Language model-based prescription information provision method and system

Abstract

A language model-based prescription information provision method according to the present invention may comprise the steps of: receiving an exercise curriculum recommendation request from a medical staff terminal into which login is performed using a medical staff account; acquiring the exercise curriculum in response to the request; and providing the acquired exercise curriculum to the medical staff terminal.

Description

Method and system for providing prescription information based on a language model

[0001] The present invention relates to a method and system for providing prescription information using a language model.

[0002] Musculoskeletal disorders refer to pain or damage occurring in the musculoskeletal system, including muscles, nerves, tendons, ligaments, bones, and surrounding tissues. Musculoskeletal disorders can affect various parts of the body, including the neck, back, arms, and legs.

[0003] According to a report by the World Health Organization (WHO), economic losses due to musculoskeletal disorders are the fourth highest among all diseases, and musculoskeletal disorders are chronic pain that affects not only daily life but also economic activities.

[0004] In principle, treatment for musculoskeletal disorders should begin with less invasive treatments, with non-pharmacological conservative treatments (e.g., exercise therapy and education, cognitive therapy, or relaxation therapy) being performed first, followed by pharmacological and surgical treatments.

[0005] Treatment guidelines actively recommend non-pharmacological conservative treatment for musculoskeletal disorders, and research on methods for performing non-pharmacological conservative treatment for musculoskeletal disorders is actively underway, particularly in the United States and Europe.

[0006] Meanwhile, with the recent rapid development of artificial intelligence (AI) technology, language models capable of natural conversations with people (e.g., ChatGPT) have emerged.

[0007] Unlike existing manually built chatbots that only provide limited responses, language models are demonstrating technological capabilities that enable natural human-like communication and provide fast, accurate information, demonstrating innovation in the artificial intelligence market.

[0008] In line with this trend, industries requiring specialized knowledge, such as the healthcare industry, are also seeing a growing need to provide services using language models.

[0009] Even in industries based on offline channels, it is becoming common to provide medical services online.

[0010] The present invention relates to a method and system for providing prescription information based on a language model that can search for and recommend appropriate prescription information to a patient.

[0011] In particular, the present invention relates to a method and system for providing prescription information based on a language model, which can search and recommend prescription information based on a large language model.

[0012] More specifically, the present invention relates to a method and system for providing prescription information based on a language model capable of training a language model so as to search for and recommend appropriate prescription information according to the patient's condition and the doctor's request.

[0013] In order to solve the problem discussed above, a method for providing prescription information based on a language model according to the present invention may include the steps of: receiving a prescription information recommendation request from a medical staff terminal logged in with a medical staff account; extracting, in response to the request, medical staff meta information of a medical staff member linked to the medical staff account and patient meta information of a patient to be treated linked to the medical staff account from a server; generating a prompt in a preset format using the medical staff meta information and the patient meta information; obtaining the prescription information using a Large Language Model (LLM) that receives the prompt as an input; and providing the obtained prescription information to the medical staff terminal.

[0014] Furthermore, the present invention further includes a step of transmitting the prescription information to at least one medical system linked to the server in response to an approval event occurring in the medical professional account for the prescription information, wherein the at least one medical system may include any one of an Electronic Health Record (EHR) system, an Electronic Medical Record (EMR) system, an Electronic Prescription System (OCS), a Clinical Information System (CIS), and a Personal Health Record (PHR) system.

[0015] Furthermore, the prescription information recommendation request is received in a state where a page of at least one medical system is provided to the medical staff terminal, and in the step of transmitting the prescription information, the prescription information can be output on the page of at least one medical system.

[0016] Furthermore, the preset prompt may include a system prompt and a user prompt, wherein the system prompt may define the role of the server as the medical staff, and the user prompt may specify the patient and the indication of the patient for whom the prescription information is to be recommended.

[0017] Furthermore, in the exercise DB of the server, exercise items for treating the indications are grouped and matched for each indication, and each of the exercise items for treating the indications has information on the difficulty level of the exercise matched therewith, and in the step of obtaining the prescription information, exercise items for treating the indications are searched for from the exercise DB, and the searched exercise items can be assigned to different days constituting a preset exercise period.

[0018] Furthermore, the search for the above exercise items may be performed by considering at least one of the prescription information of the medical staff, the prescription history, the patient's age, the patient's gender, and the patient's exercise tendencies.

[0019] Furthermore, the present invention further includes a step of obtaining an exercise program different from the exercise program using the prompt based on receiving a request for recommendation of the other prescription information from the medical staff terminal, and in the step of obtaining the other exercise program, any one of the exercise items may be excluded from the exercise program, and any other exercise item that is matched to the same group as any one of the exercise items and has the same difficulty level information may be included in the exercise program.

[0020] Furthermore, the present invention further includes a step of performing verification to verify whether the exercise program satisfies a preset verification criterion, and a step of regenerating the prescription information based on the prompt and the language model if the verification result does not satisfy the preset verification criterion, wherein the preset verification criterion may be related to at least one of the degree of relevance to the treatment of the patient's indication, the difficulty of the exercise items included in the prescription information, the age of the patient, and the exercise tendency of the patient.

[0021] Furthermore, the medical staff meta information may include information on at least one of the medical staff's specialty, hospital name, career, and past prescription history, and the patient meta information may include information on at least one of the patient's age, gender, indication, surgical history, diagnosis, pain area, pain score, past exercise prescription record, and tendency.

[0022] Meanwhile, a system for providing prescription information based on a language model according to the present invention includes a communication unit for receiving a prescription information recommendation request from a medical staff terminal logged in with a medical staff account, and a control unit for extracting, in response to the request, medical staff meta information of a medical staff member linked to the medical staff account and patient meta information of a patient to be treated linked to the medical staff account from a server, wherein the control unit can generate a prompt in a preset format using the medical staff meta information and the patient meta information, obtain the prescription information using a Large Language Model (LLM) that receives the prompt as an input, and provide the obtained prescription information to the medical staff terminal.

[0023] Meanwhile, a program according to the present invention is a program executed by one or more processes in an electronic device and stored in a computer-readable recording medium, wherein the program may include commands for performing the steps of: receiving a request for recommendation of prescription information from a medical staff terminal logged in with a medical staff account; extracting, in response to the request, from a server medical staff meta-information of the medical staff linked to the medical staff account and patient meta-information of a patient to be treated linked to the medical staff account; generating a prompt in a preset format using the medical staff meta-information and the patient meta-information; obtaining the prescription information using a Large Language Model (LLM) that receives the prompt as an input; and providing the obtained prescription information to the medical staff terminal.

[0024] As described above, the method and system for recommending prescription information based on a language model according to the present invention receives a prescription information recommendation request from a medical staff terminal logged in with a medical staff account, and in response to the request, extracts medical staff meta-information of the medical staff linked to the medical staff account and patient meta-information of a patient to be treated linked to the medical staff account from a server, and generates a prompt in a preset format using the medical staff meta-information and the patient meta-information. Through this, the present invention enables the giant language model to recommend appropriate prescription information by considering the patient's condition from the medical staff's perspective.

[0025] Furthermore, the method and system for recommending prescription information based on a language model according to the present invention obtains the prescription information using a large language model (LLM) that receives the prompt as input, and provides the obtained prescription information to the medical staff terminal, thereby allowing the patient's information and the medical staff's requirements to be input into the large language model to recommend appropriate prescription information and to have the recommended prescription information reflected in the medical system.

[0026] Figure 1 is a conceptual diagram for explaining a language model-based prescription information provision environment according to the present invention.

[0027] Figure 2 is a conceptual diagram for explaining a language model-based prescription information provision system according to the present invention.

[0028] Figures 3 and 4 are conceptual diagrams for explaining the large language model used in the present invention.

[0029] FIG. 5a and FIG. 5b are flowcharts for explaining a method for providing prescription information based on a language model according to the present invention.

[0030] FIG. 6, FIG. 7a, FIG. 7b, FIG. 7c, FIG. 7d, FIG. 7e, FIG. 8, FIG. 9, and FIG. 10 are conceptual diagrams for explaining a method for providing prescription information based on a language model provided in the present invention.

