Health propaganda and education pushing method, system and equipment based on artificial intelligence and medium

Abstract

The invention discloses a health propaganda and education pushing method, system and device based on artificial intelligence and a medium, and the method comprises the steps: collecting the whole-course doctor seeing process data of a patient based on a hospital information system, and recognizing a patient propaganda and education key node through a preset rule engine; generating a structured propaganda and education scheme based on a large language model on the basis of input patient historical medical records, current diagnosis results and propaganda and education knowledge base data; based on a multi-modal generation model, converting the structured propaganda and education scheme into propaganda and education contents in audio, video or image-text forms; a push channel is connected through an API interface, and the propaganda and education content and the check link are pushed to the patient terminal based on the patient reserved contact information; and carrying out statistics on push-related data of the propaganda and education contents, collecting feedback information of the patient for the propaganda and education contents, and optimizing and updating a preset rule engine, a propaganda and education knowledge base, a large language model and a multi-modal generation model. Dynamic enhancement and real-time generation of propaganda and education contents can be realized by using an artificial intelligence technology.

Description

Technical Field

[0001] This invention relates to the field of medical and nursing technology, and in particular to a method, system, device and medium for health education delivery based on artificial intelligence. Background Technology

[0002] In the field of healthcare, health education is a crucial link in improving patient treatment adherence and rehabilitation outcomes, and its level of intelligence and precision directly impacts the quality of medical services. Traditional health education systems mainly rely on fixed knowledge rules to generate educational content, employing single formats such as oral explanations, brochure distribution, or static video playback. This results in the delivery of homogenized information to patients with similar symptoms, leading to several prominent problems: On the one hand, content updates rely on manual maintenance, resulting in poor timeliness and a fixed, monotonous format, making it difficult to adapt to individual patient differences (such as age, education level, and disease progression). On the other hand, it is impossible to dynamically adjust the education program based on key nodes throughout the patient's entire disease course (such as pre-diagnosis appointments, in-diagnosis and post-operative care, and post-diagnosis rehabilitation), leading to insufficient targeting, low patient learning rates, and a lack of real-time feedback mechanisms, making it difficult to optimize education strategies in a closed loop and failing to meet the needs of modern medicine for precise and systematic health management.

[0003] While existing patented technologies attempt to optimize health education models to some extent, they still have significant shortcomings and deficiencies, such as: Zhejiang Lab proposed a method and system for delivering patient education information based on multi-source information fusion (CN114969557B). The method includes: Step S1: Constructing a patient education knowledge graph and pushing it to patients via an education mini-program; Step S2: Fuding and correcting the patient's basic information, medical information, eye movement information, and personality scale to obtain multi-source patient information; Step S3: Using the multi-source patient information and collected patient medication behavior data, constructing a compliance prediction model through a neural network; Step S5: Establishing a system rule base, searching for corresponding diseases and treatments in the patient education knowledge graph based on information from the system rule base, and then pushing them to patients via the education mini-program.

[0004] CN114969557B focuses on multi-source information fusion and adherence prediction, but it has limitations: First, the patient education knowledge graph relies on public knowledge, expert supplements, and electronic medical records, and its dynamic updating capability is insufficient, making it difficult to incorporate the latest clinical guidelines or patient condition changes in real time; second, the push strategy is mainly based on eye-tracking information and personality scales to build a rule base, which does not fully cover key nodes in the entire course of diagnosis and treatment (such as preoperative preparation and discharge medication guidance), and has weak adaptability to the dynamic course of the disease; third, the multimodal content is stored in a pre-set format and lacks the ability to dynamically generate content based on the real-time needs of patients, resulting in insufficient flexibility.

[0005] CN119993507A discloses an AI-based digital one-stop health education platform and its construction method, including: a health data management module for collecting user health data from multiple data sources and automatically generating user health records based on user-input personal health information; and a health profile generation module for generating personalized health profiles based on health records and conducting personalized needs analysis based on user-defined goals. While emphasizing health data management and trend prediction, it has shortcomings: First, the platform leans towards general health management and is insufficiently adapted to the specific needs of patients in medical scenarios (such as postoperative complication prevention and chronic disease prescription adjustments), resulting in weak medical professionalism and disease-specificity in the educational content; second, the educational content relies on deep learning algorithms for recommendation and does not employ large language models to dynamically generate structured solutions, making it difficult to customize personalized content in real time based on the patient's specific diagnosis; third, the feedback optimization mechanism mainly targets adjustments to recommendation strategies and lacks deep linkage between the educational knowledge base and feedback data, hindering automatic iterative updates of knowledge content.

