition processing to digitize the data and receive that data, or it can install OCR processing or image recognition processing software itself and digitize the data using OCR processing or image recognition processing and enter it into the health management (data) sheet. In the case of electronic prescriptions, the data is either directly and automatically imported or received via API (My Number Portal / pharmacy) and entered into the health management (data) sheet. API connections include, for example, Apple HealthKit, Google Fit, Fitbit API, and Garmin API. Integration methods include, for example, retrieving data in JSON format via each REST API and writing it to a health management (data) table. Automation tools include, for example, Node-RED, Zapier, and Make.

[0052] Anomaly detection in the health management (data) table data is performed by detecting specific conditions (e.g., exceeding the standard value) from the health management (data) table using, for example, a Python script, converting it into natural language (e.g., "Blood pressure on April 10th was 160 / 100, which is high"), and sending it to the SLM AI (AI agent). Available AIs include, for example, SLMs and lightweight LLMs for medical and health use (e.g., LLaMA 2-7B, Mistral 7B) that run on smartphones or small servers. The transmission method is, for example, via API or to the SLM or lightweight LLM in a local environment (LangChain, etc.). In addition, if a diagnosis such as CT, MRI, or ultrasound is performed, the image medical data is sent to an AI specialized for medical image diagnosis (AI for medical image diagnosis) for processing and analysis, and the AI ​​agent receives the analysis results and enters them into the health management (data) table sheet along with the date, and communicates them to the patient, family, and pharmacist. In the health management (data) table, for example, the vertical axis records the date and the horizontal axis records the test results.

[0053] In a chronologically recorded health management (data) sheet, if there is data exceeding the standard value, the AI ​​agent performs natural language processing on that data, sends the processing results to an AI specialized for health checkups (health checkup AI) for analysis, receives the analysis results, enters the data along with the date into the health management (data) sheet, and communicates it to the patient, family, and pharmacist. Furthermore, it performs natural language processing on changes in data from the health management (data) sheet that may exceed the standard value, sends the processing results to an AI specialized for medical use (medical AI) for analysis, enters the analysis results along with the date into the health management (data) sheet, and communicates it to the patient, family, and pharmacist. When a doctor's diagnosis is received, the AI ​​agent has the doctor's voice converted into data using speech recognition processing AI, receives that data, and enters it chronologically into the health management (data) sheet. Naturally, information related to the patient's medical diagnosis is primarily recorded. While the AI ​​agent can process the data directly, if it cannot fully analyze the content, it communicates with a specialized AI (for example, a medical AI) to clarify any unclear points and record them chronologically in a health management (data) sheet. The doctor's diagnosis can be communicated not only to the patient but also to other relevant parties (family, pharmacist, etc.).

[0054] AI agents can interact with doctors directly or indirectly (through the patient) regarding a patient's diagnosis. In the case of direct interaction, for example, the AI ​​agent is installed on the patient's smartphone, and the AI ​​agent listens to the doctor's voice, allowing the AI ​​agent to speak directly from the smartphone and engage in conversation. In the case of indirect interaction, for example, the AI ​​agent is installed on the patient's smartphone, and the AI ​​agent listens to the doctor's voice, displaying the conversation content on the patient's smartphone, allowing the patient to interact based on that. In this case, the AI ​​agent must have speech recognition capabilities (i.e., the AI ​​agent has speech recognition processing capabilities), or a separate speech recognition AI must be installed on the smartphone, and the AI ​​agent communicates with the speech recognition AI within the smartphone. Examples of speech recognition AI include Google Speech-to-Text, Whisper (OpenAI), and Azure Speech API. Examples of conversational AI agents (SLM) include open-source options (e.g., LLaMA 2, Gemma). As a method for transferring input results to a health management (data) table, workflow automation using Microsoft Power Automate / Zapier is an option.

