System and method for automatically generating personalized health education and computer program product thereof

Abstract

This invention discloses a method for automatically generating personalized health education guidelines, comprising: first, processing input medical information received from a first user to generate textual medical information; then, using a large language model to generate a medical information summary; next, processing medical records received from a second user to generate textual diagnostic information; then, using a large language model to generate a diagnostic summary; then, based on the chief complaint and past medical history in the medical information summary and the chief complaint and diagnosis in the diagnostic summary, retrieving corresponding health education content from a health education database; and finally, using a large language model to process the chief complaint and past medical history in the medical information summary, the chief complaint and diagnosis in the diagnostic summary, and the retrieved corresponding health education content to generate personalized health education guidelines. This invention further discloses a system and computer program product for automatically generating personalized health education guidelines.

Description

Technical Field

[0001] This invention relates to the automatic generation of health education or instructions, and more specifically, to a system, method, and computer program product for automatically generating personalized health education guidelines. Prior Technology

[0002] Currently, the general medical consultation process can be roughly divided into three stages: pre-consultation, during-consultation, and post-consultation. Pre-consultation stage: Patients register at the hospital counter. During registration, staff may first confirm the desired outpatient department through simple inquiries. After registration, patients wait in the waiting area for their number to be called. While waiting, some hospitals or clinics may have outpatient assistants or nurses proactively inquire about the patient's chief complaint and needs for later reference by the doctor. However, this proactive pre-consultation inquiry may be difficult to implement given the current shortage of medical staff and would also increase the workload of medical personnel.

[0003] Upon entering the consultation room, the consultation process begins. During this stage, the outpatient physician will inquire about the patient's chief complaint, symptoms, and needs. After conducting appropriate examinations, the physician will provide a diagnosis, anticipated treatment plan, medication administration instructions, and patient education. However, since the entire process is conducted through dialogue, patients may not be able to fully grasp or remember the medical advice.

[0004] After a consultation, some hospitals or clinics will ask nurses to prepare relevant health education materials (one or more copies, depending on the situation and the patient) for the patient to take home and read. Because different conditions and symptoms need to be considered, the content of the health education materials is quite detailed, and not the entire document is relevant to the patient's current symptoms. For example, separate precautions are described for stages one, two, and three of hypertension; patients need to select the sections that best suit their situation to read and understand. However, due to the complexity of the materials, patients often simply put them away without reading them.

[0005] Therefore, how to ensure that the public obtains appropriate medical advice or health education materials is a topic of active discussion in the medical community. Summary of the Invention

[0006] To address the aforementioned and other issues, this invention discloses a system, method, and computer program product for automatically generating personalized health education guidelines.

[0007] A method for automatically generating personalized health education guidelines, executed by a processor, includes: processing input symptoms received from a first user to generate symptom text data, and then processing the symptom text data using a large language model to generate a symptom summary; processing medical records received from a second user to generate diagnostic text data, and then processing the diagnostic text data using a large language model to generate a diagnostic summary; retrieving corresponding health education content from a health education database based on the chief complaint and past medical history in the symptom summary and the chief complaint and diagnosis in the diagnostic summary; and processing the chief complaint and past medical history in the symptom summary, the chief complaint and diagnosis in the diagnostic summary, and the retrieved corresponding health education content using a large language model to generate personalized health education guidelines.

[0008] A computer program product, which is loaded onto a computer to execute the aforementioned method for automatically generating personalized health education guidelines.

[0009] A computer-readable recording medium storing instructions, which can be executed by a computing device or computer through a processor and / or memory, to perform the aforementioned method of automatically generating personalized health education guidelines when the computer-readable recording medium is executed.

[0010] A system for automatically generating personalized health education guidelines includes: processing input symptoms received from a first user to generate symptom text data, and then processing the symptom text data using a large language model to generate a symptom summary; processing medical records received from a second user to generate diagnostic text data, and then processing the diagnostic text data using a large language model to generate a diagnostic summary; retrieving corresponding health education content from a health education database based on the chief complaint and past medical history in the symptom summary and the chief complaint and diagnosis in the diagnostic summary; and using a large language model to process the chief complaint and past medical history in the symptom summary, the chief complaint and diagnosis in the diagnostic summary, and the retrieved corresponding health education content to generate personalized health education guidelines.

[0011] In the system, method, and computer program product for automatically generating personalized health education guidelines, the symptom summary also includes the time of occurrence, medication status, concerns, and requests, while the diagnostic summary also includes medication, treatment, and health education.

