Bidirectional referral service method and system, computer and storage medium

Abstract

The invention provides a two-way referral service method and system, a computer and a storage medium, and the method comprises the following steps: receiving multi-mode inquiry information from a patient end, and generating a pre-inquiry result based on an artificial intelligence model according to the inquiry information; and obtaining diagnosis information from the current doctor terminal, generating a pre-referral prompt based on the dynamic resource scoring model and the classification model according to the diagnosis information, feeding back the pre-referral prompt to the current doctor terminal, and synchronizing the diagnosis information to a referral doctor terminal corresponding to the referral decision. Through intelligent prompt, a decision algorithm and department recommendation, it is ensured that a patient can be quickly and accurately referred to the most appropriate medical institution and department, the utilization efficiency of medical resources is improved, reasonable configuration of the medical resources is achieved, three-end closed-loop intercommunication covering patient referral, doctor operation and management end tracking is achieved, an optimal path for graded diagnosis and treatment is provided for the patient, and the medical quality of the patient is improved. A dynamic resource allocation center is established for medical institutions, and data-driven decision support is provided for health management departments.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular to a two-way referral service method, system, computer and storage medium. Background Art

[0002] Under the current regional medical collaboration and tiered diagnosis and treatment system, the patient referral process between primary hospitals and large comprehensive hospitals is cumbersome and information is not smooth. The following problems exist: the referral process is complicated, requiring multiple offline operations, which is time-consuming, patient information entry is cumbersome, manual operations are prone to errors, referral progress is difficult to track in real time, and communication between patients and doctors is inconvenient. Summary of the Invention

[0003] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a two-way referral service method, system, computer and storage medium, aiming to solve the technical problem of low referral efficiency in the existing technology.

[0004] In order to achieve the above objectives, in a first aspect, the present invention provides a two-way referral service method, comprising the following steps: Receive multimodal medical consultation information from the patient, generate a pre-diagnosis result based on the artificial intelligence model according to the medical consultation information, and feed it back to the patient. The pre-diagnosis result includes the pre-diagnosis department and the corresponding consulting doctor; Generate an electronic medical record based on the consultation information and the pre-consultation result, and synchronize the electronic medical record to the current doctor terminal corresponding to the consulting doctor; Obtaining diagnosis information from the current doctor, and generating a pre-referral prompt based on the diagnosis information and a dynamic resource scoring model and a classification model, and feeding it back to the current doctor, wherein the pre-referral prompt includes referral classification data obtained by the classification model and referral hospital recommendation data obtained by the dynamic resource scoring model; Receive a referral decision from the current doctor's terminal, and when the referral decision is to confirm the referral, synchronize the diagnosis information to the referring doctor's terminal corresponding to the referral decision; Receive patient hospital admission information fed back from the referring doctor to update the referral progress based on the patient hospital admission information.

[0005] According to one aspect of the above technical solution, the step of generating a pre-diagnosis result based on the artificial intelligence model according to the medical consultation information specifically includes: Based on the pre-trained speech model, the patient's symptom description is semantically understood, the symptom information is analyzed, and the corresponding diagnosis and treatment department is matched according to the symptom information; A medical knowledge graph is constructed based on medical literature and case data to reveal potential disease information based on the symptom information and generate preliminary consultation results including disease probability distribution and department recommendation ranking.

[0006] According to one aspect of the above technical solution, the symptom information includes symptom characteristics, disease codes and patient history records; The steps of matching the corresponding diagnosis and treatment department according to the symptom information specifically include: Constructing a department recommendation model based on a multilayer perceptron or a tree model, inputting the symptom information into the department recommendation model to obtain a recommended department and a matching score corresponding to the recommended department; If the matching score meets the first threshold range, a target diagnosis and treatment department is generated and corresponding medical resources are locked; If the matching score satisfies the second threshold range, multiple candidate departments and visual recommendation basis corresponding to the candidate departments are generated to generate a target diagnosis and treatment department in combination with the patient's selection; If the matching score satisfies the third threshold range, a medical question is generated to supplement the symptom information according to the patient's answer until a target diagnosis and treatment department is generated.

[0007] According to one aspect of the above technical solution, after the step of receiving the referral decision from the current doctor, the method further includes: If the referral decision is that no referral is required and the pre-referral prompt is that referral is recommended, receiving feedback information from the current doctor; The feedback information is analyzed, and parameters of the classification model are adjusted based on the analysis results.

[0008] Secondly, this application also provides a two-way referral service system, including: A pre-diagnosis module is used to receive multimodal medical information from the patient, generate a pre-diagnosis result based on the artificial intelligence model according to the medical information, and feed it back to the patient. The pre-diagnosis result includes the pre-diagnosis department and the corresponding consulting doctor; A synchronization module, configured to generate an electronic medical record based on the consultation information and the pre-consultation result, and synchronize the electronic medical record to a current doctor terminal corresponding to the consulting doctor; A prompt module is used to obtain diagnosis information from the current doctor and generate a pre-referral prompt based on the diagnosis information and the dynamic resource scoring model and classification model, and feed it back to the current doctor. The pre-referral prompt includes referral classification data obtained by the classification model and referral hospital recommendation data obtained by the dynamic resource scoring model. A referral module is configured to receive a referral decision from the current doctor, and when the referral decision is to confirm the referral, synchronize the diagnosis information to the referring doctor corresponding to the referral decision; The updating module is used to receive the patient's hospital admission information fed back from the referring doctor to update the referral progress based on the patient's hospital admission information.