[0055] Communication with external specialized AI can be received, for example, as JSON or text, recorded in a health management (data) table using VBA / Python, and notified to the patient. Sharing with family and pharmacists is done automatically through integration with services such as Microsoft Teams, LINE, and Chatwork. Dashboards can be displayed via a web application (Next.js + Firebase, etc.). For AI that translates data changes into natural language and performs trend analysis, analysis tools (time series) include, for example, Prophet (Meta), LSTM models, and Google AutoML Tables. Analysis tools (trend detection and anomaly detection) include, for example, Amazon Lookout for Metrics and AnomalyDetector (Azure). Analysis results and responses from SLM and lightweight LLM (AI agents) can be received as JSON or text, recorded in a health management (data) table using VBA / Python, and notified to the patient. As a means of sharing with family and pharmacists, automatic integration with services such as Microsoft Teams, LINE, and Chatwork is implemented, and dashboards can be displayed via a web application (Next.js + Firebase, etc.).

[0056] Medical images such as CT scans, MRI scans, gastroscopy images, colonoscopy images, lung X-rays, and ultrasound images can be analyzed in detail using specialized AI for medical image diagnosis. The AI ​​agent stores the patient's medical images on its own smartphone or PC and, as needed, sends the patient's medical images to an external server equipped with specialized AI for medical image diagnosis for analysis and diagnosis. The AI ​​agent then records the results in a health management (data) sheet. Examples of medical AI libraries include MONAI (Medical Open Network for AI) and NVIDIA Clara. Cloud utilization options include AWS HealthImaging and Google Cloud Healthcare API (DICOM compatible). For summarizing results in natural language, text extraction can be performed using, for example, CloudVision. Automatic generation of image reports (e.g., "No abnormalities found in CT image") is also possible using LLM.

[0057] The system automatically generates image reports (e.g., "Abnormality found in ** of CT image" (natural language)) and returns the results to an SLM AI (AI agent). Analysis results and responses from the SLM and lightweight LLM (AI agent) are received as JSON or text, recorded in a health management (data) table using VBA / Python, and notified to the patient. Furthermore, it automatically integrates with Microsoft Teams / LINE / Chatwork, etc., for sharing with family and pharmacists, and displays a dashboard via a web application (Next.js + Firebase, etc.). To summarize, data entry uses tools such as OCR, speech recognition, and smartwatch APIs. Data management uses tools such as Excel and Google Sheets (+ VBA / Apps Script), and automation and notifications use tools such as Power Automate, Zapier, and the LINE API. Visualization methods include tools such as Dash, Streamlit, and web applications (React + Firebase). Data security is also important; for example, personal and medical information must be handled in accordance with P-mark and ISMS standards, and the adoption of a small-scale LLM that can process data in a local environment is suitable (minimizing communication risks).

[0058] It is important for patients to take their medication according to the prescription issued by their doctor. However, patients often forget to take their medication for various reasons. Therefore, we have invented a medicine box computer equipped with an AI agent as shown in Figure 3, and we will provide further specific means to eliminate missed medication doses. We call this the medication monitoring box, which includes the aforementioned medicine box computer. The monitoring (medication) box is equipped with an AI agent for the patient, and for example, the AI ​​agent and the patient will have a conversation about medication as shown below. The monitoring (medication) box can be a terminal such as a smartphone or PC, a pet animal robot or doll equipped with a computer, or any other means that can communicate with the patient. Alternatively, the present invention can also be implemented by having an AI agent installed on an external server, and a means of communication such as conversation with that AI agent (a smartphone with a communication app is also possible). Below is an example of a dialogue between the AI ​​agent and the patient to eliminate missed medication doses.

[0059] Figure 5 is a diagram illustrating an example of a dialogue between an AI agent and a patient to eliminate (prevent) patients from forgetting to take their medication, or a flowchart showing how the AI ​​agent monitors patient medication adherence. (1) Greet and check on the patient's condition through voice interaction. Conversation via the medicine box (AI agent contacts from the medicine box: "Good morning," "How are you feeling today? Are you well?" Patient's response: "Good morning, I'm fine. I'm feeling good today too.") (2) Confirm the day of the week and date. AI agent: "What day of the week is it today? What day is it?" Patient: "It's Tuesday, April 22nd." AI agent: "That's right. Let's take your morning medication today. Please press the medication button." "Once the medication bag comes out, please check if the date written on the bag matches today's date and time." (3) Confirm the date on the medication bag and record the act of taking the medication with a camera. (Example: Confirm today's date on the medication bag and record the act of taking the medication with the camera.) AI agent: "Please hold the medication bag so that the date is visible in front of the camera, and then take the medication facing the camera. Can you see the picture of you taking the medication on the monitor?" (4) Image storage (1 week) and AI-based medication compliance check. (Example: The AI ​​checks the saved images to determine if the medication was taken.) Images of medication taken are sent to the AI. The images are stored for 1 week, and the image checking AI checks the images to confirm that the medication was taken. (5) Provide feedback to the individual based on the AI's assessment results. (Example: "You took your medicine properly!") Specifically, the results are sent to the AI ​​agent. The AI ​​agent then confirms with the individual, "I took my medicine." (6) The AI ​​agent saves the date and time of medication intake in a health management (data) table and notifies family members and pharmacists of the medication record. (Example notification: "Medication taken today at 8:10") (7) The AI ​​answers questions from the patient based on past data. If the patient is worried about forgetting to take their medication, they can ask the AI ​​agent, "Did I take my midday medication today?" The AI ​​agent checks the past medication data in the health management (data) sheet and answers the patient, "You took it at 12:50." (8) Interacting with a doll or robot. (Example: A pet robot speaks to and guides the child.)