[0006] CN120492714A discloses a method, robot, device, and medium for recommending nursing health education content, relating to the field of health education content recommendation technology. The method includes: collecting user characteristic information to be educated; matching and identifying education content based on the user characteristic information to obtain recommended education content, wherein the matching degree is analyzed using an intelligent model for recommendation; displaying the recommended education content and collecting user feedback; and updating the intelligent model based on the feedback. While it collects user characteristics through voice and facial information, it has several drawbacks: First, user characteristic collection is limited to voice questions and facial information, failing to fully integrate patient's entire course of medical data (such as examination reports, surgical records, and medication history), resulting in limited matching accuracy due to the single feature dimension; second, the education content originates from a pre-set database and relies on the intelligent model to analyze the matching degree for recommendation, lacking dynamic generation capabilities and unable to address new education needs arising from changes in the patient's condition; third, feedback updates only target the education matching and recognition device, limiting long-term optimization effects.

[0007] In summary, existing technologies have not yet solved the core problems in health education systems, such as insufficient dynamic knowledge updates, poor adaptability to all disease stages, weak personalized content generation capabilities, and incomplete feedback loops. There is an urgent need to build an intelligent health education solution that can adapt to the entire disease process, dynamically generate personalized content, and continuously optimize through real-time feedback. Summary of the Invention

[0008] To address the aforementioned issues, this invention proposes a health education delivery method, system, device, and medium based on artificial intelligence. By utilizing artificial intelligence technology, it achieves dynamic enhancement and real-time generation of education content, overcoming the limitations of traditional systems that rely on pre-prepared content and improving the personalization and timeliness of education.

[0009] The technical solution adopted in this invention is as follows: A health education delivery method based on artificial intelligence includes: Based on the hospital information system, data on the entire course of a patient's medical treatment is collected, and key nodes in patient education are identified through a preset rule engine. Based on the input patient's historical medical records, current diagnosis results, and health education knowledge base data, a structured health education plan is generated using a large language model. Based on a multimodal generation model, the structured education and outreach program is converted into educational content in audio, video, or text and image formats. By connecting to the push channel through the API interface, the educational content and viewing link are pushed to the patient's terminal based on the patient's reserved contact information; The system collects statistical data on the push of educational content, gathers patient feedback on the content, and optimizes and updates the preset rule engine, educational knowledge base, large language model, and multimodal generation model.

[0010] Furthermore, the step of identifying key patient education nodes through a preset rule engine includes: mining the time-series characteristics of patient visits through a time-series analysis algorithm, and combining them with preset medical and nursing path rules to identify education nodes under pre-visit appointments, in-visit examinations, post-operative care, discharge rehabilitation, prescription medication, and non-routine medical procedures.

[0011] Furthermore, based on the input patient's historical medical records, current diagnosis results, and education knowledge base data, a structured education plan is generated using a large language reasoning model. This includes: combining the patient's historical medical records, current diagnosis results, and education knowledge base data, establishing structured relationships through knowledge graph technology, and updating education-related data through knowledge enhancement technology.

[0012] Furthermore, the multimodal generation model includes a text-to-speech model, a text-to-image generation model, and a video synthesis model, and the generation of the educational content is based on the patient's individual preferences and / or physiological characteristics to adapt the content type.

[0013] Furthermore, the statistical data related to the push of propaganda content includes: monitoring the status of propaganda content after it is pushed; when it is detected that the content is not opened after delivery or is opened but not viewed, adjusting the push channel or push time, and repeatedly pushing the corresponding propaganda content at intervals; for preset high-priority propaganda content, adopting a multi-channel simultaneous push mode.

[0014] Furthermore, the system collects data related to the delivery of educational content and gathers patient feedback on the content, including: the delivery rate, link opening rate, and completion rate of the educational content; and patient ratings and opinions on the educational content.