[0031] Hereinafter, embodiments disclosed in this specification will be described in detail with reference to the attached drawings. Regardless of the drawing numbers, identical or similar components will be given the same reference numbers, and redundant descriptions thereof will be omitted. The suffixes "module" and "part" used for components in the following description are assigned or used interchangeably only for the convenience of writing the specification, and do not in themselves have distinct meanings or roles. In addition, when describing the embodiments disclosed in this specification, if it is determined that a specific description of a related known technology may obscure the gist of the embodiments disclosed in this specification, a detailed description thereof will be omitted. In addition, the attached drawings are only intended to facilitate easy understanding of the embodiments disclosed in this specification, and the technical ideas disclosed in this specification are not limited by the attached drawings, and should be understood to include all modifications, equivalents, and substitutes included in the spirit and technical scope of the present invention.

[0032] Terms that include ordinal numbers, such as first, second, etc., may be used to describe various components, but the components are not limited by these terms. These terms are used solely to distinguish one component from another.

[0033] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components intervening. Conversely, when a component is referred to as being "directly connected" or "connected" to another component, it should be understood that there are no other components intervening.

[0034] Singular expressions include plural expressions unless the context clearly indicates otherwise.

[0035] In this application, terms such as “include” or “have” are intended to specify the presence of a feature, number, step, operation, component, part or combination thereof described in the specification, but should be understood not to exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts or combinations thereof.

[0036] The storage unit (120) may be configured to include at least one of a patient DB (210), a medical staff DB (220), an exercise DB (230), and a prescription DB (240). In FIG. 2, the exercise DB (230) and the prescription DB (240) are depicted as being independent (or separate), but the exercise DB (230) and the prescription DB (240) may exist as a linked DB that is grouped together. For example, the prescription DB (240) may be configured to include the exercise DB (230) (i.e., the exercise DB exists within the prescription DB). In addition, as illustrated in FIG. 2, the exercise DB (230) and the prescription DB (240) may exist separately.

[0037] The patient DB (210) may store and exist patient information (patient user information or medical information), and may include, for example, at least one of the patient's name, age, gender, surgical history, disease (current medical history, past medical history, etc.), indication, diagnosis, pain site, consultation history, diagnosis result, treatment process, examination information (e.g., X-ray image, blood test information, etc.), prescribed exercise curriculum details, prescribed medicine details, patient tendencies, allergy information, patient's lifestyle, eating habits, exercise habits, exercise level and smoking status, and stress level.

[0038] The medical staff database (220) may store medical staff information (medical staff user information or medical practice information). For example, the medical staff database (220) may include at least one of the medical staff's name, age, career, specialty, hospital where they work, and research field.

[0039] The exercise DB (230) may include a plurality of exercise items, and at least some of the difficulty level, description information, exercise video, and exercise method of each exercise item may be matched to each other. The exercise DB (230) may include a plurality of exercise items grouped based on indications, and each of the plurality of exercise items may be matched with description information regarding the name, exercise method, exercise effect, etc.

[0040] The prescription database (240) may store and exist the prescription history of at least one patient or medical professional. The prescription history may include at least one of the following: the prescription date, the prescribing physician, the prescribed patient, the prescribed drug or exercise, the diagnosis result, the treatment process, and the prescription details.

[0041] In the present invention, the patient DB (210), medical staff DB (220), exercise DB (230), and prescription DB (240) can be described by naming them as a database (200) or storage unit.

[0042] The communication unit (130) may be configured to transmit or receive data. The communication unit (130) may receive a service request including a user query from a medical professional's user terminal (20) and transmit prescription information corresponding to the user request. Furthermore, the communication unit (130) may be configured to receive information required by the present invention from an external server.

[0043] The communication unit (130) can support various communication methods according to the communication standards of the communicating device (or electronic device). For example, the communication unit (120) may be configured to perform communication using at least one of WLAN (Wireless LAN), Wi-Fi (Wireless-Fidelity), Wi-Fi (Wireless Fidelity) Direct, DLNA (Digital Living Network Alliance), WiBro (Wireless Broadband), WiMAX (World Interoperability for Microwave Access), HSDPA (High Speed ​​Downlink Packet Access), HSUPA (High Speed ​​Uplink Packet Access), LTE (Long Term Evolution), LTE-A (Long Term Evolution-Advanced), 5G (5th Generation Mobile Telecommunication), Bluetooth™, RFID (Radio Frequency Identification), Infrared Data Association (IrDA), UWB (Ultra-Wideband), ZigBee, NFC (Near Field Communication), Wi-Fi Direct, and Wireless USB (Wireless Universal Serial Bus) technologies.

[0044] Meanwhile, the control unit (130) may be configured to control the overall operation of the prescription information provision system (100) related to the present invention. Specifically, the control unit (130) may process signals, data, information, etc. input or output through the components discussed above, or provide or process appropriate information and functions to the user.

[0045] As illustrated in FIG. 3, the control unit (140) can use information extracted from the database (200) (e.g., patient's age, sex, diagnosis, surgical history, painful area, pain score, past exercise, etc., 310) and a service request from medical staff (e.g., "Recommend a 4-week curriculum with low-difficulty exercises", 320) as inputs to a large language model (LLM, 142), and obtain an exercise curriculum (330) corresponding to the medical staff's request as a response from the large language model (142).

[0046] More specifically, the control unit (140) can identify the intention (requirement) of the medical staff from the user query included in the service request and extract meta information of the patient or medical staff from at least one of the patient DB (210) and the medical staff DB (220). Then, the control unit (140) can generate a prompt using the extracted meta information and search and recommend prescription information using a large language model (142) that receives the prompt as input.

[0047] Furthermore, the control unit (140) can perform a series of data processing to search and recommend prescription information corresponding to a service request from medical staff.

[0048] The control unit (140) can analyze the intention (or requirement) of the medical staff from the service request and analyze (understand or understand) the intention with which the medical staff requested the service.

[0049] The control unit (140) may utilize at least one of various artificial intelligence technologies to analyze the intent of a service request. For example, the artificial intelligence technologies may include at least one of Machine Reading Comprehension (MRC), Paraphrasing, Abstract Summarization, Text Generation (TG), Question Generation (QG), Natural Language Understanding (NLU), OCR-NLP (Textanding), and Natural Language-Based Querying (NL2SQL).

[0050] The prompt generation unit (141) may be configured to design (or configure or generate) an appropriate prompt to obtain a desired result from a language model, and perform “prompt engineering” technology to perform natural language processing tasks by utilizing a very large language model.

[0051] Here, a prompt is an input value for generating a response (or answer or result) from a language model. The prompt contains information related to instructions or commands for the language model, and the language model generates a response based on this.

[0052] The prompt generation unit (141) can understand (analyze) the requirements of medical staff from the user query included in the service request of medical staff through natural language processing, and extract information necessary for prompt generation from a database (at least one of patient DB, medical staff DB, exercise DB, and prescription DB) based on the analysis results.

[0053] The prompt generation unit (141) can generate a prompt using a service request (user query), analysis results, and extracted information so as to obtain prescription information corresponding to the medical staff's requirements from the language model as a response.

[0054] In the present invention, the prompt may include at least one of a system prompt (141a) and a user prompt (141b).

[0055] In the present invention, a “system prompt” is an instruction or command to define (grant) the role of a prescription information provision system (100), defines the persona of the medical staff who is the prescribing subject, and commands the language model to recommend prescription information according to the defined persona (ex: “>[INST] < <sys>>A rehabilitation medicine or orthopedic specialist can play a role in recommending exercises for the patient's rehabilitation." That is, the system prompt can be made to define the server's role as the aforementioned medical professional.

[0056] Furthermore, the “user prompt” is information defining the target of prescription information recommendation (or prescription) by the prescription information provision system (100) (ex: “<< / sys> > A 34 year old male with ankle pain of about 6 points and a diagnosis of simple ankle sprain may be included. That is, the user prompt may be configured to specify the patient and the indication for the patient for whom the prescription information is to be recommended (or prescribed).

[0057] The present invention provides a medical service based on a Large Language Model (LLM), and can recommend various prescription information (e.g., medicines, medical devices, medical tests, treatments, exercise curriculums, exercise programs, exercise plans, cognitive therapy, diet, counseling, etc.) so that medical staff can prescribe appropriate medicines to patients.

[0058] The present invention can be made to be plugged into a medical system using an electronic health record (EHR) and an electronic medical record (EMR).