[0060] As described above, the present invention's "medication monitoring (medication) box," which allows for conversation with patients, is a support tool for the elderly and home-care patients that integrates AI, IoT, and UX design. The AI ​​agent and / or external server that can communicate with the AI ​​agent have voice dialogue functions, camera image analysis functions, data storage functions, automatic notification functions, and / or robot collaboration functions. The implementation method step by step and the usable AI, apps, and data collaboration technologies are, overall, smart medication support through the integration of AI and IoT (AI x IoT), as follows. (1) The voice dialogue function uses small-scale LLM (SLM), speech synthesis (TTS), and speech recognition (STT) technologies that can be installed on smartphones or robots (computers) to understand patient conversations and provide accurate responses, enabling smooth dialogue with patients and providing confirmation and medication guidance through voice dialogue. Specifically, speech recognition technologies include Whisper (OpenAI), Google Speech API, and Azure Speech; dialogue engine (SLM) technologies include LLaMA 2 + LangChain + TTS (Voicevox, etc.); and speech synthesis technologies include Google TTS / open_jtalk (Japanese language support). By using a Raspberry Pi or smart speaker, a microphone and speaker can be built into the medicine box to interact with patients. Furthermore, the voice data is converted to text, an LLM response is given, and a voice dialogue loop is performed with TTS output.

[0061] (2) Camera images and their analysis functions will use image recognition AI (YOLOv8, Mediapipe, MONAI, etc.) to recognize medication taking actions and verify labels. Specifically, for image-based medication verification, camera connection will be, for example, a USB camera or a smartphone camera (with a linked app), image recognition will be, for example, YOLOv8 to "check the date and time on the medication bag," Mediapipe to "check mouth movements (medication taking behavior)," and OpenCV to perform image preprocessing and saving. For saving, for example, images will be saved locally or to Google Drive for one week and automatically classified by day. For AI-based medication verification and result response, the technology used will be, for example, an image judgment script in Python to flag successful / failed medication taking, the results will be recorded in CSV / Excel (date and time, image name, judgment), and the AI ​​agent will provide feedback such as "You took your medicine properly today." (3) The data storage function will save data locally or in the cloud using technologies such as Excel + Python / Pandas, Google Sheets, and SQLite. In addition, as a method of saving, data will be automatically recorded in a health management (data) table (using OpenPyXL / Pandas, etc.). (4) The automatic notification function uses technologies such as LINE Notify, email API, and Webhook integration to notify relevant parties (e.g., family, pharmacist, pharmacy, hospital). Furthermore, notifications can be automatically sent to family / pharmacists using the LINE Notify API, for example. For instance, a notification such as "April 22nd, 8:10 AM: Medication taken (with image)" could be sent and recorded.

[0062] In the above-mentioned dialogue between the AI ​​agent and the patient, the method for answering the question "Did you take your medicine?" would involve, for example, speech recognition → input into SLM → health management (data) table search (e.g., extracting past rows using Pandas). For interaction with animal robots or dolls, the hardware would be, for example, a Raspberry Pi (with audio input / output and Wi-Fi) + servo motors + microphone and speaker. As a platform, for example, a VUI SDK (e.g., Picovoice, Respeaker, ESP32 voice board) or a small AI engine (e.g., Edge-LLM, TinyML) would be used. As a design example, a cat-shaped robot could say "Good morning~" in the morning, proceed with the conversation, and guide the user through button presses. Interactions such as moving the ears with servos or making the eyes light up could also be added.

