An evidence-based coronary heart disease post-intervention health education AI question and answer robot system

Abstract

This invention relates to an evidence-based AI-powered question-and-answer robot system for post-PCI (percutaneous coronary intervention) health education, belonging to the field of medical artificial intelligence. It aims to address the challenges of traditional health education after PCI patients are discharged, including difficulties in maintaining continuity of care, insufficient content standardization, and inability to meet personalized, immediate consultation needs. The system comprises: an evidence-based knowledge base construction module, which forms a structured knowledge base and stores it quantitatively through systematic literature retrieval, expert consultation, and evidence quality evaluation; an intelligent question-answering core engine module, which utilizes Retrieval Enhanced Generation (RAG) technology, employs the Zhipu embedding-v3 model for semantic retrieval, and calls upon a large language model to generate answers, ensuring professionalism and accuracy; a multi-terminal adaptation and deployment module, supporting WeChat mini-programs and third-party smart voice hardware to achieve lightweight services and pure voice interaction; and a closed-loop optimization and learning module, which continuously updates the knowledge base and iterates the system based on user feedback and nursing staff annotations.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology in medical and health care, and in particular to an intelligent question-and-answer robot system based on evidence-based medicine and fusion retrieval augmented generation (RAG) technology for health education of patients after coronary intervention. Background Technology

[0002] Percutaneous coronary intervention (PCI) is one of the main treatments for coronary artery disease, but long-term management of patients after the procedure, including lifestyle modifications, medication adherence, and symptom monitoring, is crucial for prognosis. Currently, traditional health education models are limited by time and space, making it difficult to continue after discharge. While existing methods based on mobile apps and WeChat platforms offer convenience, they generally suffer from insufficient standardization of health education content, difficulty in ensuring professionalism, and inability to accurately address individual patient needs. This is especially true for elderly patients or those with limited information access, who face challenges such as complex interfaces, difficulty in text input, and questionable reliability of information. Currently, there is a lack of intelligent health education solutions that can provide timely, accurate, and personalized care, particularly suitable for post-operative home rehabilitation. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art and provide an evidence-based AI question-and-answer robot system for post-coronary interventional health education. This system ensures the professionalism and accuracy of the answers by constructing an evidence-based knowledge base and integrating advanced RAG technology. At the same time, it supports multimodal interaction and multi-terminal deployment, effectively solving the needs of post-operative patients for timely, homogeneous and personalized health education. To achieve the above objectives, the present invention adopts the following technical solution: An evidence-based AI question-and-answer robot system for post-coronary interventional health education, characterized in that it includes the following modules: 1. Evidence-based knowledge base construction module: used to systematically collect, screen, evaluate and integrate authoritative evidence for post-coronary interventional health education to form a structured knowledge base; this module includes: (1) Literature retrieval and screening unit: based on a preset retrieval strategy, automatically or semi-automatically retrieves relevant literature from domestic and foreign guideline databases, literature databases and professional websites. (2) Evidence quality assessment and standardization unit: The quality of the included literature is assessed using standard tools such as AGREE II and JBI, and the feasibility, appropriateness, clinical significance and effectiveness of the evidence are assessed through expert consultation (such as Fame evaluation) to form strong recommendation (level A) and weak recommendation (level B) evidence. (3) Knowledge transformation and storage unit: The evaluated evidence is transformed into standardized knowledge fragments that are easy for machines to understand and retrieve, and then vectorized and stored. 