[0015] Furthermore, the health education knowledge base is optimized and updated, including: inputting collected patient feedback information and updated clinical practice data into the health education knowledge base through a knowledge enhancement algorithm, and updating the nursing guidelines and precautions content in the health education knowledge base.

[0016] An AI-based health education delivery system includes: The key node identification module is configured to collect patient's entire medical process data based on the hospital information system and identify key nodes in patient education through a preset rule engine. The education program generation module is configured to generate a structured education program based on a large language model, based on the input patient's historical medical records, current diagnosis results, and education knowledge base data. The multimodal content generation module is configured to convert the structured education scheme into education content in audio, video, or graphic form based on a multimodal generation model. The intelligent push execution module is configured to connect to the push channel via API interface and push the educational content and viewing link to the patient's terminal based on the patient's reserved contact information; The effect feedback statistics module is configured to collect statistics on data related to the push of educational content, collect patient feedback on the educational content, and optimize and update the preset rule engine, educational knowledge base, large language model and multimodal generation model.

[0017] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the AI-based health education push method.

[0018] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the AI-based health education push method.

[0019] The beneficial effects of this invention are as follows: 1. This invention is based on the dynamic identification needs of key nodes throughout the patient's disease course (pre-diagnosis appointment, in-diagnosis treatment, post-operative rehabilitation, etc.), integrates multi-source information such as historical medical records, diagnostic results, and behavioral data, and generates education programs that are tailored to individual conditions and preferences through a large language model. It covers both inpatients and non-inpatient long-term treatment groups such as those with chronic diseases, and completely breaks away from the limitations of homogeneous static content.

[0020] 2. This invention can automatically convert structured educational programs into multimodal content such as text, audio, video, and graphics, adapting to the acceptance habits of patients of different ages and educational levels; it can also push content through multiple channels such as WeChat, SMS, and bedside devices, combined with strategies such as interval repetitive push and simultaneous push of high-priority content across multiple channels, significantly improving the content delivery rate, opening rate, and learning rate.

[0021] 3. This invention collects push effect data and patient feedback in real time, and optimizes the rule engine, education knowledge base and AI model in reverse to form a closed loop of the whole link, which continuously improves the accuracy and adaptability of education content.

[0022] 4. This invention replaces traditional manual lectures and manual distribution with automated content generation, push and management, reducing repetitive work for medical staff; it enables efficient application of medical knowledge by using knowledge graphs and intelligent search technology, reducing manpower and time costs for education and allowing medical resources to be more focused on core diagnosis and treatment services. Attached Figure Description

[0023] Figure 1 This is a flowchart of a health education and promotion method based on artificial intelligence, according to Embodiment 1 of the present invention.

[0024] Figure 2 This is a schematic diagram of a health education and promotion system based on artificial intelligence, according to Embodiment 2 of the present invention. Detailed Implementation

[0025] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments are now described. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention; that is, the described embodiments are only a part of the embodiments of the invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0026] Example 1 like Figure 1 As shown, this embodiment provides a health education push method based on artificial intelligence, including: Based on the hospital information system, data on the entire course of a patient's medical treatment is collected, and key nodes in patient education are identified through a preset rule engine. Based on the input patient's historical medical records, current diagnosis results, and health education knowledge base data, a structured health education plan is generated using a large language model. Based on a multimodal generative model, structured educational programs are converted into educational content in audio, video, or text and image formats. By connecting to the push channel through the API interface, the educational content and viewing link are pushed to the patient's terminal based on the patient's reserved contact information; We collect data on the push of educational content, gather feedback from patients regarding the content, and optimize and update the preset rule engine, educational knowledge base, large language model, and multimodal generation model.

[0027] It should be noted that this method, through fully automated processing, enables closed-loop management of health education from node identification, plan generation, content conversion to push feedback, significantly improving the efficiency of education work and reducing the cost of manual intervention by medical staff; the education content generated based on individual patient medical data is more targeted, effectively solving the problem of homogenization in traditional education content; the multimodal content format and multi-channel push method improve the reach of education content and patients' willingness to accept it; the continuous optimization and update mechanism can constantly adapt to changes in medical scenarios and upgrades in patient needs, maintaining good education results in the long term.