[0063] If a patient's (self's) personal AI agent is installed on a mobile device such as a smartphone, and the AI ​​agent is kept constantly ON (operating), it can continuously monitor the patient's health status and behavior from the audio and images received by the smartphone. The AI ​​agent has the function of identifying and analyzing the patient's voice, so it can understand the patient's health status based on the state of the patient's voice. For example, if the patient's voice sounds like they have a cold, or is hoarse or raspy, the AI ​​agent can determine that there is an abnormality in the throat condition and notify the patient of the abnormality and / or encourage them to contact a doctor or pharmacist. If the change is minor, it can also advise on specific medication prescriptions. If the patient is wearing a portable health monitor (e.g., electrocardiograph, blood pressure monitor, pulse meter, electroencephalograph), the AI ​​agent installed on the smartphone can obtain the data and analysis results using the communication function, and the AI ​​agent can analyze the data and inform the patient of the results. Examples of portable health monitoring devices include smartwatches, portable electrocardiographs, and health management watches.

[0064] The AI ​​agent on the smartphone and the AI ​​agent on the medicine box can share data and information with each other. In other words, the smartphone's program processing and communication functions are connected to and utilized by the medicine box. Since the medicine box AI agent is fixed, it cannot obtain data and information about the patient's movement, but by communicating with the smartphone and the medicine box, the medicine box's AI agent can also acquire the data and information held by the smartphone's AI agent. Conversely, through mutual communication, data and information held by the medicine box, as well as conversations and patient expressions transmitted through the medicine box, can be transferred to the smartphone using communication functions. A large LCD screen can be attached to the medicine box to facilitate smoother conversations and information exchange with pharmacists and AI agents on a larger screen. Communication between the smartphone and the medicine box can be done using normal wireless communication when they are far apart, and short-range communication such as Bluetooth when they are close together. By equipping the medicine box with a terminal that allows for wired connection to a smartphone, data exchange of higher capacity and speed becomes possible by connecting the two devices using a connection cable or by placing the smartphone in the terminal (port). Furthermore, simply placing the smartphone in the medicine box activates the communication function, allowing for easy interaction with pharmacists or AI agents with simple operation of the medicine box, enabling early data and information exchange. Simultaneous charging of the smartphone from the medicine box is also possible. Moreover, by combining and coordinating the AI ​​agent on the smartphone with the AI ​​agent on the medicine box, it becomes possible to monitor the patient's health status from both sides, enabling a more detailed understanding of the patient's health condition.

[0065] The medicine box of this invention can be used not only to support medication guidance by a primary care pharmacist, but also to prevent dementia in patients, slow its progression, and stabilize the patient's mental state. This medicine box can be described as an interactive medicine box, and its AI agent engages in dialogue with the patient, using voice analysis, facial expression analysis, motion analysis, and language analysis to prevent or slow the progression of dementia in elderly people living alone, or to detect it early. If the AI ​​agent detects something unusual in the patient's behavior during the dialogue, the AI ​​agent will contact the primary care pharmacist, who can then access the medicine box to check the patient's condition. Depending on the patient's condition, the pharmacist may also be able to interact with the patient in person. The AI ​​agent in the medicine box can also support medication guidance by the primary care pharmacist. The dialogue function of the AI ​​agent in the medicine box goes beyond simply taking medication; the act of dialogue itself contributes to the prevention and treatment of dementia, depression, and other conditions in patients. While medication is just about taking the medicine, the AI ​​agent's checks of daily behavior make it possible to prevent and improve health in a pre-disease state. Thus, the AI ​​agent's interaction with the medicine box enhances the health literacy of patients and their families, ultimately leading to an improvement in overall health literacy in the community. Furthermore, the healthcare provided to local residents by the interactive medicine box with AI agent functionality of this invention promotes healthy longevity among the nation's population, contributing to the containment and reduction of national healthcare costs. It is also possible to equip smartphones with similar functions, and together with the interactive medicine box with AI agent functionality, they can contribute to improving healthcare for patients, communities, and the nation.