[0012] In the system, method and computer program product for automatically generating personalized health education guidelines, the multimodal input processing module accepts voice input, handwritten text input and typing input.

[0013] In the system, method and computer program product for automatically generating personalized health education guidelines, the health education database is constructed by converting the format and content of multiple health education documents into multiple health education document blocks and storing them in the health education database.

[0014] In the system, method and computer program product for automatically generating personalized health education guidelines, the personalized health education guide generation module first uses a large language model to judge the chief complaint and past medical history in the symptom summary and the chief complaint and diagnosis in the diagnosis summary to generate at least one disease and its symptoms. Then, based on the at least one disease and its symptoms, the module retrieves the corresponding health education document block in the health education database. Each symptom corresponds to a retrieved health education document block to constitute the corresponding health education content.

[0015] In the system, method, and computer program product for automatically generating personalized health education guidelines, the symptom summary processing module, the diagnosis summary processing module, and the personalized health education guide generation module respectively transmit the symptom summary, the diagnosis summary, and the personalized health education guide to the hospital information system (HIS) integration management module.

[0016] In the system, method, and computer program product for automatically generating personalized health education guidelines, the large language model (LLM) used by the symptom summary processing module, the diagnosis summary processing module, and the personalized health education guide generation module is composed of an artificial neural network with multiple parameters and is trained using self-supervised learning or semi-supervised learning. The types and number of parameters of the large language model used by the symptom summary processing module, the diagnosis summary processing module, and the personalized health education guide generation module are different.

[0017] In other words, this invention can simultaneously support the processing of symptoms described by a first user (such as a patient or their family) and consultation conversations with a second user (such as a doctor or nurse), automatically generating a fixed-format summary. This summary content can then be integrated into the Hospital Information System (HIS), accelerating the medical consultation process. Furthermore, by combining the symptom summary described by the first user (such as a patient or their family), the diagnostic summary by the second user (such as a doctor or nurse), and based on the hospital's existing health education documents and manuals, it automatically generates personalized health education guidelines tailored to the patient, which are provided to the first user (such as a patient or their family) for reference after the consultation. Therefore, this invention solves the problem of nurses having to manually select appropriate health education documents, which is time-consuming and labor-intensive. Moreover, compared with the single-sheet health education guidelines for a single disease in the prior art, the health education guidelines generated by this invention first divide the existing health education documents for different diseases into multiple blocks, and then combine and integrate them into personalized health education guidelines suitable for the first user (such as the patient or their family) based on the chief complaint and past medical history in the symptom summary of the first user (such as the patient or their family) and the chief complaint and diagnosis in the diagnosis summary of the second user (such as the doctor or nurse). Simple Explanation of the Diagram

[0018] Figure 1 is a flowchart illustrating the method for automatically generating personalized health education guidelines according to the present invention.

[0019] Figure 2 is a schematic diagram of the architecture of the system for automatically generating personalized health education guidelines according to the present invention. Implementation

[0020] The following specific embodiments illustrate the implementation of this invention. Those skilled in the art can easily understand the other advantages and effects of this invention from the content disclosed herein. The structures, ratios, sizes, etc., illustrated in the accompanying drawings are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed herein, and are not intended to limit the implementation conditions of this invention. Therefore, any modifications, changes, or adjustments, without affecting the effects and objectives achieved by this invention, should still fall within the scope of the technical content disclosed herein.

[0021] The terms “comprising,” “including,” “having,” “containing,” or any other variation thereof, as used herein, are intended to cover non-exclusive inclusion. Unless otherwise stated, singular terms such as “a,” “one,” and “the” also apply to plural terms, while terms such as “or,” “and / or,” etc., are used interchangeably.

[0022] Please refer to Figure 1, which illustrates the flow of the method for automatically generating personalized health education guidelines according to the present invention, including steps S11 to S19, which can be executed by a processor.

[0023] In step S11, the health education documents are converted to a different format. In one embodiment, readable electronic formats such as PDF, text, images, and DOC are converted to TXT format.

[0024] In step S12, the health education documents are converted into health education document blocks. In one embodiment, each health education document is segmented with a fixed number of words (e.g., 500 words) to convert it into a different number of blocks.

[0025] In step S13, the document is stored in a health education database. In one embodiment, each block of a health education document in a fixed text format is vectorized and then stored in the health education database as a file in a fixed format (e.g., JSON format).