[0009] According to one aspect of the above technical solution, the pre-diagnosis module is specifically used to: Based on the pre-trained speech model, the patient's symptom description is semantically understood, the symptom information is analyzed, and the corresponding diagnosis and treatment department is matched according to the symptom information; A medical knowledge graph is constructed based on medical literature and case data to reveal potential disease information based on the symptom information and generate preliminary consultation results including disease probability distribution and department recommendation ranking.

[0010] According to one aspect of the above technical solution, the symptom information includes symptom characteristics, disease codes and patient history records; The pre-diagnosis module is further specifically used for: Constructing a department recommendation model based on a multilayer perceptron or a tree model, inputting the symptom information into the department recommendation model to obtain a recommended department and a matching score corresponding to the recommended department; If the matching score meets the first threshold range, a target diagnosis and treatment department is generated and corresponding medical resources are locked; If the matching score satisfies the second threshold range, multiple candidate departments and visual recommendation basis corresponding to the candidate departments are generated to generate a target diagnosis and treatment department in combination with the patient's selection; If the matching score satisfies the third threshold range, a medical question is generated to supplement the symptom information according to the patient's answer until a target diagnosis and treatment department is generated.

[0011] According to one aspect of the above technical solution, the system further includes: A feedback module, configured to receive feedback information from the current doctor if the referral decision is that no referral is required and the pre-referral prompt is that referral is recommended; The feedback information is analyzed, and parameters of the classification model are adjusted based on the analysis results.

[0012] In a third aspect, an embodiment of the present application provides a computer comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the computer program, it implements a bidirectional referral service method as described in the first aspect.

[0013] In a fourth aspect, an embodiment of the present application provides a storage medium on which a computer program is stored, and when the program is executed by a processor, a bidirectional referral service method as described in the first aspect is implemented.

[0014] Compared with the existing technology, the beneficial effects of the present invention are: through intelligent prompts, decision-making algorithms and department recommendations, it ensures that patients can be quickly and accurately referred to the most appropriate medical institutions and departments, improves the utilization efficiency of medical resources, and realizes the rational allocation of medical resources. It covers the three-end closed-loop interconnection of patient referral, doctor operation and management-end tracking, ensures the transparency and traceability of referral, effectively solves the pain points of resource mismatch, response delay, data islands and other pain points in traditional referral, provides patients with the optimal path for tiered diagnosis and treatment, establishes a dynamic resource allocation center for medical institutions, and provides data-driven decision support for health management departments. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a flow chart of a two-way referral service method according to a first embodiment of the present invention; Figure 2 This is a structural block diagram of a two-way referral service system in a second embodiment of the present invention; Figure 3 is a schematic diagram of the hardware structure of a computer in the third embodiment of the present application; The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0016] To facilitate understanding of the present invention, the present invention will be described more fully below with reference to the accompanying drawings. The drawings illustrate several embodiments of the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and comprehensive understanding of the present invention.

[0017] It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be an intermediate element. When an element is referred to as being "connected to" another element, it may be directly connected to the other element or there may be an intermediate element. The terms "vertical," "horizontal," "left," "right," and similar expressions used herein are for illustrative purposes only.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0019] Example 1 See also Figure 1 , which shows a bidirectional referral service method in a first embodiment of the present invention, comprising the following steps: Step S100: Receive multimodal medical consultation information from the patient, generate a pre-consultation result based on the artificial intelligence model based on the consultation information, and feed it back to the patient. The pre-consultation result includes the pre-consultation department and the corresponding consulting doctor. Specifically, in this embodiment, the patient terminal has functions such as one-click login with a mobile phone account, entry of basic information and symptoms, real-time query of referral progress, SMS notification of referral, and query of referral records.

[0020] In some application scenarios of this embodiment, the patient logs in to the system with one click, and the system automatically reads the WeChat account mobile phone number to avoid manual input errors. The patient fills in relevant personal information, appointment referral time, and symptom information, and clicks Submit to initiate the referral. For patients who are not sure which department they need to be referred to, they can call the referral AI assistant, and the AI ​​will complete the pre-interview and relevant referral records and intelligent recommendations for departments. Through OCR ID card and medical insurance card recognition technology: automatically collect patient information to improve input efficiency and accuracy. The program is based on embedded development of mobile lightweight applications. There is no need to install additional apps. Both patients and doctors can operate it conveniently.

[0021] Preferably, in this embodiment, the step of generating a pre-diagnosis result based on the artificial intelligence model according to the medical inquiry information specifically includes: Based on the pre-trained speech model, the patient's symptom description is semantically understood, symptom information is analyzed, and the corresponding diagnosis and treatment department is matched based on the symptom information. Natural language processing technology can help the AI ​​system understand the text and voice descriptions input by the patient. The patient's voice is converted into text through the speech recognition engine, and the text is semantically analyzed and information extracted using natural language processing (NLP) technology to accurately capture the patient's symptom description. The AI ​​pre-diagnosis system can combine multiple input methods, such as text, voice, and images, to adapt to the usage habits of different patients. Through multimodal information processing, AI can more comprehensively understand the patient's symptom description.