[0066] The AI ​​agent in the medicine box utilizes camera, speaker, and microphone functions to provide visual, auditory, and audio capabilities. Using a 360-degree panoramic camera, the AI ​​agent can track the patient even if they move freely within its field of view (for example, around the medicine box). Therefore, the AI ​​agent can monitor and analyze the patient's movements to determine if they are experiencing abnormal conditions, such as the progression of dementia. If the patient falls and becomes immobile, the AI ​​agent can use the speaker function to notify family members or nearby individuals. Even if no one is nearby, the communication function can contact others (such as a pharmacist, ambulance, or caregiver), enabling early intervention in the patient's condition. The microphone function also collects sound from 360 degrees, so if a patient is within its range and makes a sound, the AI ​​agent can monitor and analyze the sound to understand the patient's condition. If there are abnormalities in the patient's voice, such as slurred speech or stammering, the AI ​​agent can recognize the abnormality and notify relevant parties.

[0067] The healthcare check system of the present invention, which has a personalized AI agent for each individual (patient), is equipped with a personalized AI agent in a personalized mobile information terminal (smartphone, etc.) or medicine box (box), making it applicable to remote treatment and remote medication management. For example, decentralized clinical trials (DCTs) are a system that uses digital technology to allow participation in clinical trials from home or nearby medical institutions, reducing the frequent hospital visits that were unavoidable in conventional clinical trials and acquiring data through wearable devices and online consultations. The personalized AI agent equipped in the personalized mobile information terminal (smartphone, etc.) or medicine box (box) of the present invention can be applied to these decentralized clinical trials. Specifically, in online consultations and telemedicine using video calls, when communicating remotely with doctors and pharmacists using a mobile information terminal (smartphone, etc.) or medicine box (box), the patient's AI agent can also participate as an advisor or assistant to the patient, providing support to the patient and assisting in communicating the patient's condition to doctors and pharmacists. 24-hour biometric data, such as heart rate and activity levels, continuously measured using wearable devices, can be acquired by the patient's AI agent via communication from the wearable device. The patient's AI agent can then notify and transmit this objective data to medical institutions and other relevant parties.

[0068] The patient's AI agent stores the investigational drug in the medication box and provides it to the patient when needed, managing the patient's medication status, preventing missed doses or overdoses, and allowing for monitoring of medication adherence in daily life. Accurate data on drug use is reported from the patient's AI agent to relevant medical institutions, doctors, and pharmacists, enabling the clinical trial organization to obtain accurate trial data. Furthermore, patients can participate in the trial while going about their daily lives without much awareness, lowering the barrier to participation and making it easier to take part. Using the AI ​​agent of this invention eliminates the need for long-distance hospital visits or taking time off work, and by using it in conjunction with a wearable device, real-time data can be automatically collected 24 hours a day, 365 days a year. Moreover, in today's world of concentrated specialized medical care, medical institutions that can provide advanced and sophisticated medical care are limited, but using the AI ​​agent of this invention enables remote consultations and other services at these limited medical institutions, making it possible to achieve early improvement of the patient's condition and rapid improvement of the trial content through real-time feedback.

[0069] Furthermore, by using the AI ​​agent of the present invention, clinical trials and treatments can be conducted in a state close to normal daily life, allowing for a more realistic evaluation of drug effects. For pharmaceutical companies and medical facilities, DCT and telemedicine using the AI ​​agent of the present invention make it easier to recruit patients from a wide range of areas, attract more participants to clinical trials, efficiently shorten the trial period, and accelerate the development of new drugs. In particular, it becomes possible to acquire real-world data (realistic medical data collected under actual living conditions) obtained in daily life. Conventional DCT and telemedicine have problems such as requiring lectures on the content to family members or caregivers, requesting them to perform operations on behalf of others, and requiring 24-hour telephone support. However, the AI ​​agent of the present invention can instantly grasp the content of DCT and telemedicine, accurately perform the corresponding tasks (especially medication administration), and communicate with relevant organizations as needed. Furthermore, in DCT and telemedicine, investigational drugs may be dispensed from pharmacies or drugstores, and pharmacists may follow up with subjects (patients). However, the AI ​​agent of the present invention can easily communicate with pharmacies and pharmacists, so when a subject (patient) encounters problems related to medication, the AI ​​agent can support the subject (patient) and resolve those problems based on the guidance of the pharmacist or pharmacy.