[0026] Steps S11-S13 constitute the data preparation stage. Hospitals and clinics typically provide patient education guidelines based on verified content, rather than readily available online articles. Most hospitals maintain their own patient education document database, which can be directly provided to patients when needed. This database is then converted into a vector database to facilitate quick retrieval of relevant documents through vector comparison during subsequent searches, thus completing the preliminary preparation of patient education data.

[0027] In step S14, the input symptoms received from the patient are processed to generate textual symptom data.

[0028] In step S15, the textual data of symptoms is processed using a large language model to generate a symptom summary.

[0029] Steps S14 and S15 constitute the pre-diagnosis stage. Patient-inputted symptoms (e.g., text input, verbal description) are processed to generate a verbatim symptom transcript. Using an LLM model, the transcript is transformed into a fixed-format symptom summary, including six themes: chief complaint, time of occurrence, past medical history, medication history, concerns, and requests. This symptom summary can then be displayed on the physician's computer or tablet via the HIS for reference before the consultation.

[0030] In step S16, the consultation records received from the physician are processed to generate diagnostic text data.

[0031] In step S17, diagnostic text data is processed using a large language model to generate a diagnostic summary.

[0032] Steps S16 and S17 constitute the consultation phase. The physician's consultation process is recorded, such as through verbatim transcription or by the physician inputting diagnostic text. The LLM model is used to generate a fixed-format physician diagnostic summary from the verbatim transcript, including five themes: chief complaint, diagnosis, medication, treatment, and health education. Furthermore, the diagnostic summary can also be used by the physician's information system (HIS) for reference when writing diagnostic records.

[0033] In step S18, based on the chief complaint and past medical history in the symptom summary and the chief complaint and diagnosis in the diagnosis summary, the corresponding health education content is retrieved from the health education database.

[0034] In step S19, a large language model is used to process the chief complaint and past medical history in the symptom summary, the chief complaint and diagnosis in the diagnosis summary, and the retrieved corresponding health education content to generate personalized health education guidelines.

[0035] Steps S18 and S19 constitute the post-diagnosis stage. The chief complaint and past medical history from the pre-diagnosis symptom summary are combined with the chief complaint and diagnosis from the in-diagnosis diagnostic summary. After vector transformation, the data is searched in the health education database to find relevant health education content. Based on this, LLM (Local Health Management) is used to generate personalized health education guidelines from the three data sets: symptom summary, diagnostic summary, and retrieved health education content. Furthermore, these personalized health education guidelines can be integrated into the hospital's internal procedures through the HIS (Hospital Information System) for future reference by patients.

[0036] Therefore, the method for automatically generating personalized health education guidelines in this invention is based on generative AI technology, combining the three major processes of hospital visits (pre-visit, during-visit, and post-visit). It unifies the processing of multimodal input through a front-end, supporting both pre-visit patient self-reports and in-visit dialogue content. A large-scale language model generates a fixed-format summary, and extracts portions of the patient's self-reported symptoms and the doctor's diagnosis summary. Based on a health education database, it generates personalized post-visit health education guidelines. In addition to providing more personalized health education content, the system and streamlined processes help reduce the administrative burden on medical staff, allowing them to dedicate more time and energy to clinical care.

[0037] Furthermore, the present invention provides a computer program product that executes one or more of the above methods after being loaded onto a computer. In addition, the computer program (product) can be stored on a recording medium or directly transmitted and provided over a network; that is, the computer program (product) is something containing a computer-readable program and is not limited to any particular external form. The computer includes, but is not limited to, electronic devices with processors, such as servers.

[0038] Furthermore, the present invention also provides a computer-readable recording medium, which is applied in a computing device or computer having a processor and / or memory. The computer-readable recording medium stores instructions, and the computing device or computer can execute the computer-readable recording medium through the processor and / or memory to perform the aforementioned methods and / or content when executing the computer-readable recording medium. The computer-readable recording medium (e.g., hard disk, floppy disk, optical disk, USB flash drive) stores the computer program (product). In one embodiment, the computer-readable recording medium is a non-transitory computer-readable recording storage medium.

[0039] Please refer to Figure 2, which illustrates the architecture of the system for automatically generating personalized health education guidelines according to the present invention. The system includes a health education data conversion module 21, a health education database 22, a multimodal input processing module 23, a symptom summary processing module 24, a diagnosis summary processing module 25, a personalized health education guide generation module 26, and a hospital information system integration and management module 27.