[0022] A medical knowledge graph is constructed based on medical literature and case data to reveal potential disease information based on the symptom information and generate pre-diagnosis results that include disease probability distribution and department recommendation ranking. Natural language processing technology is used to extract key knowledge from a large amount of medical literature and case data to construct a medical knowledge graph. These knowledge graphs can help AI systems more accurately judge patients' conditions and recommend appropriate diagnosis and treatment plans. Furthermore, patients input symptom descriptions (e.g., "chest pain with shortness of breath"), which are then semantically understood by pre-trained language models (such as Wenxin Yiyan, ChatGPT, and DeepSeek), converting natural language into standardized medical terminology (e.g., "acute coronary syndrome"). Using big data analytics, large amounts of medical data are processed and analyzed to reveal underlying disease patterns and patterns, helping to improve diagnostic accuracy.

[0023] Preferably, in some application scenarios of this embodiment, natural language processing technology can also be used to analyze the emotional tone of the patient's speech and judge the patient's emotional state. This helps doctors better understand the patient's psychological state and provide corresponding psychological support.

[0024] Furthermore, in this embodiment, the symptom information includes symptom characteristics, disease codes, and patient history records; The steps of matching the corresponding diagnosis and treatment department according to the symptom information specifically include: Construct a department recommendation model based on a multilayer perceptron or tree model, and input the symptom information into the department recommendation model to obtain recommended departments and the matching scores corresponding to the recommended departments. Specifically, for the above-mentioned department recommendation model, a department recommendation model based on a multilayer perceptron or gradient boosting tree (GBDT) is constructed. The input layer dimension of the model is consistent with the length of the symptom feature vector, and the output layer uses a softmax function to generate the department matching probability distribution; input: symptom characteristics, disease code, and patient history; output: recommended department and matching score; learning objective: Based on historical referral data, optimize model parameters and improve recommendation accuracy.

[0025] If the matching score meets the first threshold range, a target diagnosis and treatment department is generated and the corresponding medical resources are locked. In this embodiment, the first threshold range is preferably greater than 90%. If the matching score is greater than 90%, the department is directly recommended. If the matching score satisfies the second threshold range, multiple candidate departments and visual recommendation evidence corresponding to the candidate departments are generated to generate a target diagnosis and treatment department in combination with the patient's selection. In this embodiment, the second threshold range is preferably 70%-90%, providing multiple candidate departments for the patient to choose from. If the match score meets the third threshold, a pre-diagnosis question is generated to supplement symptom information based on the patient's responses, until the target diagnosis and treatment department is determined. In this embodiment, the third threshold is preferably less than 70%, prompting the patient to provide more symptom information. Through multiple rounds of interaction with the patient, symptom information is gradually refined (e.g., "Is it accompanied by chest tightness?"). The pre-diagnosis system uses machine learning and natural language processing technologies to simulate a doctor's questioning process and gradually sort out the patient's condition.

[0026] Based on a rich corpus of medical literature, medical records, questions and answers, etc., the symptom descriptions entered by the patient are intelligently analyzed to determine the urgency of the condition, the doctor's areas of expertise, and the doctor's level. In addition, the disease knowledge graph in the medical knowledge base is used to help doctors quickly and comprehensively understand the patient's condition. Based on the information provided by the patient, the system combines the hospital's internal historical medical records and pre-consultation information to establish a unified knowledge base and customize the rule base, such as time mutual exclusion rules, sequence conflict rules, etc., to accurately recommend corresponding departments and doctors. In addition, this solution takes into account the patient's symptoms and medical history, and combines the patient's gender, age, disease site and symptoms and other characteristics, and uses medical AI and natural language processing technology to perform semantic analysis and intelligently match departments and doctors in the medical knowledge base. The system can also recommend corresponding referral departments, doctors, treatment programs, etc. according to the patient's specific situation. Step S200 generates an electronic medical record based on the consultation information and the pre-consultation results, and synchronizes the electronic medical record to the current doctor's client corresponding to the consulting physician. Using technologies such as speech recognition, natural language understanding, and image recognition, the system collects information such as the patient's chief complaint, current medical history, past medical history, allergy history, menstrual and marital history, and personal history. Before submitting a referral application, the system guides the patient to describe their symptoms and medical history to generate a pre-consultation electronic medical record.

[0027] Step S300, obtain the diagnosis information from the current doctor side, and generate a pre-referral prompt based on the diagnosis information and the dynamic resource scoring model and classification model and feedback it to the current doctor side. The pre-referral prompt includes the referral classification data obtained by the classification model and the referral hospital recommendation data obtained by the dynamic resource scoring model.

[0028] It should be noted that in some application scenarios of this embodiment, the classification model is a multi-classifier based on the XGBoost algorithm, the input features include diagnosis codes, vital signs time series data and referral indication rule matching results, and the output referral level label (referral recommended (urgent, ordinary), no referral required) and confidence score.