[0028] Preferably, when identifying key patient education nodes through a preset rule engine, a time-series analysis algorithm is first used to deeply mine the patient's medical behavior data, extracting time-series features to clarify the sequence of patient visits, the duration of each step, and the relationships between different steps. Simultaneously, the mined time-series features are combined with preset medical and nursing pathway rules, which cover the health management requirements and education needs at each stage of the routine medical process.

[0029] Specifically, for routine medical procedures such as pre-diagnosis appointments, in-diagnosis examinations, post-operative care, discharge rehabilitation, and prescription medication, the rule engine accurately identifies key nodes in each procedure that require education based on time series characteristics and medical care pathway rules. For non-routine medical procedures, such as emergency visits and treatment adjustments after sudden changes in the patient's condition, the rule engine analyzes the special nature and urgency of the medical behavior and combines it with corresponding emergency nursing guidelines to identify key education nodes in special scenarios, ensuring that the education needs of patients throughout the entire course of their illness are covered without omission.

[0030] It should be noted that by combining time-series analysis algorithms with medical and nursing pathway rules, dynamic tracking and accurate judgment of the patient's medical process can be achieved. This not only identifies the education nodes in the routine medical process but also adapts to the special needs of unconventional medical scenarios, significantly improving the comprehensiveness and accuracy of identifying key education nodes. The node identification process strictly follows medical and nursing standards, ensuring the synchronization and adaptability of education work with the treatment process. This lays a solid foundation for the generation of personalized education plans and avoids the problem of poor results caused by inappropriate timing of education.

[0031] Preferably, when generating a structured education plan, the following steps are taken: First, individual data from the patient's medical history, including past medical history, treatment history, and medication history; core information from the current diagnosis, such as the diagnosis and treatment recommendations; and general knowledge from the education knowledge base, such as medical common sense, nursing guidelines, and rehabilitation plans. This multi-source data is then input into a knowledge graph technology module. The knowledge graph technology establishes relationships between different data sources, such as the relationship between the patient's medical history and current condition, and the relationship between the current diagnosis and the corresponding nursing plan, forming a structured data network.

[0032] Meanwhile, this embodiment introduces knowledge enhancement technology to capture the latest research results in the medical field, updated clinical practice content, and newly added knowledge in the education knowledge base in real time, supplementing and updating the established data associations to ensure the timeliness and completeness of the data. The data, after being associated with the knowledge graph and updated with knowledge enhancement, is input into the large language model. The large language model generates a standardized education plan that includes modules such as education objectives, core content, implementation suggestions, and precautions, according to preset structured format requirements.

[0033] It should be noted that by establishing structured connections between multi-source data through knowledge graph technology, the generated education plans can fully integrate individual patient circumstances with general medical knowledge, significantly improving the relevance and scientific rigor of the plans. The application of knowledge augmentation technology enables dynamic updates of education-related data, ensuring that the generated education plans keep pace with changes in medical development and clinical practice, avoiding inaccuracies caused by outdated knowledge. The structured plan format makes the education content clear and organized, facilitating patient understanding and providing a standardized framework for subsequent multimodal content conversion.

[0034] Preferably, the multimodal generation model includes core sub-models such as a text-to-speech model, an image-to-text generation model, and a video synthesis model. These sub-models work together to convert and generate educational content. Before generating the educational content, the model obtains the patient's interests, preferences, and information receiving habits through channels such as the patient's historical interaction data and individual information registration. At the same time, it collects physiological characteristics that may affect information reception, such as the patient's vision, hearing, and education level.

[0035] Preferably, the presentation format of the educational content is adapted to the patient's preferences and physiological characteristics: for patients who prefer auditory reception or have poor vision, the text-to-speech model converts the text content in the structured educational plan into clear and fluent audio content; for educational content that prefers visual reception or requires intuitive display, the image-text generation model transforms core knowledge into richly illustrated materials, while the video synthesis model creates vivid and engaging video content for complex treatment procedures, rehabilitation movements, and other content. Through these methods, multimodal educational content tailored to the individual circumstances of different patients is customized.