[0040] The health education data conversion module 21 is used to convert the format and content of multiple health education documents into multiple health education document blocks and store them in the health education database 22. In one embodiment, all health education documents and manuals that a hospital or clinic may provide to patients are organized, limited to readable electronic formats such as PDF and TXT. Other formats such as images and DOC are converted to other formats to facilitate reading by subsequent modules. Through the health education data conversion module 21, health education documents can be converted to TXT format, and each health education document is segmented with a fixed number of characters (e.g., 500 characters) to convert health education documents of different lengths into different numbers of blocks. Through the health education data conversion module 21, each block of the fixed TXT format health education document is vectorized and stored in a fixed JSON format as a file, completing the construction of the health education database 22. As shown below.

[0041] Table 1: Example Explanation.

[0042] Table 2: Example Data.

[0043] The multimodal input processing module 23 processes the input symptoms received from the patient to generate textual symptom data, and processes the consultation records received from the physician to generate textual diagnostic data. In one embodiment, the multimodal input processing module 23 can uniformly process user interface input, such as: spoken descriptions via voice recognition and text input via typing. Before the consultation, the patient uses different media (such as recording devices and keyboards) to have the multimodal input processing module 23 process different types of input symptoms to generate a verbatim transcript of the condition. During the consultation, the physician records the interaction with the patient and uses different media (such as recording devices and keyboards) to have the multimodal input processing module 23 process the interaction records to generate a verbatim transcript of the diagnosis, diagnostic results, and other content.

[0044] The symptom summary processing module 24 uses a large language model to process the symptom text data generated by the multimodal input processing module 23 to generate a symptom summary. In one embodiment, the symptom summary processing module 24 integrates different LLMs to generate a fixed-format symptom text data (verbal transcript) into a symptom summary, as shown in Tables 3 and 4.

[0045] Table 3: Examples

[0046] Table 4: Example Data.

[0047] The diagnostic summary processing module 25 uses a large language model to process the diagnostic text data generated by the multimodal input processing module 23 to generate a diagnostic summary. In one embodiment, the diagnostic summary processing module 25 integrates different LLMs to generate a fixed-format diagnostic summary from the verbatim transcript of the diagnostic text data, as shown in Tables 5 and 6.

[0048] Table 5: Example Explanation.

[0049] Table 6: Example Data.

[0050] The personalized health education guidance generation module 26 retrieves corresponding health education content from the health education database 22 based on the chief complaint and past medical history in the symptom summary and the chief complaint and diagnosis in the diagnosis summary. Specifically, the module first uses a large-scale language model to determine at least one disease and its symptoms, and then retrieves corresponding health education document blocks from the health education database 22 based on these at least one disease and its symptoms. Each symptom corresponds to one health education document block to constitute the corresponding health education content. Thus, the personalized health education guidance generation module 26 uses a large-scale language model to process the chief complaint and past medical history in the symptom summary generated by the symptom summary processing module 24, the chief complaint and diagnosis in the diagnosis summary generated by the diagnosis summary processing module 25, and the retrieved corresponding health education content to generate personalized health education guidance.

[0051] In one embodiment, the personalized health education guidance generation module 26 obtains relevant information from the symptom summary and diagnosis summary, namely, the textual content of the chief complaint and past medical history in the symptom summary, and the chief complaint and diagnosis in the diagnosis summary. First, it uses LLM to understand and judge different diseases and symptoms (e.g., 1. hypertension 2. diabetes 3. obesity). Then, it performs vector transformation on each of the aforementioned different numbers of diseases, and simultaneously searches the health education database 22 to find matching health education document blocks. Each symptom corresponds to one health education document block. Next, using the "chief complaint and past medical history in the symptom summary," the "chief complaint and diagnosis in the diagnosis summary," and "all the aforementioned health education document blocks," the personalized health education guidance is generated through the LLM model.

[0052] In one embodiment, the personalized health education guidance generation module 26 can search the health education database 22 for different numbers of symptoms, and then use the retrieved health education content to compile and organize it to generate personalized health education materials, with special summary reminders for different aspects and possible conflicts, as shown in Tables 7 and 8.

[0053] Table 7: Health Education Guidelines for Each Disease (only excerpted for illustrative purposes).

[0054] Table 8: Compiled Health Education Guidelines

[0055] In addition, the large language model (LLM) used by the symptom summary processing module 24, the diagnosis summary processing module 25, and the personalized health education guidance generation module 26 is composed of artificial neural networks with multiple parameters and is trained using self-supervised learning or semi-supervised learning. The types and number of parameters of the large language model used by the symptom summary processing module 24, the diagnosis summary processing module 25, and the personalized health education guidance generation module 26 are different.