[0029] For the referral of patients with acute chest pain, the input diagnosis information is: ST segment elevation on electrocardiogram, elevated troponin, and the output is: referral level: urgent (confidence level 0.95); In some application scenarios of this embodiment, after the patient fills out their medical condition, the outpatient doctor can access the "pre-consultation electronic medical record" to review the pre-consultation conversation record, structured pre-consultation report, etc., to further understand the patient's condition and obtain diagnostic information. The system will continuously optimize the content of the pre-consultation report based on the patient's dynamic feedback to ensure the accuracy and completeness of the information.

[0030] Preferably, in this embodiment, the expression of the dynamic resource scoring model is: ; Where S is the hospital score corresponding to the referral hospital recommendation data, is the dynamic weight of the i-th indicator, Where n is the quantitative value of each indicator, and n is the number of indicators. Indicators include both static and dynamic indicators. Static indicators include specialist matching, number of beds, bed occupancy rate, disease cure rate, treatment costs, and number of physicians. Dynamic indicators include real-time bed availability, physician workload, transportation conditions (such as hospital distance), and historical referral success rates. Data processing is performed on a big data platform (such as Hadoop / Spark) to update dynamic indicators in real time. Static and dynamic data are standardized to facilitate subsequent model calculations. Weight optimization: Machine learning (such as random forest or LightGBM) is used to optimize weight assignment based on historical referral data. Dynamic weight coefficients are dynamically adjusted through reinforcement learning. When the classification model outputs a referral level of "urgent," the specialist matching weight in the dynamic resource scoring model is increased by 20%, and hospitals with emergency green channels are prioritized for matching.

[0031] Specifically, in this embodiment, the data sources of the above-mentioned diagnostic information are electronic medical record systems (EMR), inspection and testing systems (LIS, PACS), previous medical records, etc. Based on this, key information such as the patient's medical history, symptoms, signs, test results (such as blood routine, liver and kidney function, etc.), imaging reports (such as CT / MRI), medication records, etc. are extracted. The data is preprocessed for subsequent use, including cleaning, deduplication and normalization of data, elimination of outliers, and unification of feature formats.

[0032] In addition to ranking pre-referral hospitals based on a dynamic resource scoring model, pre-referral notifications also analyze electronic medical records uploaded by physicians, i.e., diagnostic data. Specifically, using labeled datasets, a classification model is constructed using random forest, XGBoost, or deep learning models (such as LSTM) to predict whether a patient meets referral criteria. Input: patient feature vector (such as symptom severity, test results, etc.); output: referral recommendation (such as "recommended referral to cardiology" or "no referral required"). Referral indication rules are also defined based on industry guidelines and expert experience (e.g., myocardial infarction with chest pain requires urgent referral). This model combines rules and data to improve prediction accuracy.

[0033] More specifically, the trigger point in the patient diagnosis and treatment process is when the doctor enters patient information (symptoms, test results, etc.), and the system conducts a real-time assessment to determine whether referral criteria are met. Prompt logic includes, for example, if an indicator exceeds a limit, a real-time pop-up window prompting "urgent referral suggestion"; if suspected (e.g., an indicator approaches a critical value): a prompt "further examination or referral recommended"; analysis results are displayed in the form of charts or radar charts to help doctors quickly understand the basis for the prompt. Multi-dimensional data analysis replaces sole reliance on doctor experience, improving the scientific nature of decision-making. By comprehensively considering multiple dimensions of information such as the patient's symptoms, medical history, and test results, doctors can provide scientific referral recommendations, thereby improving the accuracy and timeliness of referral decisions, reducing unnecessary referrals or delayed referrals due to inaccurate subjective judgments, and enhancing the quality of medical services.

[0034] Preferably, in this embodiment, after the step of receiving the referral decision from the current doctor, the method further includes: If the referral decision is that no referral is required and the pre-referral prompt is that referral is recommended, receiving feedback information from the current doctor; The feedback information is analyzed, and parameters of the classification model are adjusted based on the analysis results.

[0035] In this embodiment, for example Specifically, doctors make final referral decisions based on AI-generated recommendations. At the same time, the system records referral results and feedback to continuously optimize the model's performance advantages and differences. The specific steps are as follows: Establish a feedback mechanism: First, a systematic feedback mechanism needs to be established so that doctors can easily provide feedback on AI-generated referral recommendations. This includes integrating feedback tools into medical information systems so that doctors can submit feedback directly after using the AI ​​system.

[0036] Collect feedback regularly: Collect feedback from doctors regularly to understand their satisfaction with AI-generated referral recommendations and suggestions for improvement. This feedback can be quantitative (such as a scoring system) or qualitative (such as open comments).

[0037] Analyze and process feedback: Detailed analysis of collected feedback is performed to identify areas where the AI ​​system performs poorly or contains errors. For example, a doctor may indicate that the AI's recommendations for certain types of cases are inaccurate or omit important diagnostic information.

[0038] Model adjustment and optimization: Based on the feedback, the AI ​​model is adjusted and optimized. This may include updating the training dataset, adjusting algorithm parameters, or introducing new features to improve the accuracy and reliability of the model.