[0036] It should be noted that the diversified design of the multimodal generation model breaks through the limitations of traditional single-text education, providing patients with richer information reception options; the content adaptation mechanism based on individual patient preferences and physiological characteristics ensures that the education content can be better accepted and understood by patients, effectively improving patients' interest in learning and mastery of the education knowledge; personalized content adaptation for patients with special physiological conditions reflects the humanistic care of health education, expands the coverage of education work, and improves the overall effectiveness of education work.

[0037] Preferably, after the educational content is pushed out, the push status is monitored in real time. Through data interaction with the push channel, key information such as the delivery status of the educational content, the reception status of the patient's terminal, and whether the patient has opened and viewed the content is obtained. When it is detected that the educational content has been delivered but the patient has not opened it, or has opened it but has not viewed it completely, an adjustment mechanism is activated. Based on data such as the patient's daily active time and preferred receiving channels, a suitable push time and push channel are reselected, and the corresponding educational content is pushed out repeatedly at intervals to avoid frequent pushes that may cause patient resentment.

[0038] Preferably, for pre-set high-priority educational content, such as postoperative emergency care precautions, knowledge related to warnings of disease deterioration, and preparation requirements before important examinations, the system adopts a multi-channel simultaneous push mode, simultaneously pushing information to patients through multiple channels such as SMS, WeChat mini-programs, and hospital bedside devices, ensuring that patients can receive and view key educational information in a timely manner.

[0039] It should be noted that by monitoring and dynamically adjusting the push status in real time, the problem of the push content stopping immediately has been effectively solved, significantly improving the actual delivery rate and complete viewing rate of the content. The intermittent repetitive push mechanism ensures the effectiveness of the education while taking into account the user experience of patients and avoiding excessive disturbance. The multi-channel simultaneous push mode of high-priority content ensures that key health information can reach patients in a timely manner, providing an important guarantee for patients' disease management and treatment cooperation, and reducing the health risks caused by information omissions.

[0040] Preferably, the data statistics module comprehensively analyzes the data related to the effectiveness of the push of the educational content, including key indicators such as the delivery rate of the educational content (the proportion of the number of successfully delivered content to the patient's terminal out of the total number of pushes), the link opening rate (the proportion of the number of patients clicking to view the link out of the number of successfully delivered content), and the viewing completion rate (the proportion of the number of patients who fully view the educational content out of the number of clicks to view content), to form an intuitive push effect data report.

[0041] Preferably, after the patient has finished viewing the educational content, the system collects the patient's rating of the educational content by popping up a feedback form or setting an evaluation button on the patient's terminal. The rating dimensions may include the practicality, comprehensibility, and relevance of the content. In addition, the system provides an entry point for patients to express their opinions, allowing them to input textual feedback, such as supplementary needs for the educational content or suggestions for expression, thus comprehensively collecting the patient's subjective feedback.

[0042] It should be noted that comprehensive push notification effectiveness statistics enable medical staff and system administrators to clearly understand the actual implementation of the education work, accurately judge the reach of the education content and the level of patient participation; the collection of patient ratings and written opinions directly reflects patients' satisfaction with the education content and their actual needs, providing a real and reliable basis for subsequent education program optimization, content updates and push strategy adjustments, and promoting continuous improvement of education work.

[0043] Preferably, the collected patient feedback information is categorized and organized, and valuable content is extracted, such as knowledge points not covered by the patient, unclear parts of the content, and health knowledge that the patient wishes to supplement; at the same time, updated data in clinical practice is collected, including new diagnosis and treatment technologies, adjustments to nursing standards, updates to disease prevention and control points, and other medical-related content.

[0044] Preferably, the aforementioned patient feedback information and updated clinical practice data are input into a knowledge enhancement algorithm. This algorithm filters, integrates, and structures the data, extracting core knowledge points. The processed knowledge data is then input into an educational knowledge base. The educational knowledge base, according to preset update rules, supplements, modifies, and improves relevant content such as nursing guidelines, health precautions, and disease rehabilitation knowledge, ensuring that the content in the educational knowledge base can respond in real-time to patient needs and changes in clinical practice.