[0056] The hospital information system integration management module 27 can receive symptom summaries, diagnosis summaries, and personalized health education guidelines from the symptom summary processing module 24, the diagnosis summary processing module 25, and the personalized health education guide generation module 26, respectively. In one embodiment, the hospital information system integration management module 27 can uniformly handle the integration needs with hospital and clinic information systems. After the symptom summary processing module 24 completes the patient's condition summary, it is presented on the physician's computer or display medium (such as a tablet) through information system integration, providing the physician with a reference before the consultation. During the consultation, the physician can quickly ask questions on key issues, accelerating the consultation speed and improving quality. It can also be used through the integrated information system for the physician's diagnostic record of the patient's chief complaint, simplifying the physician's manual workflow. After the physician's diagnosis summary processing module 25 completes the diagnosis summary, it is provided to the physician for reference when writing the diagnosis record through information system integration, simplifying the physician's manual workflow. After the personalized health education guide generation module 26 completes the generation of health education guidelines, it is integrated into the hospital's workflow through the information system, providing reference for patients. For example, when patients pick up their medication, health education documents are printed out simultaneously.

[0057] In one embodiment, each module of the present invention can be software, hardware, or firmware; if it is hardware, it can be a processing unit, processor, or computer host with data processing and computing capabilities; if it is software or firmware, it can include instructions executable by a processing unit, processor, computer, or computer host, and can be installed on the same hardware device or distributed across different multiple hardware devices.

[0058] Therefore, this invention solves the problem of nurses having to manually select appropriate health education documents, which is time-consuming and labor-intensive. Moreover, compared with the single-sheet health education guidelines for a single disease in the prior art, the health education guidelines generated by this invention first divide the existing health education documents for different diseases into multiple blocks, and then combine and integrate them into personalized health education guidelines suitable for the first user (such as the patient or their family) based on the chief complaint and past medical history in the symptom summary of the first user (such as the patient or their family) and the chief complaint and diagnosis in the diagnosis summary of the second user (such as the doctor or nurse).

[0059] In summary, the system, method, and computer program product for automatically generating personalized health education guidelines of the present invention have the following advantages:

[0060] 1. Before visiting a hospital or clinic, patients can describe their reasons for seeking medical attention and symptoms through various input media (such as voice or text input). This invention can automatically generate a highly readable symptom summary in a fixed format. By integrating with the hospital or clinic's HIS system, doctors can quickly grasp the patient's condition through the summary before the consultation, thus accelerating the consultation and diagnosis process.

[0061] 2. The doctor's consultation process is recorded through different input media (such as voice consultation and doctor's text input). The invention automatically generates a highly readable diagnostic summary in a fixed format and completes the recording by integrating with the hospital's HIS system.

[0062] 3. Combining the patient's self-reported symptoms and physician's diagnosis, and based on the hospital's existing health education documents and manuals, the system automatically generates personalized health education guidelines for the patient. Furthermore, unlike the current system where only manual selection of suitable health education documents is possible, this system addresses the problem of separate documents that are difficult for patients to understand.

[0063] The above embodiments are merely illustrative of the effects of this application and are not intended to limit the scope of this application. Anyone skilled in the art can modify and alter the above embodiments without departing from the spirit and scope of this application. Therefore, the scope of protection of this application should be as set forth in the following patent application claims.

[0064] 21: Health Education Material Conversion Module 22: Health Education Resource Database 23: Multimodal Input Processing Module 24: Symptom Summary Processing Module 25: Diagnostic Summary Processing Module 26: Personalized Health Education Guidance Generation Module 27: Hospital Information System Integration and Management Module S11~S19: Steps