[0039] Human-machine collaboration: Physician participation is crucial in the AI-generated referral process. Physicians can review and revise AI recommendations based on their clinical experience and expertise, ensuring the final diagnostic recommendations are both accurate and reliable.

[0040] Continuous learning and updating: AI systems should be able to continuously learn and improve from new clinical data and physician feedback. This helps maintain the model's advancement and adaptability.

[0041] Data augmentation and model iteration: Utilizing the feedback data provided by doctors, we retrain and iterate the model. For example, we can add case data related to the feedback to enrich the training set, thereby improving the model's performance in specific scenarios.

[0042] Step S400: receiving a referral decision from the current doctor's end; when the referral decision is to confirm the referral, synchronizing the diagnosis information to the referring doctor's end corresponding to the referral decision.

[0043] In some application scenarios of this embodiment, the doctor registers a relevant account for the first time, fills in the medical institution, department, name, and the system automatically obtains the phone number; logs in to the referral system through the system, selects the referral type (outpatient, emergency, inpatient), selects the referral institution and referral department, patient basic information, appointment referral time, diagnosis and related medical record information, and clicks Submit Referral. The AI ​​assistant can provide referral indicators and referral department (doctor) recommendations through intelligent analysis. Doctors in the medical alliance can adjust the referral department or doctor according to the suggestions and initiate referrals with one click.

[0044] Step S500: Receive patient admission information from the referring physician and update the referral progress based on the patient admission information. After initiating a referral application, the system will review each step of the process. The referral-related approving user will receive real-time text messages and system notifications informing them of the referral progress and appointment time. After approval, the referred patient will receive a timely admission notification. The system will also generate a referral form for the patient to verify upon arrival.

[0045] In some application scenarios of this embodiment, the system also includes a management terminal for interacting with the patient side and the doctor side. The management terminal monitors the referral volume of each medical institution in real time, generates data reports, analyzes referral time and patient flow, and provides a decision-making basis for the allocation of medical resources. At the same time, real-time referral data, including the distribution of referral institutions, the number of referred patients, and the average time taken for referrals, are displayed on the large screen within the hospital. The management terminal includes the following key technical points and improvements: OCR ID card and medical insurance card recognition technology: automatically collects patient information and improves data entry efficiency and accuracy.

[0046] Mobile lightweight application: Based on embedded development of mobile lightweight application, there is no need to install additional apps, and both patients and doctors can operate it conveniently.

[0047] Self-service registration function: There is no need to collect referral user information offline. Doctors in the medical alliance can self-register in the system and initiate referral applications in a timely manner.

[0048] Patient referral end: Two-way referral opens a hospital referral green channel for patients.

[0049] Doctor referral terminal: This system is not only used by lower-level medical alliances, but all employees of the medical alliance at this level can use the referral system to initiate upward referral applications, which can further expand the source of patients for the hospital.

[0050] Real-time SMS and WeChat notifications: Patients and doctors receive referral status promptly to avoid information delays.

[0051] Data visualization analysis: Display regional referral flows through PC, mobile and large screen terminals, providing intuitive decision-making support for management.

[0052] Implementation plan for intelligent referral indication prompts: Through in-depth analysis of the patient's medical data, the system can intelligently prompt whether referral is needed, helping doctors make more accurate decisions.

[0053] In some application scenarios of this embodiment, the specific business logic is as follows: 1. Patient selects department and enters information: Patients first select their department in the intelligent pre-consultation system and enter their personal information into their health records. This step ensures that the patient's basic information and medical history are accurately recorded and processed by the system.

[0054] 2. Patient Description of Symptoms: Patients use the chat interface to describe their symptoms, including specific manifestations, causes, and location. This information can be written or input via voice. The system then uses natural language processing technology to convert the patient's description into medical terms.

[0055] 3. Intelligent Diagnosis and Symptom Analysis: Based on the symptom information provided by the patient, the AI ​​system simulates a doctor's consultation process, gradually deepening the patient's condition through multiple rounds of dialogue. The system analyzes the patient's symptoms in conjunction with the medical knowledge base and generates preliminary diagnostic recommendations.

[0056] 4. Recommended Departments and Doctors: Based on the patient's symptoms and condition analysis, the AI ​​system recommends the most appropriate department and doctor specializing in the condition. The recommended results are displayed to the patient in a list, helping them quickly find the appropriate referral department and doctor.

[0057] 5. Generate electronic medical records and synchronize them to the doctor's computer: After the patient completes the pre-consultation, the system automatically generates an electronic medical record and synchronizes the pre-consultation results to the attending doctor's computer. Doctors can directly view the patient's pre-consultation information without manual input, thereby improving diagnosis and treatment efficiency.

[0058] 6. Doctor Confirmation and Further Diagnosis and Treatment: The attending doctor will conduct further examinations or treatment based on the patient's specific circumstances and the preliminary consultation results provided by the AI ​​system. If a referral is required, the doctor can submit a referral request in the system and specify the appropriate referral department and doctor.

[0059] 7. Patient receives referral suggestion: The system will notify the patient of the referral suggestion. The patient can choose to accept the referral. The system will automatically fill out the referral application and submit the application with one click.