[0070] The AI ​​agent in the system of the present invention is capable of managing patients' health, monitoring their health status, and providing detailed support to patients (including health maintenance and improvement), thus fulfilling the role of a primary care pharmacist. Therefore, the present invention can make a significant contribution to addressing the challenge of improving the quantity and quality of primary care pharmacists. The AI ​​agent system of the present invention can centrally manage daily health data and medication guidance history at all times (24 hours a day, 365 days a year), and can quickly notify and share abnormal symptoms and drug interaction risks with pharmacists and family members, enabling prevention of symptom worsening and early intervention for patients. The AI ​​agent has medication history and reminder functions, and can instantly detect missed doses, overdoses, and accidental ingestion of medications. It can also comprehensively analyze vital data and lifestyle data (e.g., sleep, exercise, diet) to detect changes from acute to chronic. The system of the present invention can also automatically share information on medication history, health abnormalities, and drug interaction risks with the (primary care) pharmacist. The AI ​​agent can not only remind patients of medication times but also record and provide feedback on medication behavior and actions. The AI ​​agent of the present invention is conversational and can engage in conversations that are easy for patients to understand, and can respond and act solely through conversation, making it easy for the elderly and children to use. By using the AI ​​agent of the present invention, patient constitution and medication data can be centrally managed, and family members, pharmacists, and medical staff can respond in real time and share information. Furthermore, if the system of the present invention is installed in a robot, for example, it is possible to realize an AI agent that is always attentive to the patient. It goes without saying that if the content described in a certain part of this specification can be applied without contradiction to other parts not described therein, then that content can be applied to those other parts. Moreover, the above embodiments are examples and can be implemented with various modifications without departing from the gist of the invention, and it goes without saying that the scope of the rights of the present invention is not limited to the above embodiments or examples. [Explanation of symbols]

[0071] 10 AI agent, 11 computer, 12 camera (function), 13 Microphone (function), 14 Speaker (function), 15 Authentication device (function), 16 Patient, 17 Pharmacist, 19 Display screen, 21 External (data) server, 22 Big data, 23 Medical information, 24 Pharmaceutical (drug) information, 25 Welfare information, 26 Caregiving information, 32 Lid (cover), 33 Medicine box, 34 Image monitor (display screen), 35 Image monitor (display screen), 36 antennas, 42 cameras (functions), 43 microphones (functions), 44 Speaker (function), 45 Authentication device (function), 47 Medicine, 48 Medicine,

Claims

1. A healthcare check system using a personal AI agent implemented on a computer, (a) Means for outputting text or voice conversations by a chatbot and recording the contents thereof, (b) means of obtaining user health management information, (c) AI processing means in which an AI agent analyzes the user's health condition and provides advice on the health condition based on the information obtained in (a) and (b). A healthcare check system characterized by having an AI agent that is a generative AI that has learned specialized knowledge in the medical and pharmaceutical fields using a small-scale language model.

2. The healthcare check system according to claim 1, characterized in that the AI ​​agent periodically connects to an external server to learn specialized knowledge regarding the medical and pharmaceutical fields of the external.

3. The healthcare check system according to claim 1, characterized in that the healthcare check system having the personal AI agent is connected to a healthcare check system having an AI agent dedicated to pharmacists or pharmacies (collectively referred to as pharmacists, etc.), and that the personal and / or personal AI agent can exchange information with the AI ​​agent dedicated to pharmacists, etc.

4. The healthcare check system according to claim 1, characterized in that the computer is a smartphone, a personal computer, or a small server.

5. The healthcare check system according to claim 1, characterized in that the health management information is data on prescriptions, facial color, voice, and / or vital signs (blood pressure, pulse, body temperature, etc.).

6. The healthcare check system according to claim 1, characterized by providing a user interface that allows patients, pharmacists, chatbots, and patients' families to participate on the same screen and converse with each other.

7. The healthcare check system according to claim 1, further comprising a facial recognition module, which enables access to the healthcare check system with a single click through facial recognition of a patient or pharmacist.

8. The healthcare check system according to claim 1, further comprising an image recognition means that uses the camera function of the computer to photograph a paper medium containing examination data provided by a medical institution, and digitizes the captured image by image recognition.

9. The healthcare check system according to claim 1, characterized by comprising data management means that form storage links with multiple data servers and realize data backup and data preservation in the event of a disaster.