[0059] Furthermore, the present invention can provide a service that searches for a patient's indications and a prescription target for treatment using at least one of an electronic health record and an electronic medical record, and recommends prescription information including the searched information to medical staff.

[0060] For convenience of explanation, the present invention will be described as an example of recommending prescription information to medical professionals based on the indications of a patient requiring exercise therapy. The indications of a patient requiring exercise therapy may be musculoskeletal disorders. The present invention, based on a language model, searches for exercise items for treating musculoskeletal disorders and recommends an exercise curriculum (or exercise program or exercise plan) containing the searched exercise items to the medical professional as a prescription target.

[0061] In the present invention, an exercise item may mean an exercise movement or a type of exercise, and an exercise curriculum may include at least one exercise item and an exercise method of the exercise item (e.g., exercise schedule, exercise order, number of exercise times, exercise time, etc.).

[0062] Furthermore, the “language model-based prescription information providing system” according to the present invention may be named a “prescription information providing system,” and the system may be named a “server” or a “platform.”

[0063] Hereinafter, a method and system for providing prescription information based on a language model according to the present invention will be described in more detail with reference to the attached drawings. Fig. 1 is a conceptual diagram for explaining an environment for providing an exercise curriculum based on a language model according to the present invention. Fig. 2 is a conceptual diagram for explaining a system for providing prescription information based on a language model according to the present invention, Figs. 3 and 4 are conceptual diagrams for explaining a large language model used in the present invention, Figs. 5a and 5b are flowcharts for explaining a method for providing prescription information based on a language model according to the present invention, and Figs. 6, 7a, 7b, 7c, 7d, 7e, 8, 9, and 10 are conceptual diagrams for explaining a method for providing prescription information based on a language model provided in the present invention.

[0064] In the present invention, a large language model can be used to recommend prescription information for the medical service request. For example, as illustrated in FIG. 1, the present invention can receive a medical service request related to an exercise curriculum prescription (e.g., “Recommend a 3-week exercise curriculum”, 1) from a medical professional. In the present invention, a large language model can be used to recommend an exercise curriculum (2) suitable for prescribing to a patient according to the medical service request. As another example, the present invention can receive a prescription medical service request related to a medicine prescription (e.g., “Prescribe a medicine to treat a 1-year-old girl’s cold symptoms”) from a medical professional, and in response to the request, a large language model can be used to recommend prescription information including a medicine to be prescribed to the patient.

[0065] Medical professionals can request prescription information recommendations by entering the minimum information required for a prescription. The present invention extracts various information for recommending prescription information in response to a service request from a medical information server, uses the extracted information to generate prompts to be input into a large-scale language model, and uses the prompts as input to a pre-trained language model to search and recommend prescription information.

[0066] Here, the medical information server can contain various medical information necessary for medical practice or generated by medical institutions, such as electronic health records (EHRs) and electronic medical records (EMRs). Furthermore, the medical information server can provide medical services, such as diagnosis, prescriptions, treatment, and management of patient health, based on this medical information.

[0067] The present invention can provide services in conjunction with the program (software, etc.) of the prescription information provision system (100) or other medical services. When linked with other medical services, the present invention can provide services by recommending prescription information through communication with the other medical services, or by providing additional (or expanded) functions of the prescription information provision system (100) to the other medical services (e.g., plugins).

[0068] These medical services may be provided by a medical system. The medical system may include at least one of an Electronic Health Record (EHR) system, an Electronic Medical Record (EMR) system, an Electronic Ordering Communication System (OCS), a Clinical Information System (CIS), and a Personal Health Record (PHR) system.

[0069] For example, as illustrated in FIG. 1, the present invention can provide an icon (or GUI, Graphical User Interface, 3) linked to a prescription information provision system (100) while another medical service is being provided on a user terminal of a medical professional. The icon (3) can be provided to the user terminal so as to overlap a screen (ex: “prescription search screen” 10) provided by another medical service. In the present invention, in response to the icon (3) being selected, a page (or screen, 4) linked to the prescription information provision system (100) can be provided to the medical professional terminal. In this case, the present invention can overlap a page (4) of the prescription information provision system (100) with an area of ​​a screen (10) of another medical service, and receive a service request (1) requesting prescription information through the page (4).

[0070] In the present invention, a “service request (or query)” may include the requirements of medical staff, and through analysis of the service request, it is possible to specify which patient is targeted and which prescription is desired. In Fig. 1, “Recommend a 3-week exercise curriculum” includes a prescription type (exercise curriculum) and a prescription period (or a 3-week exercise curriculum period), and although not shown, “Recommend an exercise curriculum for a patient who dislikes repeating the same movement” includes a patient tendency (dislikes repeating the same movement). In the present invention, a large language model can be utilized to search for and recommend prescription information according to the requirements of medical staff, and provide it.

[0071] Meanwhile, as illustrated in FIG. 2, the prescription information provision system (100) according to the present invention may be configured to include at least one of an input unit (110), a storage unit (120), a communication unit (130), and a control unit (140) that performs a series of controls related to provision of prescription information based on a large language model (LLM, 142).

[0072] The input unit (110) may be configured to receive a service request (prescription information request or query) from a medical professional (U). In the present invention, the service request may be input in various formats such as text, voice, video, and image, and the input unit (110) may perform a function by utilizing the hardware configuration of the user terminal (20) of the medical professional (U). For example, the input unit (110) may receive a user request through at least one of the keyboard, mouse, touch screen, and microphone of the user terminal (20).

[0073] In the present invention, the user terminal may also be referred to as an electronic device, and may include a mobile phone, a smart phone, a laptop computer, a digital broadcasting terminal, a PDA (personal digital assistant), a PMP (portable multimedia player), a navigation device, a slate PC, a tablet PC, an ultrabook, a wearable device (e.g., a smartwatch, a smart glass, an HMD (head mounted display)), etc.

[0074] Medical personnel (U) can utilize the prescription information recommendation service based on the large language model provided by the present invention using the user terminal (20) they possess. The medical personnel (U) can prescribe the prescription information recommended by the present invention to a patient, and the patient can proceed with exercise therapy based on the prescribed prescription information through the user terminal they possess.

[0075] In this case, the user terminal (20) can be understood as a terminal logged in with a user account of a medical staff (U) or a terminal logged in with a user account of a patient. Accordingly, in the present invention, “medical staff” and “medical staff account”, “patient” and “patient account” can be used interchangeably. In addition, in the present invention, a user terminal of a medical staff (or a user terminal logged in with a medical staff account) can be named a “medical staff terminal”, and a user terminal of a patient (or a user terminal logged in with a patient account) can be named a “patient terminal”.

[0076] The storage unit (120) can store various information required by the prescription information providing system (100) according to the present invention. The storage unit (120) can be provided within the prescription information providing system (100) according to the present invention or can be provided in an external server (or a database 200 of an external server). In the present invention, the storage unit (120) can be understood as having no physical space limitations as long as it is a space where information required to search and recommend prescription information based on a large language model (142) is stored. Accordingly, in the present invention, the storage unit (120) can be described interchangeably with a database (DB) and a memory.

[0077] Meanwhile, the large language model (142) may include at least one of a prescription information generation unit (142a), a prescription information verification unit (142b), and a learning unit (142c). In this case, the prescription information generation unit (142a) and the prescription information verification unit (142b) may be named in various ways depending on the type of prescription information (e.g., exercise or medicine) recommended by the medical staff. For example, if there is a request for an exercise curriculum recommendation, the prescription information generation unit (142a) and the prescription information verification unit (142b) may also be named an exercise curriculum generation unit and an exercise curriculum verification unit.

[0078] The prescription information generation unit (142a) may be configured to use the prompt generated by the prompt generation unit (141) as input to the macro language model (142), and obtain prescription information corresponding to the service request as a response from the macro language model (142). In addition, the prescription information verification unit (142b) may be configured to verify whether the prescription information generated by the prescription information generation unit (142a) corresponds to the request of the medical staff.

[0079] The learning unit (142c) may be configured to perform learning on the large language model (142) so that prompts corresponding to service requests and prescription information according to the prompts are generated as responses from the large language model (142).