2. Intelligent question answering core engine module: Based on retrieval augmented generation (RAG) technology, it is used to understand user questions and generate accurate answers; this module includes: (1) User interaction interface unit: Supports two input methods, text and voice, and two output methods, text and speech synthesis (TTS). (2) Semantic understanding and retrieval unit: The user input questions are vectorized and encoded, and semantic similarity is retrieved in the vector database to recall the most relevant knowledge fragments. (3) Answer generation and optimization unit: Using a large language model, the retrieved knowledge fragments are combined with user questions to generate natural, fluent and professional answers; when the question exceeds the scope of the knowledge base, a unified prompt message is returned. 3. Multi-terminal adaptation and deployment module: Used to deploy the core functions of the system to different user terminals; This module supports: (1) Lightweight front-end deployment of WeChat mini program, which is convenient for patients to access via mobile phone. (2) Smart hardware expansion deployment, which adapts to third-party smart voice hardware through cloud voice service to achieve pure voice interaction and meet the needs of elderly or patients with inconvenient operation. 4. Closed-loop optimization and learning module: Used for continuous improvement of the system; This module includes: (1) Feedback collection unit: Records user interaction data and provides a management backend for nursing staff to mark the quality of questions and answers. (2) Knowledge base update unit: Supports manual or semi-automatic updating of knowledge base content according to the latest clinical guidelines or expert opinions. Further, in the evidence-based knowledge base construction module, the knowledge base content covers three core question clusters: basic disease knowledge, lifestyle management, and disease symptom management.Furthermore, in the intelligent question-answering core engine module, the Zhipu embedding-v3 model is used for text vectorization, and Deep seek v3 is used as the generation model. Furthermore, in the multi-terminal adaptation and deployment module, the intelligent hardware expansion deployment scheme constructs a cloud-based voice service middle layer, receives the audio stream from the intelligent hardware, performs speech-to-text (STT), calls the core question-answering API, and performs text-to-speech (TTS) in sequence, and returns the voice stream to the hardware for playback, thereby achieving decoupling between the hardware and the core service. Furthermore, the user interaction interface unit is optimized for elderly users, including supporting large font display, simplifying the operation process, and enhancing the voice interaction function. After adopting the above technical solutions, the beneficial effects of this invention are: 1. Patient level: (1) Improve patients' health knowledge level: The application of AI question-answering robots helps patients understand postoperative lifestyle adjustments, disease symptom management, and other knowledge. Personalized and accurate health education information is provided for patients after coronary intervention. Based on evidence transformation knowledge and artificial intelligence technology, it provides comprehensive and accurate health education knowledge after coronary intervention through interaction. (2) Enhance patient satisfaction: AI question-answering robots provide humanized health education services and improve the medical experience. The robot can provide regular follow-up and reminder services to ensure that patients have timely follow-up visits and medication, and reduce the risk of complications. 2. Medical care level: (1) Reduce the workload of medical staff: AI question-answering robots can undertake some health education tasks, reducing the workload of medical staff. The robot provides standardized health education content to achieve homogeneous education. (2) Improve the efficiency of health education: AI question-answering robots can quickly respond to patients' education needs, provide timely and accurate information, and improve the efficiency and coverage of health education. Based on patients' feedback, optimize the education content and methods to improve the pertinence and effectiveness of education. Attached Figure Description