10. The system includes means for analyzing the patient's gaze, facial expressions, and voice responses. The healthcare check system according to claim 1, characterized in that it checks the patient's medication status and health condition by the means and records the results.

11. The healthcare check system according to claim 1, characterized by comprising a speech recognition text conversion function (means) for converting speech during medication guidance into text, and means for storing the converted text data.

12. The healthcare check system according to claim 1, characterized in that it includes means for reading medication information displayed on a medication check calendar using a camera on a computer and confirming the medication status.

13. The healthcare check system according to claim 3, characterized in that it has at least two display screens, one of which displays the user's own information and images captured by a camera, and the other display screen is capable of displaying a pharmacist, etc. (meaning a pharmacist and / or a pharmacy), or an AI agent dedicated to pharmacists, etc.

14. The healthcare check system according to claim 1, characterized in that the healthcare check system having a personal AI agent is equipped with a medicine box for storing the patient's medication, and the patient's medication status can be managed by photographing the patient taking the medication with the camera.

15. The healthcare check system according to claim 14, characterized in that it has a function to record the date and time when a patient takes medication, and to notify the patient when that date and time arrives, or to warn the patient if they do not take the medication at that time.

16. The healthcare check system according to claim 14, characterized in that the medicine box has a lid that opens automatically when the patient takes the medicine.

17. The healthcare check system according to claim 14, characterized in that it has a mechanism that allows stored medicines to be automatically removed from the outside of the medicine box.

18. The healthcare check system according to claim 1, characterized in that when an abnormality occurs in the patient's body, the AI ​​agent observes the affected area (symptom) of the body with the camera function of the computer and advises the patient on whether the abnormal change is due to a side effect of the drug.

19. The healthcare check system according to claim 1, characterized in that the method of taking medication is displayed on a screen of the computer, and the patient is instructed to take the medication.

20. The healthcare check system according to claim 1 or 14, characterized in that the AI ​​agent reads the health checkup results form and / or the physical examination results form and / or prescription by OCR processing, and records the digitized health and / or physical and / or medication data in a health management data table in chronological order.

21. The healthcare check system according to claim 20, characterized in that the AI ​​agent takes in electronic prescription data or obtains medication-related data via API (My Number Portal and / or pharmacy) and records the medication-related data in the health management data table in chronological order.

22. A healthcare check system according to claim 1 or 14, characterized in that the AI ​​agent reads a health management data table in which health-related data (health data) and physical examination results (physical examination results) are recorded in chronological order, and if there is an abnormality in the health data or physical examination results, displays a warning on the health management data table and / or communicates the patient's abnormality to the patient's family and / or the patient's pharmacist.

23. The healthcare check system according to claim 20, characterized in that the health management data sheet contains the patient's medication schedule, the AI ​​agent informs the patient to take their medication according to the schedule, takes a picture of the patient taking their medication with a camera, records that the patient has taken their medication in the health management data sheet, and / or notifies the patient's family and / or the patient's pharmacist that the patient has taken their medication.

24. The healthcare check system according to claim 20, characterized in that when vital data such as blood pressure, pulse rate, and blood glucose level are obtained from a smartwatch or the like, an AI agent performs voice recognition processing and / or natural language processing on the voice data or digital data directly or via API and records the date and data in the health management data table.

25. The healthcare check system according to claim 20, characterized in that, when a prescription is provided on paper, the AI ​​agent can instruct an external AI capable of OCR processing and image recognition processing to digitize the prescription and receive that data, or can equip itself with OCR processing and image recognition processing software to digitize the prescription itself using OCR processing and image recognition processing and enter it into the health management data sheet.

26. The healthcare check system according to claim 20, characterized in that the detection of anomalies in the data of the health management data table involves detecting specific conditions (e.g., exceeding a standard value) from the health management data table, converting them into natural language, and transmitting them to an AI agent.

27. The healthcare check system according to claim 20, characterized in that, in detecting anomalies in the data of the health management data table, the AI ​​agent performs natural language processing on the data, sends the processing results to an AI specialized for health checkups (health checkup AI) for analysis, receives the analysis results, enters the data into the health management data table along with the date, and transmits it to the patient, family, and pharmacist.