[0060] Specifically, referral situations include: 1. Patients are referred to higher-level hospitals After visiting a primary care hospital, Patient A was recommended by the doctor to be transferred to a higher-level hospital within a medical alliance for further examination. The doctor entered the patient's medical history and needs into the referral system and submitted the referral. The AI ​​assistant then conducted a comprehensive analysis and provided a referral prompt.

[0061] After receiving the SMS notification, the approval department of the superior medical alliance hospital confirms the patient's relevant information, makes relevant preparations for the reception, and notifies the patient to come to the hospital for treatment.

[0062] Patient A received a text message notification and went to the superior medical alliance hospital for treatment.

[0063] The management side simultaneously records the referral process and updates the data in the statistical report.

[0064] 2. Lower-level hospitals accept referral patients After patient B completes treatment at the superior medical alliance hospital, if further rehabilitation is required in the later stage, the superior medical alliance doctor will use the doctor's end to transfer patient B's medical records and rehabilitation suggestions to the grassroots hospital through the system to ensure that the patient continues rehabilitation treatment in the grassroots hospital and avoid occupying resources of the large hospital.

[0065] 3. Patients directly submit referral applications to higher-level hospitals After patient C followed the two-way referral system through social software, he felt unwell later. The AI ​​assistant pre-diagnosed and recommended a department. He submitted an application for hospitalization in a department of a higher-level hospital through the system. After the higher-level hospital reviewed the application, he went through the referral green channel, and a dedicated person contacted the patient to confirm the patient's visit to the hospital and assisted the patient in completing the relevant admission procedures.

[0066] The difference between this application and the existing technology is that it is developed based on WeChat mini-programs, does not require additional application downloads, and has a low usage threshold; it integrates OCR ID card / medical insurance card recognition technology to realize automatic information entry and reduce manual operations; it has comprehensive two-way referral and management functions, realizing information exchange between superior and subordinate hospitals and closed-loop management of patient flow; AI-assisted management, through AI-assisted analysis, provides more effective assistance for referrals; it has a large-screen decision-making panel to intuitively display referral data and assist hospital management in optimizing resource allocation; refined reporting tools: provide fine-grained data analysis for the management side, such as performance statistics by quarter, disease type, and referring doctor; dynamic map regional distribution: obtain user GPS positioning through the referral initiator, add referral dynamic map to the decision-making screen, and intuitively display patient flow, referral coverage area and hot spot distribution; open interface: compatible with more types of medical information systems (such as LIS, PACS, etc.) to achieve comprehensive data interconnection; modular design: provides customized modules for hospitals of different sizes to meet the different needs from small medical alliances to large regional collaboration. Allow hospitals to customize referral workflows according to their own needs, such as adjusting approval nodes or optimizing admission processes.

[0067] The advantages are: Improved referral efficiency: The entire process from patient information entry to referral is seamlessly connected, significantly shortening referral time. Reduced operational difficulty: Based on the WeChat ecosystem, users do not need to learn or download additional applications, which is convenient and fast. Innovation in referral model: Not limited to traditional doctor-side referral, the hospital has developed staff referral and patient-side referral modules, providing green channel services to further improve patient satisfaction and the efficiency of the referral system. Intelligent management: Real-time data visualization helps the hospital make scientific decisions and optimize resource allocation. Improved efficiency in the use of medical resources: Through intelligent prompts, decision-making algorithms and department recommendations, patients can be quickly and accurately referred to the most appropriate medical institutions and departments, improving the efficiency of medical resource utilization. Promote the implementation of the tiered diagnosis and treatment system: Through intelligent means, guide patients to make reasonable referrals, promote the implementation of the tiered diagnosis and treatment system, and achieve rational allocation of medical resources. Full-process closed-loop management: Covering the three-end closed-loop interconnection of patient referral, doctor operation and management-end tracking, to ensure transparent and traceable referrals.

[0068] In summary, the two-way referral service method in the above-mentioned embodiment of the present invention ensures that patients can be quickly and accurately referred to the most appropriate medical institutions and departments through intelligent prompts, decision-making algorithms and department recommendations, improves the utilization efficiency of medical resources, and realizes the rational allocation of medical resources. It covers the three-end closed-loop interconnection of patient referral, doctor operation and management-end tracking, ensures the transparency and traceability of referral, effectively solves the pain points of resource mismatch, response delay, data islands and other pain points in traditional referral, provides patients with the optimal path for tiered diagnosis and treatment, establishes a dynamic resource allocation center for medical institutions, and provides data-driven decision support for health management departments.