[0080] As illustrated in (b) of FIG. 4, the learning unit (142c) may configure learning data (400) so that appropriate prescription information is recommended to a specific patient according to the role assigned to the prescription information provision system (100). In this case, the learning unit (142c) may configure learning data (400) so that the know-how of prescribing prescription information to a patient is reflected based on the actual prescription history of medical staff by referring to at least one of the patient DB (210), the medical staff DB (220), the exercise DB (230), and the prescription DB (240). The learning unit (142c) may use the learning data (400) to perform learning of the large language model (LLM, 142). The learning data (400) may be composed of sample data sets (410, 420, 430) in a preset format (data structure).

[0081] The format of the sample data set (410, 420, 430) may be predefined to include data corresponding to at least one category (attribute or field). As illustrated in (a) of FIG. 4, the category may include at least some of the following categories: i) “System Role Definition” category (C1), ii) “Recommendation Target Information” category (C2), iii) “Recommendation Target” category (C3), and iv) “Recommendation Result” category (C4).

[0082] The learning unit (142c) can train the large language model (142) to derive at least one of a prompt corresponding to a service request and a response according to the prompt as a result through learning on a sample data set (410, 420, 430).

[0083] For example, the sample data set (410, 420) is. i) Data values ​​(411, 421) corresponding to the “System Role Definition” category (C1) “As a rehabilitation medicine or orthopedic specialist, take on the role of recommending exercises for the patient’s rehabilitation”, “As a rehabilitation medicine specialist who has worked at OO Hospital for 10 years and is in charge of geriatric rehabilitation.”, ii) Data values ​​(412, 422) corresponding to the “Recommendation Target Information” category (C2) “A 34-year-old male, with ankle pain of about 6 points and diagnosed with simple ankle sprain.”, “A 71-year-old female, with a history of lumbar disc surgery, dislikes exercising for more than 20 minutes. She has average flexibility for her age group.”, iii) Data values ​​(413, 423) corresponding to the “Recommendation Target” category (C3) “Recommend a 3-week exercise curriculum”, “Recommend a 10-week exercise curriculum.”, and iv) Data values ​​(414) corresponding to the “Recommendation Result” category (C4) “[ / INST] [“Ankle lifts”, ...]”, “The following exercise curriculum is recommended. It can include bending the waist 10 times~~~~”.

[0084] In the present invention, for the convenience of explanation, the prompt generation unit (141), the prescription information generation unit (142a), the prescription information verification unit (142b), and the learning unit (142c) are not separately distinguished, but are all named as “control unit (140)” and LLM (142) for explanation.

[0085] Meanwhile, the large language model (LLM, 142) can be configured to generate a prompt corresponding to a service request and an exercise curriculum according to the prompt as a response based on learning about the learning data (400). Hereinafter, a language model-based recommendation information providing system (100) according to the present invention will be described in more detail together with the attached drawings. As described above, the present invention can provide various types of prescription information (e.g., medicine, exercise, diet, counseling, cognitive therapy, etc.), and below, for the convenience of explanation, a method of recommending an “exercise curriculum (or exercise program)” as prescription information will be described as an example.

[0086] In the present invention, a process of receiving a prescription information recommendation request from a medical staff terminal logged in with a medical staff account may be performed (S510, see FIG. 5a).

[0087] The control unit (140) can receive a service request including a user query input from a medical staff member from a user terminal (20) logged in with a medical staff member account.

[0088] As described above, in the present invention, a service request including a user query can be directly received from medical staff through the prescription information provision system (100), or a service request can be received from medical staff through another medical service linked to the prescription information provision system (100).

[0089] For example, the control unit (140) may provide a service page provided by the prescription information provision system (100) to the medical staff's user terminal and receive a service request entered into the service page from the medical staff. As another example, as illustrated in FIG. 1, when the function of the prescription information provision system (100) is executed while another medical service (e.g., screen 10 of another medical service) is being provided on the medical staff's user terminal, the control unit (140) may provide an input field (Input Field, 5) for inputting service request information to the user terminal.

[0090] In the present invention, a "service request" may include a user query from a medical professional. For example, as illustrated in Figure 6, a service request (610) may include a user query related to exercise curriculum recommendation, such as "Recommend a 3-week exercise curriculum." A service request may consist of various data types, such as text, voice, video, image, or signal.

[0091] In the present invention, in response to the above request, a process of extracting medical staff meta information of a medical staff member linked to the medical staff member account and patient meta information of a patient to be treated linked to the medical staff member account from the server may be performed (S520, see FIG. 5a).

[0092] When a service request is received, the control unit (140) can specify the medical staff to prescribe the exercise curriculum and the patient to whom the exercise curriculum will be prescribed. The control unit (140) can generate a system prompt that assigns the role of a specific medical staff member to the prescription information provision system (100) and a user prompt that specifies (describes) the patient to whom the exercise curriculum will be prescribed.

[0093] The control unit (140) can obtain an exercise curriculum recommended to a patient described as a specific medical professional as a response by using the system prompt and the user prompt as input to the large language model (142).

[0094] In the present invention, the role of the prescription information provision system (100) may correspond to an actual medical staff or a virtual medical staff.

[0095] The control unit (140) can specify the role (or persona) of the prescription information provision system based on at least one of the user queries included in the medical staff account and service request.

[0096] The control unit (140) can specify the role of the prescription information provision system (100) based on which medical professional's user account the service request was received from. For example, the control unit (140) can specify Doctor A as the role of the prescription information provision system (100) based on the service request being received from Doctor A's user account.

[0097] In addition, the control unit (140) can specify the role of the prescription information provision system (100) through analysis of the user query included in the service request. For example, the control unit (140) can extract information that can identify the medical staff, such as the name, specialty, and career of the medical staff, from the user query, and specify the role of the prescription information provision system (100) based on the extracted information.

[0098] Furthermore, the control unit (140) can primarily specify the role of the prescription information provision system (100) based on the user account of the medical staff, and secondarily change the role of the prescription information provision system that is primarily specified based on the user query.

[0099] More specifically, the control unit (140) can refer to the medical staff DB (220) to identify the medical staff corresponding to the user account and requesting the service. The control unit (140) can compare the information (e.g., name, gender, age, career, specialty, hospital where the medical staff works, etc.) of the medical staff requesting the service with the medical staff corresponding to the user query.

[0100] As a result of the comparison, if the medical staff who requested the service and the medical staff corresponding to the user query are the same, the control unit (140) can specify the role of the prescription information provision system (100) for the medical staff who requested the service. For example, let's assume that a service request including the user query "As a rehabilitation medicine specialist, please recommend an exercise curriculum" is received from a specialist A in the Department of Rehabilitation Medicine. Based on the fact that the specialist A who requested the service and the medical staff included in the user query are the same "rehabilitation medicine specialist," the control unit (140) can specify the role of the prescription information provision system (100) as specialist A.

[0101] In contrast, if the comparison results show that the medical staff who requested the service and the medical staff who responded to the user query are different, the control unit (140) can specify the role of the prescription information provision system (100) based on the user query. For example, let's assume that a service request saying "As an orthopedic specialist with 30 years of experience, please recommend an exercise curriculum" is received from the rehabilitation medicine specialist A. The control unit (140) can specify the role of the prescription information provision system (100) as a virtual "orthopedic specialist with 30 years of experience" rather than the rehabilitation medicine specialist A.

[0102] When the role of the prescription information provision system (100) is specified, the control unit (140) can extract medical staff meta information (620) from the medical staff DB (220), as illustrated in FIG. 6, to generate a system prompt to be input into the large language model (142).

[0103] In the present invention, "medical staff meta-information" refers to information used to implement the role of the prescription information provision system (100), and can be understood as information related to specific medical staff. For example, medical staff meta-information may include at least one of the following: a specific medical staff member's specialty, hospital name, career history, age, gender, hospital where the member works, and research field.

[0104] When the control unit (140) specifies the medical staff who requested the service as the role of the prescription information provision system (100), the control unit (140) can extract medical staff information matched to the medical staff who requested the service as meta information from the medical staff DB (220). Even if the control unit (140) does not specifically receive information about the medical staff who requested the service, the control unit (140) can extract medical staff meta information from the medical staff DB (220) to enable the prescription information provision system (100) to perform the role of the corresponding medical staff. For example, a specialist A in rehabilitation medicine can input a user query such as “recommend an exercise curriculum” without any specific information about himself, and the control unit (140) can extract medical staff meta information related to specialist A from the medical staff DB and assign the role of specialist A to the prescription information provision system (100).