Claims

1. A method for automatically generating personalized health education guidelines, executed by a processor, the method comprising: Before diagnosis, the input symptoms received from the patient are processed to generate textual symptom data, and then a large language model is used to process the textual symptom data to generate a symptom summary. During the consultation, the system processes the consultation records received from the physician to generate diagnostic text data. This diagnostic text data is then processed using a large-scale language model to generate a diagnostic summary. Post-consultation, based on the chief complaint and past medical history in the symptom summary, and the chief complaint and diagnosis in the diagnostic summary, corresponding health education content is retrieved from the health education database. Furthermore, the system uses a large-scale language model to process the chief complaint and past medical history in the symptom summary, the chief complaint and diagnosis in the diagnostic summary, and the retrieved corresponding health education content to generate personalized health education guidance. The construction of this health education database includes: converting each health education document into a text format, dividing it into multiple health education document blocks according to the number of words, and then converting each health education document block into a vector and storing it in the health education database. The storage format structure of each health education document stored in the health education database includes a title, source, text content, blocks, and vector values. The process involves using a large-scale language model to determine the chief complaint and past medical history in the symptom summary, as well as the chief complaint and diagnosis in the diagnosis summary, to generate at least one disease and its symptoms. This is then converted into a vector representing at least one disease and its symptoms. This vector is then compared semantically with the corresponding health education document blocks in the health education database to retrieve at least one health education document block that corresponds to the at least one disease and its symptoms. Furthermore, the dietary control and exercise recommendations from this at least one health education document block are summarized and merged using the large-scale language model. A conflict analysis is then performed on the at least one disease and its symptoms, along with the dietary control and exercise recommendations from the at least one health education document block. This ensures that the personalized health education guidance includes consolidated dietary control, consolidated regular exercise, and reminders following the conflict analysis.

2. The method as described in claim 1, wherein, The symptom summary also includes the time of onset, medication status, concerns and requests, while the diagnostic summary includes medications, treatments and health education.

3. The method as described in claim 1, wherein, The large-scale language model used is composed of artificial neural networks with multiple parameters and is trained using self-supervised learning or semi-supervised learning. The types and number of parameters in the large-scale language model used vary.

4. A computer program product, which is loaded onto a computer to perform the method described in any one of claims 1 to 3.

5. A system for automatically generating personalized health education guidelines, comprising: The multimodal input processing module processes the input symptoms received from the patient before the consultation to generate symptom text data, and processes the consultation records received from the physician during the consultation to generate diagnostic text data. The symptom summary processing module processes the symptom text data generated by the multimodal input processing module using a large language model before the consultation to generate a symptom summary. The diagnostic summary processing module processes the diagnostic text data generated by the multimodal input processing module using a large language model during the consultation to generate a diagnostic summary. The health education data conversion module converts various health education documents into text format, then segments them into multiple health education document blocks based on word count. Each health education document block is then vectorized and stored in a health education database. The storage format structure of each health education document in the database includes a title, source, text content, blocks, and vector values. The personalized health education guidance generation module, after diagnosis, retrieves corresponding health education content from the database based on the chief complaint and past medical history in the symptom summary and the chief complaint and diagnosis in the diagnosis summary. A large-scale language model is then used to process the chief complaint and past medical history generated by the symptom summary processing module, the chief complaint and diagnosis generated by the diagnosis summary processing module, and the retrieved corresponding health education content to generate personalized health education guidance. The personalized health education guidance generation module utilizes a large-scale language model to judge the chief complaint and past medical history in the symptom summary and the chief complaint and diagnosis in the diagnosis summary to generate at least one disease and its symptoms. This is then converted into a vector of at least one disease and its symptoms. The vector of at least one disease and its symptoms is then compared semantically with the corresponding health education document blocks in the health education database to retrieve the health education document blocks that correspond to the at least one disease and its symptoms. The personalized health education guidance generation module uses a large-scale language model to summarize and merge the dietary control and exercise recommendations of the at least one health education document block to perform conflict analysis on the at least one disease and its symptoms and the dietary control and exercise recommendations of the at least one health education document block. This ensures that the personalized health education guidance includes the consolidated dietary control, the consolidated regular exercise, and the reminders after conflict analysis.

6. The system as described in claim 5, wherein, The symptom summary also includes the time of onset, medication status, concerns and requests, while the diagnostic summary includes medications, treatments and health education.

7. The system as described in claim 5 further includes a hospital information system integration and management module, wherein, The symptom summary processing module, the diagnosis summary processing module, and the personalized health education guidance generation module transmit the symptom summary, the diagnosis summary, and the personalized health education guidance to the hospital information system integration and management module, respectively.

8. The system as described in claim 5, wherein, The large language models used by the symptom summary processing module, the diagnosis summary processing module, and the personalized health education guidance generation module are composed of artificial neural networks with multiple parameters and are trained using self-supervised learning or semi-supervised learning. The types and number of parameters of the large language models used by the symptom summary processing module, the diagnosis summary processing module, and the personalized health education guidance generation module are different.

Images