[0069] Example 2 The second embodiment of the present application further provides a two-way referral service system, which is used to implement the embodiments and preferred embodiments, and will not be repeated hereafter. As used below, the terms "module," "unit," "subunit," etc. may refer to a combination of software and / or hardware that implements a predetermined function. Although the systems described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0070] like Figure 2 As shown, the system includes: a pre-diagnosis module 100, a synchronization module 200, a prompt module 300, a referral module 400 and an update module 500; The pre-diagnosis module 100 is used to receive multimodal medical information from the patient, generate a pre-diagnosis result based on the artificial intelligence model according to the medical information, and feed it back to the patient. The pre-diagnosis result includes the pre-diagnosis department and the corresponding doctor. The synchronization module 200 is used to generate an electronic medical record based on the consultation information and the pre-consultation result, and synchronize the electronic medical record to the current doctor terminal corresponding to the consulting doctor; The prompt module 300 is used to obtain diagnosis information from the current doctor, and generate a pre-referral prompt based on the diagnosis information and the dynamic resource scoring model and the classification model, and feed it back to the current doctor. The pre-referral prompt includes referral classification data obtained by the classification model and referral hospital recommendation data obtained by the dynamic resource scoring model. The referral module 400 is used to receive a referral decision from the current doctor, and when the referral decision is to confirm the referral, synchronize the diagnosis information to the referring doctor corresponding to the referral decision; The updating module 500 is used to receive the patient's hospital admission information fed back from the referring doctor, so as to update the referral progress based on the patient's hospital admission information.

[0071] Preferably, in this embodiment, the pre-diagnosis module 100 is specifically used to: Based on the pre-trained speech model, the patient's symptom description is semantically understood, the symptom information is analyzed, and the corresponding diagnosis and treatment department is matched according to the symptom information; A medical knowledge graph is constructed based on medical literature and case data to reveal potential disease information based on the symptom information and generate preliminary consultation results including disease probability distribution and department recommendation ranking.

[0072] Preferably, in this embodiment, the symptom information includes symptom characteristics, disease codes and patient history records; The pre-diagnosis module 100 is further configured to: Constructing a department recommendation model based on a multilayer perceptron or a tree model, inputting the symptom information into the department recommendation model to obtain a recommended department and a matching score corresponding to the recommended department; If the matching score meets the first threshold range, a target diagnosis and treatment department is generated and corresponding medical resources are locked; If the matching score satisfies the second threshold range, multiple candidate departments and visual recommendation basis corresponding to the candidate departments are generated to generate a target diagnosis and treatment department in combination with the patient's selection; If the matching score satisfies the third threshold range, a medical question is generated to supplement the symptom information according to the patient's answer until a target diagnosis and treatment department is generated.

[0073] Preferably, in this embodiment, the system further includes: A feedback module, configured to receive feedback information from the current doctor if the referral decision is that no referral is required and the pre-referral prompt is that referral is recommended; The feedback information is analyzed, and parameters of the classification model are adjusted based on the analysis results.

[0074] It should be noted that each module can be a functional module or a program module, and can be implemented by software or hardware. For modules implemented by hardware, each module can be located in the same processor; or each module can be located in different processors in any combination.

[0075] The third embodiment of the present application provides a computer. It can be understood that the principles mentioned in the bidirectional referral service system in this embodiment correspond to the bidirectional referral service method in the first embodiment of the present application. For details of the relevant principles not described, please refer to the first embodiment and will not be elaborated here.

[0076] The computer may include a processor 81 and a memory 82 storing computer program commands.

[0077] Specifically, the processor 81 may include a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.

[0078] Memory 82 may include a large-capacity memory for data or commands. By way of example, and not limitation, memory 82 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 82 may include removable or non-removable (or fixed) media. Where appropriate, memory 82 may be internal or external to the data processing device. In certain embodiments, memory 82 is non-volatile memory. In certain embodiments, memory 82 includes read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically alterable ROM (EAROM) or a flash memory (FLASH), or a combination of two or more of these. Under appropriate circumstances, the RAM can be a static random access memory (SRAM) or a dynamic random access memory (DRAM), where the DRAM can be a fast page mode dynamic random access memory (FPMDRAM), an extended data out dynamic random access memory (EDODRAM), a synchronous dynamic random access memory (SDRAM), etc.

[0079] The memory 82 may be used to store or cache various data files that need to be processed and / or used for communication, as well as possible computer program commands executed by the processor 81 .

[0080] The processor 81 implements any one of the bidirectional referral service methods in the above embodiments by reading and executing the computer program commands stored in the memory 82 .

[0081] In some embodiments, the computer may further include a communication interface 83 and a bus 80. Figure 3 As shown, the processor 81, the memory 82, and the communication interface 83 are connected via a bus 80 and communicate with each other.

[0082] The communication interface 83 is used to implement communication between the various modules, devices, units, and / or devices in the embodiments of the present application. The communication interface 83 can also implement data communication with other components such as: external devices, image / data acquisition equipment, databases, external storage, and image / data processing workstations.

[0083] The bus 80 includes hardware, software, or both, and couples computer components together. The bus 80 includes, but is not limited to, at least one of the following: a data bus, an address bus, a control bus, an expansion bus, and a local bus. By way of example and not limitation, bus 80 may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an InfiniBand interconnect, a Low Pin Count (LPC) bus, a memory bus, a Micro Channel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local Bus (VLB) bus, or other suitable buses, or a combination of two or more of these. Bus 80 may include one or more buses, where appropriate. Although embodiments herein describe and illustrate a particular bus, this application contemplates any suitable bus or interconnect.

[0084] In addition, in conjunction with the two-way referral service method in the above embodiments, a fourth embodiment of the present application provides a readable storage medium having computer program commands stored thereon; when the computer program commands are executed by a processor, any one of the two-way referral service methods in the above embodiments is implemented.