[0105] In addition, the control unit (140) can extract information on other medical staff corresponding to a specific medical staff member designated as a role of the prescription information provision system (100) from the medical staff DB (220) as meta information. For example, the control unit (140) can extract information on other medical staff members corresponding to “rehabilitation medicine specialists like specialist A” as meta information from the medical staff DB (220).

[0106] Furthermore, when a virtual medical staff is specified as a role of the prescription information provision system (100), the control unit (140) can extract medical staff information corresponding to the virtual medical staff as meta information from the medical staff DB (220). For example, when a virtual “orthopedic specialist with 30 years of experience” is specified as a role of the prescription information provision system (100), the control unit (140) can extract various meta information associated with the virtual “orthopedic specialist with 30 years of experience” by referencing the exercise DB (230).

[0107] Meanwhile, when a service request is received from a medical staff account, the control unit (140) can specify which patient among the patients associated with the medical staff account is the service request.

[0108] The control unit (140) can determine which patient the service request is for based on the patient's prescription process during which the service request was received. For example, assume that a service request is received while a prescription page associated with Patient B is being provided on the medical staff's user terminal. The control unit (140) can determine that the service request is for the specific patient based on the fact that the service request was received.

[0109] In this case, the control unit (140) can receive patient information such as name and personal information of the patient along with the service request.

[0110] In addition, the control unit (140) can identify which patient the service request is for by analyzing the user query included in the service request. The user query may include information that can identify the patient (e.g., “Recommend an exercise curriculum for patient Kim Cheol-su”), and the control unit (140) can identify the patient to whom the exercise curriculum is to be prescribed based on the user query. For another example, the user query may include information that can identify the patient’s indication or tendency (e.g., “Recommend an exercise curriculum for a patient with an ankle sprain”), and the control unit (140) can identify the patient to whom the exercise curriculum is to be prescribed based on the user query.

[0111] When a patient to be prescribed an exercise curriculum is specified, the control unit (140) can extract patient meta information (630) from the refund DB (210), as illustrated in FIG. 6, to generate a user prompt to be input into the large language model (142).

[0112] In the present invention, “patient meta information” is information for describing a patient to be prescribed an exercise curriculum, and may include, for example, at least one of patient tendencies (e.g., whether they like exercise, what their personality is, what their lifestyle habits are, etc.), surgical history, age, gender, exercise prescription record, diagnosis, pain area, pain score, and past exercise.

[0113] In the present invention, a process of generating a prompt in a preset format can be performed using the medical staff meta information and the patient meta information (S530, see FIG. 5a).

[0114] The control unit (140) can generate a prompt for deriving an exercise curriculum recommendation response from a large language model by using at least some of the user query and extracted meta information included in the service request.

[0115] As described above, in the present invention, the prompt may include at least one of a system prompt and a user prompt.

[0116] In the present invention, a “system prompt” is an instruction or command to define (assign) the role of a prescription information provision system (100), defines the persona of a medical staff prescribing an exercise curriculum, and commands a language model to recommend an exercise curriculum according to the defined persona (ex: “>[INST] < <sys>>It can be done by a rehabilitation medicine or orthopedic specialist who recommends exercises for the patient's rehabilitation." The role of the server can be defined as the above medical staff.

[0117] Furthermore, the "user prompt" defines the patient profile and patient treatment information (ex: "<) that define the target for recommendation (or prescription) of the exercise curriculum by the prescription information provision system (100).< / sys> > A 34 year old male with ankle pain of about 6 points and a diagnosis of simple ankle sprain may be included. That is, the user prompt may be configured to specify the patient and the indication for which the exercise curriculum is to be recommended.

[0118] The control unit (140) can generate a prompt in a preset format to be input into the macro language model, including at least one of a system prompt and a user prompt. By utilizing the prompt in the preset format and the macro language model, the control unit (140) can assign a specific medical professional role to the prescription information provision system (100) and obtain an exercise curriculum for treating a specific patient's indication as a response from the macro language model assigned the specific medical professional role.

[0119] In the present invention, the preset format of the prompt may be configured to include at least one category (attribute or field) and a data value in the category so as to correspond to the learning data (see reference numeral “400” in (a) of FIG. 4) that the giant language model learns.

[0120] As illustrated in FIG. 6, the category may include at least some of the following categories: i) “System Role Definition” category (C1), ii) “Recommended Target Information” category (C2), and iii) “Recommended Target” category (C3).

[0121] The control unit (140) can generate a prompt (640) by arranging the categories (C1, C2, C3) and data values ​​(641, 642, 643) corresponding to the categories according to a preset format.

[0122] Here, the “System Role Definition” category (C1) and its corresponding data value (641) are commands or instructions for defining the role of the prescription information provision system (100), and may also be understood as “system prompts” in the present invention. Furthermore, the “Recommended Patient Information” category (C2) and its corresponding data value (642) are information describing a patient to whom an exercise curriculum will be prescribed (or recommended), and may also be understood as “user prompts” in the present invention.

[0123] The control unit (140) can sequentially arrange specific categories and data values ​​corresponding to the specific categories so that the large language model (142) can recognize which data value corresponds to which category. For example, as illustrated in FIG. 6, the control unit (140) can sequentially arrange the data value (641) “A rehabilitation medicine or orthopedic specialist plays a role in recommending exercise for a patient’s rehabilitation” in the “System Role Definition” category (C1) to generate a prompt (640). The large language model (142) that receives the prompt (640) can recognize that the data value (641) “A rehabilitation medicine or orthopedic specialist plays a role in recommending exercise for a patient’s rehabilitation” is data belonging to the “System Role Definition” category (C1).

[0124] In addition, the control unit (140) can generate data values ​​(641, 642, 643) corresponding to the categories (C1, C2, C3) using keywords matched to the categories. For example, the data value (641) corresponding to the “system role definition” category (C1) may include the keyword “role (641a)”, and the data value (642) corresponding to the “recommendation target” category (C1) may include the keyword indication name (ex: “ankle sprain”, 642a). In addition, the data value (643) corresponding to the “recommendation target” category (C3) may include the keyword “recommendation (643a).

[0125] Meanwhile, in the preset prompt (640), the “Recommended Target” category (C3) and its corresponding data value (643) may include a script corresponding to the target to be recommended. The control unit (140) may generate a data value (643) corresponding to the recommended target category using a user query (e.g., “Recommend a 3-week exercise curriculum”) included in the service request (610).

[0126] The control unit (140) can analyze the user query to analyze the intent of the service request, and analyze (understand or understand) the intent of the medical staff in requesting the service. The control unit (140) can utilize at least one of various artificial intelligence technologies to analyze the intent of the service request. For example, the artificial intelligence technology may include at least one of Machine Reading Comprehension (MRC), Paraphrasing, Abstract Summarization, Text Generation (TG), Question Generation (QG), Natural Language Understanding (NLU), OCR-NLP (Textanding), and Natural Language-based Query (NL2SQL).

[0127] In the present invention, a process of acquiring the exercise curriculum can be performed using a Large Language Model (LLM) that receives the above prompt as input (S540, see FIG. 5a). In addition, in the present invention, a process of providing the acquired prescription information to the medical staff terminal can be performed (S550, see FIG. 5a).

[0128] The large language model (142) can be understood as a language model trained to generate an exercise curriculum according to the patient's condition and the medical staff's requirements based on learning data (400).

[0129] The control unit (140) inputs and processes prompts including system prompts and user prompts into a large language model, thereby assigning the role of a specific medical staff to the exercise processing system (100) and generating an exercise curriculum for the patient as a specific medical staff.

[0130] For example, as illustrated in FIG. 6, the control unit (140) can input data values ​​corresponding to the “system role definition” category (C1) (e.g., “A rehabilitation medicine or orthopedic specialist recommends exercise for the patient’s rehabilitation” 641) into a large language model to assign a role according to the data value to the prescription information provision system (100). In addition, the control unit (140) can input a prompt into the large language model (142) to specify a patient to be prescribed an exercise curriculum based on data values ​​corresponding to the “recommendation target” category (C2) (e.g., a 4-year-old male, ankle pain of about 6 points and diagnosis of simple ankle sprain 642), and to recommend an exercise curriculum according to data values ​​(643) corresponding to the “recommendation target” category (C3) for the specific patient.