28. The healthcare check system according to claim 20, characterized in that when a diagnosis such as CT, MRI, or ultrasound is performed, the image medical data is sent to an AI specialized for medical image diagnosis (AI for medical image diagnosis) for analysis, the AI ​​agent receives the analysis results and enters them into the health management data sheet along with the date, and transmits them to the patient, family, and pharmacist.

29. The healthcare check system according to claim 20, characterized in that, upon receiving a diagnosis from a doctor, the AI ​​agent has the doctor's voice converted into data using speech recognition processing AI, and then receives that data and records it chronologically in a health management data sheet.

30. The healthcare check system according to claim 20, characterized in that, when a patient receives a diagnosis from a doctor, an AI agent having a voice recognition processing function digitizes the doctor's voice and records it chronologically in a health management data sheet.

31. The healthcare check system according to claim 20, characterized in that the AI ​​agent communicates with an AI specialized in that field (for example, a medical AI) regarding the results of a doctor's diagnosis of a patient, clarifies any unclear points, and records them chronologically in a health management data table.

32. The healthcare check system according to claim 1 or 14, characterized in that the AI ​​agent has a voice dialogue function, a camera image analysis function, a data storage function, an automatic notification function, and / or a robot collaboration function, and uses the said functions to prevent patients from forgetting to take their medication.

33. The healthcare check system according to claim 32, characterized in that the AI ​​agent uses the voice dialogue function to prevent patients from forgetting to take their medication through conversation with the patient.

34. The healthcare check system according to claim 33, characterized in that the AI ​​agent uses the camera image and analysis function to photograph the patient taking medication and determines from the image whether the patient has taken the medication.

35. The healthcare check system according to claim 34, characterized in that the AI ​​agent saves the date and time of medication administration in the health management data table based on the judgment result, and further notifies the patient's family and / or pharmacist of the medication record.

36. The healthcare check system according to claim 32, characterized in that when the AI ​​agent receives a confirmation question from the patient regarding whether or not they are taking medication, it looks up past medication data in the health management data table and answers the patient regarding whether or not they are taking medication.

37. The healthcare check system having a dedicated AI agent for each individual (patient) is equipped with a dedicated mobile information terminal (hereinafter referred to as a smartphone), and the dedicated AI agent for each patient is installed on the smartphone. The healthcare check system according to claim 1, characterized in that it has a function to understand the patient's health status using an AI agent installed on the patient's smartphone and notify the patient if there is an abnormality.

38. The healthcare check system having a dedicated AI agent for each individual (patient) is equipped with a medicine box for storing the patient's medication, and the dedicated AI agent for each patient is installed in the medicine box. The healthcare check system according to claim 1, characterized in that it has a function to understand the patient's health status using an AI agent installed in the medicine box and to notify the patient if there is an abnormality.

39. The aforementioned healthcare check system, which has an AI agent dedicated to the individual (patient), is further equipped with a personal mobile information terminal (hereinafter referred to as a smartphone), and the patient-dedicated AI agent is installed on the smartphone. The healthcare check system according to claim 38, characterized in that it has a function to understand the patient's health status using an AI agent installed on the patient's smartphone and notify the patient if there is an abnormality.

40. The healthcare check system according to claim 39, characterized in that the patient's AI agent installed on a smartphone and the patient's AI agent installed on a medicine box are able to communicate with each other and have a function to share the patient's health information held by both AI agents.

41. The healthcare check system according to claim 39, characterized in that the patient's AI agent installed on a smartphone and / or the patient's AI agent installed on a medicine box are capable of interacting with the patient and have the function of understanding the patient's cognitive function through interaction.

42. The healthcare check system according to claim 38, characterized in that the patient's AI agent installed in the medicine box has the function of monitoring and analyzing the movements of patients around the medicine box using a camera function to understand the patient's abnormal condition.

43. The healthcare check system according to claim 38, characterized in that the patient's AI agent installed in the medicine box has the function of monitoring and analyzing the voice of patients around the medicine box using a microphone function to understand the patient's abnormal condition.

44. The healthcare check system according to claim 38, characterized in that, in remote treatment or remote clinical trials, the patient's AI agent installed in the medicine box has the function of managing the patient's medication status and health condition, and transmitting the medication data and health condition to the remote treatment or remote clinical trial institution for communication.

Images