[0004] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Figure 1 This is a schematic diagram of the system interface display status in this invention. Detailed Implementation

[0005] See Figure 1As shown, the technical solution adopted in this specific implementation is to provide a multi-channel and convenient interactive experience, especially considering the usage habits of postoperative patients and elderly users. Its specific usage methods are mainly divided into the following two types: 1. Scanning the QR code to enter the WeChat mini-program: This allows for text, image, and voice interaction. This method is suitable for patients who are proficient in using smartphones, providing a private and flexible interaction approach. Entry and activation: Obtaining the QR code: When patients are hospitalized or discharged from the cardiology department, medical staff will provide them with a promotional leaflet or card printed with a WeChat mini-program QR code. Figure 2 This is the QR code for the mini-program. Scan to enter: Patients can use the "Scan" function of WeChat on their mobile phones to scan this QR code to directly access the AI ​​Q&A robot mini-program interface, without needing to download or install any additional applications. 1. See [link to interactive question-asking method] Figure 3 (1) Text question (main method): After entering the mini program, there will be a clear input box on the interface. Patients can directly type their questions in the input box, such as: "What fruits can I eat after surgery?" or "What should I do if I feel chest tightness at night?", and then click the "send" button. (2) Voice question (convenient method): To facilitate input, the mini program interface also has a "microphone" icon button. Patients only need to press and hold this button to start recording and speak their questions, such as "How should I exercise?". After releasing the button, the system will automatically convert the speech into text and send it. 2. Obtaining and understanding the answer. (1) Multi-form feedback: After receiving the question, the system will generate the answer within 1-2 seconds based on the evidence-based knowledge base. Text display: The answer will be displayed in clear text form on the dialogue interface. (2) Voice broadcast: At the same time, the system will read the answer in clear and natural voice through text synthesis (TTS) technology, which is especially friendly to users with poor eyesight or who want to reduce the burden of reading. (3) Interface design: The mini program interface has been optimized for the elderly, such as Figure 1 As shown, it features large fonts, high-contrast colors, and a simple button layout to ensure easy operation and reading for users. II. Voice-activated smart hardware: Enabling pure voice interaction. Figure 4This method is designed for elderly patients who are not good at operating smartphones. It aims to seamlessly integrate services into daily life scenarios through the most natural voice interaction. 1. Device preparation and activation (1) Hardware configuration: Hospitals or communities may deploy dedicated smart voice hardware in public rest areas, rehabilitation centers, or for specific patient families. The device has been pre-configured by technicians and connected to the system's cloud voice service. (2) Wake up the device: Patients do not need to perform any manual operation. They only need to say the preset wake-up word (e.g., "Hello Smart Star") near the device (usually within 5 meters). The device indicator light will light up or emit a prompt sound, indicating that the device has been woken up and is ready to receive instructions. 2. Voice dialogue process (1) Ask questions directly: After being woken up, patients do not need to look for their mobile phones or scan codes. They can directly ask questions in natural language, such as: "Hello Smart Star, what exercises can I do today?" (2) Automatic processing and voice reply: The device's microphone array will collect voice and transmit the audio stream to the cloud voice service in real time via the network. The cloud service first performs speech-to-text (STT) and then sends the text question to the core AI question-answering engine. After the engine generates the text answer, it is then synthesized into a speech stream using a text-to-speech (TTS) engine and transmitted back to the smart hardware. The device plays the answer aloud through its built-in speaker, for example: "Hello, here are some activities suitable for patients with coronary heart disease..." The beneficial effects of this invention using the above technical solution are as follows: For patients accustomed to using mobile phones, they can access a mini-program by scanning a QR code and enjoy a private Q&A experience with rich graphics and text that can be repeatedly viewed. For patients seeking convenience or having difficulty operating mobile phones (especially the elderly), they can wake up the dedicated hardware with their voice to achieve barrier-free interaction and obtain professional guidance simply by speaking. These two methods ensure that patients receive consistent, accurate, and timely health education services regardless of the channel used, truly realizing health management anytime, anywhere. The above description is only used to illustrate the technical solution of this invention and is not intended to limit it. Other modifications or equivalent substitutions made by those skilled in the art to the technical solution of this invention, as long as they do not depart from the spirit and scope of the technical solution of this invention, should be covered within the scope of the claims of this invention.

Claims

1. An evidence-based AI question-and-answer robot system for health education after coronary intervention, characterized in that, It includes: The evidence-based knowledge base construction module is used to systematically collect, screen, evaluate, and integrate authoritative evidence for health education after coronary intervention to form a structured knowledge base; The intelligent question-answering core engine module, built on retrieval-enhanced generation (RAG) technology, is used to understand user questions and generate accurate answers; the multi-terminal adaptation and deployment module is used to deploy the core functions of the system to different user terminals; and the closed-loop optimization and learning module is used for continuous improvement of the system.

2. The system according to claim 1, characterized in that, The evidence-based knowledge base construction module includes a literature retrieval and screening unit, an evidence quality evaluation and standardization unit, and a knowledge transformation and storage unit.

3. The system according to claim 1, characterized in that, The intelligent question-answering core engine module includes a user interaction interface unit, a semantic understanding and retrieval unit, and an answer generation and optimization unit.

4. The system according to claim 3, characterized in that, The user interaction interface unit supports text and voice input / output and is optimized for elderly users.

5. The system according to claim 1, characterized in that, The multi-terminal adaptation and deployment module supports front-end deployment of WeChat mini programs and extended deployment of smart hardware.

6. The system according to claim 5, characterized in that, The smart hardware expansion deployment uses a cloud-based voice service middleware layer to achieve the conversion and bridging between the audio stream collected by the smart hardware and the core question-and-answer service.

7. The system according to claim 1, characterized in that, The closed-loop optimization and learning module includes a feedback collection unit and a knowledge base update unit.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the functions of the system as described in any one of claims 1-7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the functions of the system as described in any one of claims 1-7.

Images