[0131] The control unit (140) can use the large language model (142) and the above prompt to generate an appropriate exercise curriculum that takes into account the patient's indication status as a specific medical professional.

[0132] As previously described, in the present invention, the “exercise curriculum” may include at least one exercise item and an exercise method (e.g., exercise schedule, exercise order, number of exercise sessions, exercise time, etc.) for the exercise item. In addition, the “exercise item” may refer to an exercise movement or type of exercise.

[0133] The control unit (140) can retrieve, from the exercise database (230), exercise items necessary for treating the indications of a patient to whom an exercise curriculum is prescribed. In this case, the control unit (140) can retrieve at least one exercise item necessary for treating the indications of the patient, taking into account the indications, tendencies, age, gender, etc. of the patient described in the prompt, based on the role of the prescription information provision system (100) included in the prompt.

[0134] In the exercise DB (230) of the present invention, at least some of a plurality of exercise items, the difficulty level of each exercise item, explanatory information, exercise images, and exercise methods may be matched to each other. In the exercise DB (230), a plurality of exercise items may be grouped based on indications, and each of the plurality of exercise items may be matched with explanatory information regarding a name, exercise method, exercise effect, etc.

[0135] The control unit (140) can generate an exercise curriculum for the patient using the multiple exercise items found. In this case, the control unit (140) can generate an exercise curriculum including the exercise order, exercise time, exercise date, exercise frequency, exercise method, etc. of multiple exercise items for the treatment of the patient's indications, based on the role of the prescription information provision system (100) and the description of the patient.

[0136] The giant language model (142) can search for exercise items that reflect the patient's condition, exercise prescription situation, medical staff's requirements, etc. among all exercise items included in the exercise DB (230) based on the learning data (400), and can generate an appropriate exercise curriculum using the searched exercise items. More specifically, the giant language model can be trained to generate an exercise curriculum for a preset prescription period for the treatment of a patient's indication specified by a user prompt, as a role defined by a system prompt.

[0137] The giant language model (142) can search for at least one exercise item by referring to the exercise DB (230) and generate an exercise curriculum by assigning the exercise item to each of a plurality of different days that constitute a preset prescription period.

[0138] The control unit (140) can specify a prescription period from a service request. If the service request (e.g., "Prescribe a 3-week exercise curriculum") includes a prescription period, the control unit (140) can specify the corresponding period as the prescription period for the exercise curriculum. Furthermore, the control unit (140) can specify an appropriate prescription period for prescribing an exercise curriculum to a patient based on the service request and the patient's meta-information specified from the patient database (210).

[0139] As illustrated in FIG. 6, the control unit (140) can obtain a target data set (650) including a pair of a prompt and a result (exercise curriculum) according to the prompt from the large language model (142). More specifically, the target data set (650) can include i) a data value (651) corresponding to the “system role definition” category (C1), ii) a data value (652) corresponding to the “recommendation target information” category (C2), iii) a data value (653) corresponding to the “recommendation target” category (C3), and iv) a data value (654) corresponding to the “recommendation result” category (C4). Here, “the data value corresponding to the recommendation result category (C4) (ex: “[ / INST] [“Ankle lift”, ...]”, 654)” can include information on the exercise curriculum according to the prompt.

[0140] In the present invention, obtaining an exercise curriculum (or response) according to a prompt from a large language model can be understood as obtaining a data value (654) corresponding to a target data set (650) or a “recommendation result” category (C4) included in the target data set (650).

[0141] Meanwhile, in the exercise DB (230) of the present invention, exercise items for treating the indication are grouped and matched for each indication, and each of the exercise items for treating the indication may be matched with information on the level of difficulty of the exercise.

[0142] The control unit (140) can control the large language model (142) to search for exercise items for treating the patient's indications from the exercise DB (230) according to the prompt, and to generate an exercise curriculum by assigning the searched exercise items to different days or weeks that constitute a preset exercise period.

[0143] In this case, the control unit (140) can control the large language model to search for exercise items by considering at least one of the medical staff's exercise curriculum prescription history, the patient's age, the patient's gender, and the patient's exercise tendencies according to the prompt, and to allocate the searched exercise items to generate an exercise curriculum.

[0144] Meanwhile, the control unit (140) can transmit prescription information to at least one medical system linked to the server in response to an approval event occurring in the medical staff account for the exercise curriculum.

[0145] As described above, the medical system in the present invention may include any one of an electronic health record (EHR) system, an electronic medical record (EMR) system, an electronic ordering communication system (OCS), a clinical information system (CIS), and a personal health record (PHR) system for patients.

[0146] The control unit (140) can receive the medical staff terminal while the page of at least one medical system is provided. In this case, the control unit (140) can transmit the prescription information to the medical system so that the prescription information is displayed on the page of at least one medical system.

[0147] More specifically, the control unit (140) can generate prescription information recommendation information (also referred to as a report) corresponding to the medical staff's service request information based on the response obtained from the large language model. Furthermore, the control unit (140) can provide the prescription information recommendation information to the medical staff's user terminal.

[0148] As described above, the prescription information providing system (100) according to the present invention can provide various types of prescription information (e.g., medicines, medical devices, medical tests, treatments, exercise curriculums, exercise programs, exercise plans, cognitive therapy, diets, counseling, etc.) as estimated information. For example, as illustrated in FIG. 7a, based on receiving a request for a medicine prescription from a medical professional, the control unit (140) can generate a prompt (700a) for recommending medicine prescription information. Then, the control unit can use the prompt (700a) as an input to a large language model (142) to provide the medical professional with prescription information (700) including an appropriate medicine to be prescribed to the patient. For another example, as illustrated in FIG. 7b, based on receiving a request for an exercise curriculum prescription from a medical professional, the control unit (140) can generate a prompt (710a) for recommending exercise curriculum prescription information. And the control unit (140) can provide the medical staff with prescription information (710) including an exercise curriculum to be prescribed to the patient by using the above prompt (710a) as input to the large language model (142).

[0149] Meanwhile, the control unit (140) can generate the format of the prescription information in various types (types, formats, properties). For example, the control unit (140) can provide exercise curriculum recommendation information as at least one of a prescription type (see reference numeral “700” in FIG. 7a and reference numeral “710” in FIG. 7b), another medical service type (see reference numeral “720” in FIG. 7c), a medical service type of the prescription information provision system (100) (see reference numeral “730” in FIG. 7d), and a medical service type provided to a patient (e.g., an application type installed in a patient terminal, see reference numeral “740” in FIG. 7e).

[0150] The control unit (140) can specify which type of recommended information to provide the exercise curriculum as based on a user query included in the service request (“Show me the exercise curriculum in a prescription format report”).

[0151] In addition, the control unit (140) can specify the recommended information type of the exercise curriculum based on the medical service (e.g., other medical service, medical service of the prescription information provision system (100), etc.) being provided (or executed) to the medical staff terminal. That is, the control unit (140) can provide the exercise curriculum in the same format (or chart) as the service page currently provided to the medical staff.

[0152] When the control unit (140) determines which type of recommended information for the exercise curriculum to provide to the medical staff, it can generate a prompt from the large language model (142) so that the exercise curriculum according to the specified recommended information type is obtained.

[0153] The control unit (140) can generate a prompt to include information about the recommended information type of the exercise curriculum. Furthermore, the control unit (140) can control a large language model to obtain an exercise curriculum of a specific recommended information type based on the prompt.

[0154] For example, suppose a service request includes a user query specifying the recommended information type as “prescription.” The control unit (140) illustrated in FIG. 7b can include a command indicating an exercise curriculum of the “prescription” type (e.g., “Show the exercise curriculum in a prescription-type report”, 710b) in a prompt (710a) input to the large language model.

[0155] For another example, assume that another medical service (e.g., EHR medical service or EMR medical service) is being provided from a medical staff terminal. As illustrated in FIG. 7c, the control unit (140) can include a command (e.g., “Enter the exercise curriculum in the EMR system chart and display it”) indicating an exercise curriculum of the “other medical service” type in the prompt (720a) input to the large language model. (720b)

[0156] As another example, let's assume that a medical service of a prescription information provision system (100) is being provided from a medical staff terminal. As illustrated in FIG. 7d, the control unit (140) may include a command (e.g., "Show the exercise curriculum in the medical staff's digital therapeutic prescription", 720c) indicating an exercise curriculum of the "medical service of the prescription information provision system (100)" type in a prompt (730a) input to the large language model.