[0045] It should be noted that by combining patient feedback with clinical practice data, the health education knowledge base is dynamically updated, enabling it to continuously absorb new knowledge and avoid becoming outdated. The knowledge enhancement algorithm-based update method ensures the accuracy, structure, and practicality of the new knowledge, improving the quality of the health education knowledge base. The continuous optimization of the knowledge base provides richer and more accurate knowledge support for the generation of subsequent health education programs, ensuring the professionalism and effectiveness of health education from the source.

[0046] Example 2 like Figure 2 As shown, this embodiment provides a health education and outreach system based on artificial intelligence, including: The key node identification module is configured to collect patient's entire medical process data based on the hospital information system and identify key nodes in patient education through a preset rule engine. The education program generation module is configured to generate a structured education program based on a large language model, based on the input patient's historical medical records, current diagnosis results, and education knowledge base data. The multimodal content generation module is configured to convert structured missionary programs into missionary content in audio, video, or text and graphics formats based on a multimodal generation model. The intelligent push execution module is configured to connect to the push channel via API interface and push the educational content and viewing link to the patient's terminal based on the patient's reserved contact information; The effect feedback statistics module is configured to collect statistics on data related to the push of educational content, collect patient feedback on the educational content, and optimize and update the preset rule engine, educational knowledge base, large language model and multimodal generation model.

[0047] Preferably, the key node identification module establishes a stable data transmission channel with the hospital information system to collect real-time data on the entire patient's medical process, including data from each stage such as registration, examinations, treatment procedures, and discharge settlement. The module is equipped with a pre-set rule engine. Based on pre-set logic such as medical and nursing pathways and disease management guidelines, the rule engine analyzes and processes the collected data to accurately identify the key nodes in the patient's medical history where health education is needed, and synchronizes the identification results to the health education plan generation module.

[0048] Preferably, the education plan generation module receives node information output by the key node identification module, while simultaneously collecting individual medical data such as the patient's historical medical records and current diagnosis results, and retrieving relevant medical knowledge from the education knowledge base. The module establishes connections between multi-source data using knowledge graph technology, updates data content using knowledge enhancement technology, and inputs the processed comprehensive data into a large language model to generate a structured education plan adapted to the patient's current condition and education nodes.

[0049] Preferably, the multimodal content generation module can obtain the structured education plan output by the education plan generation module, select the corresponding multimodal generation sub-model (text-to-speech, image-text generation, video synthesis, etc.) according to the content characteristics of the plan and the individual characteristics of the patient, convert the structured plan into diversified education content, and perform format standardization processing on the generated content to ensure that it is compatible with different push channels.

[0050] Preferably, the intelligent push execution module establishes connections with various push channels such as SMS, WeChat mini-programs, and bedside smart devices through API interfaces, extracts the patient's reserved contact information and preferred push methods, accurately pushes the standardized educational content and viewing links output by the multimodal content generation module to the patient's terminal, and provides real-time feedback on the push status to the effect feedback statistics module.

[0051] Preferably, the effect feedback statistics module receives push status data from the intelligent push execution module in real time, and calculates key indicators such as delivery rate, open rate, and view completion rate. Simultaneously, it collects patient ratings and feedback on the educational content through patient terminals. The module summarizes and analyzes the collected data, sending optimization instructions to the key node identification module, educational plan generation module, multimodal content generation module, and educational knowledge base to adjust the operating parameters and data content of each module, achieving closed-loop optimization of the system.

[0052] It should be noted that the above modules have clear division of labor and work together to build a complete intelligent health education system. This system automates the entire process from node identification, plan generation, content conversion, precise push notifications to feedback optimization, significantly reducing the workload of medical staff. Real-time data interaction between modules ensures the timeliness and continuity of education, enabling the content to be accurately adapted to the patient's disease progression and individual needs. The closed-loop optimization mechanism allows the system to continuously iterate and upgrade based on actual operating results, constantly improving the accuracy, effectiveness, and patient satisfaction of education, providing strong support for the patient's full-course health management.

[0053] Example 3 This embodiment is based on embodiment 1: This embodiment provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the AI-based health education push method of Embodiment 1. The computer program can be in the form of source code, object code, executable file, or some intermediate form.