[0085] The technical features of the above-described embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0086] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A two-way referral service method, characterized in that: The following steps are involved: Receive multimodal medical consultation information from the patient, generate a pre-diagnosis result based on the artificial intelligence model according to the medical consultation information, and feed it back to the patient. The pre-diagnosis result includes the pre-diagnosis department and the corresponding consulting doctor; Generate an electronic medical record based on the consultation information and the pre-consultation result, and synchronize the electronic medical record to the current doctor terminal corresponding to the consulting doctor; Obtaining diagnosis information from the current doctor, and generating a pre-referral prompt based on the diagnosis information and a dynamic resource scoring model and a classification model, and feeding it back to the current doctor, wherein the pre-referral prompt includes referral classification data obtained by the classification model and referral hospital recommendation data obtained by the dynamic resource scoring model; Receive a referral decision from the current doctor's terminal, and when the referral decision is to confirm the referral, synchronize the diagnosis information to the referring doctor's terminal corresponding to the referral decision; Receive patient hospital admission information fed back from the referring doctor to update the referral progress based on the patient hospital admission information.

2. The two-way referral service method according to claim 1, characterized in that: The steps of generating a pre-diagnosis result based on the artificial intelligence model according to the medical consultation information specifically include: Based on the pre-trained speech model, the patient's symptom description is semantically understood, the symptom information is analyzed, and the corresponding diagnosis and treatment department is matched according to the symptom information; A medical knowledge graph is constructed based on medical literature and case data to reveal potential disease information based on the symptom information and generate preliminary consultation results including disease probability distribution and department recommendation ranking.

3. The two-way referral service method according to claim 2, characterized in that: The symptom information includes symptom characteristics, disease codes and patient history records; The steps of matching the corresponding diagnosis and treatment department according to the symptom information specifically include: Constructing a department recommendation model based on a multilayer perceptron or a tree model, inputting the symptom information into the department recommendation model to obtain a recommended department and a matching score corresponding to the recommended department; If the matching score meets the first threshold range, a target diagnosis and treatment department is generated and corresponding medical resources are locked; If the matching score satisfies the second threshold range, multiple candidate departments and visual recommendation basis corresponding to the candidate departments are generated to generate a target diagnosis and treatment department in combination with the patient's selection; If the matching score satisfies the third threshold range, a medical question is generated to supplement the symptom information according to the patient's answer until a target diagnosis and treatment department is generated.

4. The two-way referral service method according to claim 1, characterized in that: After receiving the referral decision from the current doctor, the method further includes: If the referral decision is that no referral is required and the pre-referral prompt is that referral is recommended, receiving feedback information from the current doctor; The feedback information is analyzed, and parameters of the classification model are adjusted based on the analysis results.

5. A two-way referral service system, characterized in that: include: A pre-diagnosis module is used to receive multimodal medical information from the patient, generate a pre-diagnosis result based on the artificial intelligence model according to the medical information, and feed it back to the patient. The pre-diagnosis result includes the pre-diagnosis department and the corresponding consulting doctor; A synchronization module, configured to generate an electronic medical record based on the consultation information and the pre-consultation result, and synchronize the electronic medical record to a current doctor terminal corresponding to the consulting doctor; A prompt module is used to obtain diagnosis information from the current doctor and generate a pre-referral prompt based on the diagnosis information and the dynamic resource scoring model and classification model, and feed it back to the current doctor. The pre-referral prompt includes referral classification data obtained by the classification model and referral hospital recommendation data obtained by the dynamic resource scoring model. A referral module is configured to receive a referral decision from the current doctor, and when the referral decision is to confirm the referral, synchronize the diagnosis information to the referring doctor corresponding to the referral decision; The updating module is used to receive the patient's hospital admission information fed back from the referring doctor to update the referral progress based on the patient's hospital admission information.

6. The two-way referral service system according to claim 5, characterized in that: The pre-diagnosis module is specifically used for: Based on the pre-trained speech model, the patient's symptom description is semantically understood, the symptom information is analyzed, and the corresponding diagnosis and treatment department is matched according to the symptom information; A medical knowledge graph is constructed based on medical literature and case data to reveal potential disease information based on the symptom information and generate preliminary consultation results including disease probability distribution and department recommendation ranking.

7. The two-way referral service system according to claim 6, characterized in that: The symptom information includes symptom characteristics, disease codes and patient history records; The pre-diagnosis module is further specifically used for: Constructing a department recommendation model based on a multilayer perceptron or a tree model, inputting the symptom information into the department recommendation model to obtain a recommended department and a matching score corresponding to the recommended department; If the matching score meets the first threshold range, a target diagnosis and treatment department is generated and corresponding medical resources are locked; If the matching score satisfies the second threshold range, multiple candidate departments and visual recommendation basis corresponding to the candidate departments are generated to generate a target diagnosis and treatment department in combination with the patient's selection; If the matching score satisfies the third threshold range, a medical question is generated to supplement the symptom information according to the patient's answer until a target diagnosis and treatment department is generated.

8. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the bidirectional referral service method according to any one of claims 1 to 4 is implemented.

9. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the bidirectional referral service method as described in any one of claims 1 to 4 is implemented.

Images