[0157] Furthermore, the control unit (140) can control the provision of an exercise recommendation curriculum acquired from a large language model on a patient's terminal, rather than on a medical professional's terminal. As illustrated in FIG. 7e, the control unit (140) can generate a prompt (740a) to include a command (e.g., "Show in patient application", 740b) that causes the exercise curriculum to be provided on the patient's terminal.

[0158] The control unit (140) can obtain exercise curricula of various types based on prompts input into the large language model. As described above, the control unit (140) can obtain information about the exercise curriculum from the large language model or obtain a target data set including the exercise curriculum.

[0159] These exercise curricula or target data sets may contain different data depending on the recommended information type for the exercise curricula. The control unit (140) can recommend various types of exercise curricula to the medical staff terminal based on data obtained from the large language model.

[0160] For example, as illustrated in FIG. 7b, the control unit (140) can provide the exercise curriculum (711) as a prescription type (710). In addition, as illustrated in FIG. 7c, the control unit (140) can output the exercise curriculum (721) to another medical service type (e.g., patient prescription information chart, etc., 720) in conjunction with another medical service (e.g., EHR medical service or EMR medical service). In addition, as illustrated in FIG. 7d, the control unit (140) can recommend the exercise curriculum (731) in the format of a medical service type (e.g., digital therapeutic agent, 730) provided by the prescription information provision system (100). In addition, as illustrated in FIG. 7e, the control unit (140) can control the exercise curriculum (741) to be output to the patient terminal as an application type (740) installed on the patient terminal.

[0161] Meanwhile, the control unit (140) can perform verification of prescription information to ensure reliability of the prescription information.

[0162] The control unit (140) can verify whether the exercise program satisfies preset verification criteria. If the verification result indicates that the preset verification criteria are not met, the prescription information can be regenerated based on the prompt and the language model. The preset verification criteria may be related to at least one of the degree of relevance to the patient's indication treatment, the difficulty of the exercise items included in the prescription information, the patient's age, and the patient's exercise tendency.

[0163] More specifically, as illustrated in FIG. 8, the control unit (140) can generate prescription information by utilizing a large language model (S810). The control unit (140) can perform verification on the generated prescription information.

[0164] The control unit (140) can verify whether the prescription information generated from the large language model is prescription information corresponding to the treatment of the patient's disease and determine whether such verification has been completed (820).

[0165] Based on the determination that the prescription information has been verified, the control unit (140) may provide the prescription information to at least one of the medical staff and the patient (S830). In this case, the control unit (140) may provide the verified prescription information to the patient account (or patient terminal) based on the completion of confirmation (or receipt of medical staff verification information) of the verified prescription information from the medical staff account.

[0166] Alternatively, if the control unit (140) determines, based on the verification result, that the prescription information does not correspond to the patient's indication for treatment, it may generate new prescription information using a large language model (S830). The control unit (140) may verify other prescription information and, based on the verification result, determine whether to provide the other prescription information to at least one of the medical staff and the patient. In this case, the control unit (140) may receive a prescription information regeneration request from the medical staff account (S822), and, based on the received regeneration request, may re-perform the above-described series of processes so that prescription information different from other prescription information is generated.

[0167] There may be various methods for verifying prescription information. The control unit (140) can refer to a database (at least one of the patient DB (210), the medical staff DB (220), the exercise DB (230), and the prescription DB (240)) to verify the actual treatment progress of the patient according to the prescription information. Based on the confirmation result that the treatment progress of the patient's indication is improving according to the prescription information, the control unit (140) can verify that the corresponding prescription information corresponds to the treatment of the patient's indication. On the other hand, based on the confirmation that the treatment progress of the patient's indication is worsening according to the prescription information, the control unit (140) can verify that the corresponding exercise curriculum does not correspond to the treatment of the patient's indication.

[0168] Meanwhile, verification of the exercise curriculum can be performed by medical staff. As illustrated in (a) of Fig. 9, the control unit (140) can provide the exercise curriculum (911) to a specific area (910) on the medical staff's user terminal. In addition, the control unit (140) can provide a function icon for requesting exercise curriculum prescription (e.g., "Prescribe as is", 912) and a function icon for requesting another exercise curriculum (e.g., "Receive recommendation for another exercise curriculum", 913) around the exercise curriculum (911).

[0169] The control unit (140) can prescribe an exercise curriculum (911) to a patient based on the selection of a function icon (912) linked to a prescription request from a medical staff user terminal.

[0170] On the other hand, the control unit (140) can generate an exercise curriculum different from the exercise curriculum (911) using a prompt based on the selection of a function icon (913) linked to a request for a different exercise curriculum from the user terminal of the medical staff.

[0171] The control unit (140) can obtain an exercise program different from the above exercise program by using a prompt and a large language model (142) based on receiving a request for recommendation of the other exercise curriculum from the medical staff terminal.

[0172] In this case, the control unit (140) can obtain a different exercise curriculum by excluding one of the exercise items from the exercise program and updating the exercise curriculum by including another exercise item that is matched to the same group as one of the exercise items and has the same difficulty level information in the exercise DB (230).

[0173] More specifically, the control unit (140) can obtain a different exercise curriculum from the macro language model (142) by using the prompt used to generate the above exercise curriculum (911) as input to the macro language model. When the same prompt is input, the macro language model can generate an exercise curriculum different from the previously generated exercise curriculum. In this case, the control unit (140) can regenerate an exercise curriculum that differs in at least one of the exercise items, exercise order, exercise time, exercise number of times, and exercise method included in the previously generated exercise curriculum.

[0174] Furthermore, the control unit (140) may receive a request from the medical staff to change (replace or replace) at least one exercise item (911a) among the plurality of exercise items included in the exercise curriculum (911). As illustrated in (a) of FIG. 9, when a specific exercise item (911a) among the plurality of exercise items included in the exercise curriculum (911) is selected, and a function icon (913) linked to another exercise curriculum request is selected, the control unit (140) may receive a request to change the specific exercise item (911a) from the user terminal of the medical staff. In response to the change request, the control unit (140) may change the specific exercise item (911a) to another exercise item (911b). Meanwhile, although the plurality of exercise items included in the exercise curriculum (911) are all illustrated as being the same in FIG. 9 (a), it is to be understood that the plurality of exercise items may be different from each other.

[0175] As illustrated in (b) of FIG. 9, the control unit (140) may generate a prompt (920) that commands (or instructs) a change of a specific exercise item (911a) in response to a request for replacing the specific exercise item (911a). More specifically, the control unit (140) may modify the prompt (920) in a preset format so that the target exercise item (921a) to be changed is specified, such as “Recommend a different exercise curriculum. Change the chair heel raise to a different exercise.”

[0176] The control unit (140) can obtain another exercise curriculum with a specific exercise item changed by using the modified prompt (920) as input to the large language model. The control unit (140) can control the large language model to search for another exercise item that has the same indication treatment purpose and the same difficulty level as the exercise item to be changed from the exercise DB (230), and to regenerate the exercise curriculum by including the other exercise item searched for the exercise item to be changed.

[0177] As illustrated in (c) of FIG. 9, in the exercise DB (230), indications (931) and difficulty levels (933) may be matched to each exercise item (932). For example, the indication “ankle sprain (931a)” may be matched to the exercise items of chair-supported heel lift (932a), standing ankle movement (932b), and ankle passive dorsiflexion (932c), and difficulty level information (933a to 933c) may be matched to each of the exercise items (932a to 932c). The control unit (140) can control the large language model (142) to generate a new exercise item by including another exercise item (ex: standing and moving the ankle, 932b) that has the same indication (ankle sprain, 931a) as the exercise item to be changed ("holding on a chair and lifting the heel", 932a) and the same difficulty level (ex: "difficulty level 1", 933a, 933b) by referring to the exercise DB (230).

[0178] Meanwhile, the control unit (140) can generate a prompt using various information and obtain an exercise curriculum to be prescribed to a patient from a large language model using the prompt.