[0054] Example 4 This embodiment is based on embodiment 1: This embodiment provides a computer-readable storage medium storing a computer program. When executed by a processor, this computer program implements the AI-based health education push method of Embodiment 1. The computer program can be in the form of source code, object code, executable file, or some intermediate form. The storage medium includes: any entity or device capable of carrying computer program code, recording media, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0055] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the concept described herein through the above teachings or related technologies or knowledge. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

[0056] It should be noted that, for the sake of simplicity, the foregoing method embodiments are described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

Claims

1. A health education delivery method based on artificial intelligence, characterized in that, include: Based on the hospital information system, data on the entire course of a patient's medical treatment is collected, and key nodes in patient education are identified through a preset rule engine. Based on the input patient's historical medical records, current diagnosis results, and health education knowledge base data, a structured health education plan is generated using a large language model. Based on a multimodal generation model, the structured education and outreach program is converted into educational content in audio, video, or text and image formats. By connecting to the push channel through the API interface, the educational content and viewing link are pushed to the patient's terminal based on the patient's reserved contact information; The system collects statistical data on the push of educational content, gathers patient feedback on the content, and optimizes and updates the preset rule engine, educational knowledge base, large language model, and multimodal generation model.

2. The health education delivery method based on artificial intelligence according to claim 1, characterized in that, The process of identifying key patient education nodes through a preset rule engine includes: mining the time-series characteristics of patient visits through a time-series analysis algorithm, and combining them with preset medical and nursing path rules to identify education nodes under pre-visit appointments, in-visit examinations, post-operative care, discharge rehabilitation, prescription medication, and non-routine medical procedures.

3. The health education delivery method based on artificial intelligence according to claim 1, characterized in that, Based on the input patient's historical medical records, current diagnosis results, and education knowledge base data, a structured education plan is generated using a large language reasoning model. This includes: combining the patient's historical medical records, current diagnosis results, and education knowledge base data, establishing structured relationships through knowledge graph technology, and updating education-related data through knowledge enhancement technology.

4. The health education delivery method based on artificial intelligence according to claim 1, characterized in that, The multimodal generation model includes a text-to-speech model, a text-to-image generation model, and a video synthesis model. The generation of the educational content is based on the patient's individual preferences and / or physiological characteristics to adapt the content type.

5. The health education delivery method based on artificial intelligence according to claim 1, characterized in that, The statistical data related to the push of propaganda content includes: monitoring the status of propaganda content after it is pushed; when it is detected that the content is not opened after delivery or is opened but not viewed, the push channel or push time is adjusted, and the corresponding propaganda content is pushed repeatedly at intervals; for preset high-priority propaganda content, a multi-channel simultaneous push mode is adopted.

6. The health education delivery method based on artificial intelligence according to claim 1, characterized in that, The system collects data related to the push of educational content and gathers patient feedback on the content, including: the delivery rate, link opening rate, and completion rate of the educational content; and patient ratings and opinions on the educational content.

7. The health education delivery method based on artificial intelligence according to claim 1, characterized in that, The education knowledge base is optimized and updated by: inputting collected patient feedback information and updated clinical practice data into the education knowledge base through a knowledge enhancement algorithm, and updating the nursing guidelines and precautions in the education knowledge base.

8. A health education and outreach system based on artificial intelligence, characterized in that, include: The key node identification module is configured to collect patient's entire medical process data based on the hospital information system and identify key nodes in patient education through a preset rule engine. The education program generation module is configured to generate a structured education program based on a large language model, based on the input patient's historical medical records, current diagnosis results, and education knowledge base data. The multimodal content generation module is configured to convert the structured education scheme into education content in audio, video, or graphic form based on a multimodal generation model. The intelligent push execution module is configured to connect to the push channel via API interface and push the educational content and viewing link to the patient's terminal based on the patient's reserved contact information; The effect feedback statistics module is configured to collect statistics on data related to the push of educational content, collect patient feedback on the educational content, and optimize and update the preset rule engine, educational knowledge base, large language model and multimodal generation model.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the health education push method based on artificial intelligence as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the health education push method based on artificial intelligence as described in any one of claims 1-7.

Images