[0179] As previously described, the control unit (140) can use medical staff meta-information to specify the role of the prescription information provision system (100), and can use patient profiles and patient treatment information to specify patients to be prescribed (or recommended) an exercise curriculum. In addition, the control unit (140) can use various information necessary for recommending an optimal exercise curriculum, such as a prescription history prescribed by past medical staff, feedback information received from medical staff, and patient tendency information (e.g.).

[0180] The control unit (140) can generate a prompt in a preset format using various information. In the present invention, the prompt can include at least some of i) the “system role definition” category (C1), ii) the “recommendation target information” category (C2), and iii) the “recommendation target” category (C3). The control unit (140) can configure the plurality of categories (C1, C2, C3) using various information. For example, the control unit (140) can configure the data values ​​of the “recommendation target information” category (C2) using patient tendency information. Furthermore, the control unit (140) can generate a prompt by including a new category and a data value corresponding to the new category in addition to the categories (C1, C2, C3).

[0181] For example, as illustrated in FIG. 10, the control unit (140) can extract past prescription history information (e.g., prescription content, prescription reason, post-prescription follow-up information, etc., 910) of the medical staff who requested the service from at least one of the medical staff DB (220) and the prescription DB (240). In addition, the control unit (140) can utilize the past prescription history information (910) of the medical staff who requested the service to configure a data value (1020) corresponding to the “system’s prescription history” category (C5). In this case, the control unit (140) can also utilize the medical staff’s past prescription history information (910) to configure a data value corresponding to the “system’s role definition” category (C1).

[0182] As another example, the control unit (140) can extract patient tendency information (1030) from at least one of the patient DB (210) and the user query, and use the extracted patient tendency information to configure a data value (ex: “This patient hates repeating the same action multiple times”, 1040) of the “information on target to be recommended” category (C2).

[0183] In this way, the control unit (140) generates a prompt (1050) using various information, and can obtain a patient-tailored exercise curriculum by using this prompt (1050) as input to a large language model. As described above, the control unit (140) can also obtain a target data set (1060) consisting of multiple categories and data values ​​corresponding to the categories.

[0184] As described above, the method and system for recommending an exercise curriculum based on a language model according to the present invention can receive an exercise curriculum recommendation request from a medical staff terminal logged in with a medical staff account, and in response to the request, extract medical staff meta-information of the medical staff linked to the medical staff account and patient meta-information of a patient to be treated linked to the medical staff account from a server, and generate a prompt in a preset format using the medical staff meta-information and the patient meta-information. Through this, in the present invention, the giant language model can recommend an appropriate exercise curriculum considering the patient's condition from the medical staff's perspective.

[0185] Furthermore, the method and system for recommending an exercise curriculum based on a language model according to the present invention obtains the exercise curriculum using a large language model (LLM) that receives the prompt as input, and provides the obtained exercise curriculum to the medical staff terminal, thereby allowing the patient's information and the medical staff's requirements to be input into the large language model to recommend an appropriate exercise curriculum and to have the recommended exercise curriculum reflected in the medical system.

[0186] Meanwhile, computer-readable media include all types of recording devices that store data that can be read by a computer system. Examples of computer-readable media include hard disk drives (HDDs), solid-state disk drives (SSDs), silicon disk drives (SDDs), ROMs, RAMs, CD-ROMs, magnetic tapes, floppy disks, and optical data storage devices.

[0187] Furthermore, the computer-readable medium may include a storage device, such as a server or cloud storage, accessible to electronic devices via communication. In this case, the computer can download the curriculum according to the present invention from the server or cloud storage via wired or wireless communication.

[0188] Furthermore, in the present invention, the computer described above is an electronic device equipped with a processor, i.e., a CPU (Central Processing Unit), and there is no particular limitation on its type.

[0189] Meanwhile, the above detailed description should not be construed as limiting in any respect and should be considered illustrative. The scope of the present invention should be determined by a reasonable interpretation of the appended claims, and all modifications within the equivalent scope of the present invention are intended to be included within the scope of the present invention.

Claims

1. A step of receiving a prescription information recommendation request from a medical staff terminal logged in with a medical staff account; In response to the above request, a step of extracting medical staff meta information of a medical staff member linked to the medical staff member account and patient meta information of a patient to be treated linked to the medical staff member account from the server; A step of generating a prompt in a preset format using the above medical staff meta information and the above patient meta information; A step of obtaining the prescription information using a large language model (LLM) that receives the above prompt as input; and A method for providing prescription information based on a language model, characterized in that it includes a step of providing the obtained prescription information to the medical staff terminal.

2. In paragraph 1, In response to an approval event occurring in the medical professional account for the prescription information, further comprising a step of transmitting the prescription information to at least one medical system linked to the server, At least one of the above medical systems, A method for providing prescription information based on a language model, characterized in that it includes any one of an electronic health record (EHR) system, an electronic medical record (EMR) system, an electronic prescription system (OCS, Ordering Communication System), a clinical information system (CIS), and a personal health record (PHR) system for patients.

3. In paragraph 2, The above prescription information recommendation request is, At the medical staff terminal, at least one page of the medical system is received in a state in which it is provided, In the step of transmitting the above prescription information, A method for providing prescription information based on a language model, characterized in that the prescription information is output on a page of at least one medical system.

4. In paragraph 1, The above preset prompts include system prompts and user prompts, The above system prompt defines the role of the above server as the above medical staff, A method for providing prescription information based on a language model, characterized in that the user prompt is configured to specify the patient and the indication of the patient for whom the prescription information is recommended.

5. In paragraph 4, In the exercise DB of the above server, For each indication, exercise items for treating the above indications are grouped and matched, For each of the exercise items for the above indication treatment, information on the difficulty level of the exercise is matched and exists. In the step of obtaining the above prescription information, Search for exercise items for the treatment of the above indication from the above exercise DB, A method for providing prescription information based on a language model, characterized in that the searched exercise items are assigned to different days (Days) that constitute a preset exercise period and an exercise program is prescribed.

6. In paragraph 5, A method for providing prescription information based on a language model, characterized in that the search for the above exercise item is performed by considering at least one of the prescription information prescription history of the medical staff, the patient's age, the patient's gender, and the patient's exercise tendency.

7. In paragraph 5, Further comprising a step of obtaining prescription information different from the above prescription information using the prompt based on receiving a request for recommendation of other prescription information from the above medical staff terminal, In the step of obtaining the above other prescription information, A method for providing prescription information based on a language model, characterized in that any one of the above exercise items is excluded from the exercise program, and another exercise item that matches the same group as any one of the above exercise items and has the same difficulty level information is included in the prescription information.

8. In paragraph 1, A step of verifying whether the above prescription information satisfies the preset verification criteria; and If the verification result does not satisfy the above-described verification criteria, the step of regenerating the prescription information based on the prompt and the language model is further included. The above-mentioned verification criteria are: A method for providing prescription information based on a language model, characterized in that it is related to at least one of the degree of relevance to the indication treatment of the patient, the difficulty of the exercise items included in the prescription information, the age of the patient, and the exercise tendency of the patient.

9. A communication unit that receives a request for prescription information recommendation from a medical staff terminal logged in with a medical staff account; and In response to the above request, a control unit is included that extracts medical staff meta information of a medical staff member linked to the medical staff member account and patient meta information of a patient to be treated linked to the medical staff member account from the server, The above control unit, Using the above medical staff meta information and the above patient meta information, a prompt in a preset format is generated, By using a large language model (LLM) that receives the above prompt as input, the above prescription information is obtained, A language model-based prescription information provision system characterized in that the acquired prescription information is provided to the medical staff terminal.

10. A program that is executed by one or more processes in an electronic device and stored in a computer-readable recording medium, The above program is, A step of receiving a prescription information recommendation request from a medical staff terminal logged in with a medical staff account; In response to the above request, a step of extracting medical staff meta information of a medical staff member linked to the medical staff member account and patient meta information of a patient to be treated linked to the medical staff member account from the server; A step of generating a prompt in a preset format using the above medical staff meta information and the above patient meta information; A step of obtaining the prescription information using a large language model (LLM) that receives the above prompt as input; and A program stored in a computer-readable recording medium, characterized in that it includes commands for performing a step of providing the acquired prescription information to the medical staff terminal.

Images