Patient service information push method and device
By using AI technology to conduct in-depth analysis of post-diagnosis patient data, accurate classification and personalized management needs identification are achieved, solving the problems of lack of specificity and cumbersome operation in existing systems, and improving the efficiency and timeliness of post-diagnosis patient management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING JINGDONG TUOXIAN TECH CO LTD
- Filing Date
- 2026-03-19
- Publication Date
- 2026-06-12
Smart Images
Figure CN122201849A_ABST
Abstract
Description
Technical Field
[0001] The embodiments disclosed herein relate to the field of internet healthcare technology, specifically to a method and apparatus for pushing patient service information. Background Technology
[0002] Existing post-diagnosis patient management technologies primarily focus on basic information recording and general management functions, providing standardized follow-up reminders and message sending. These technologies lack in-depth analysis of patient condition characteristics, recovery stages, and historical treatment data, failing to achieve accurate patient classification and needs matching. Furthermore, the user interface design lacks specificity, requiring doctors to switch between multiple menu entry points to initiate follow-up visits and set up follow-up inquiries.
[0003] For example, after a patient completes hypertension diagnosis and treatment and reports back, the existing system can only record the patient's basic information and diagnosis results. Doctors need to manually search for the patient in the system, select the "follow-up" function from among many functional modules, set the follow-up time and content, and then send it to the patient. For patients with different conditions such as diabetes and coronary heart disease who report back at the same time, the system cannot automatically distinguish their management needs. Doctors need to repeat tedious operations, which is not only time-consuming and laborious, but also prone to untimely management due to the cumbersome operation. At the same time, due to the lack of analysis on the dynamics of the patient's condition, doctors find it difficult to accurately determine whether the patient needs further questioning about the details of the condition, which may lead to the omission of key health information and affect subsequent diagnosis and treatment decisions. Summary of the Invention
[0004] The embodiments of this disclosure provide a method and apparatus for pushing patient service information.
[0005] In a first aspect, embodiments of this disclosure provide a method for pushing patient service information, comprising: determining at least one recommended action for each of one or more managed patients based on their respective medical data; outputting options corresponding to each recommended action for one or more managed patients on a doctor's operation page; in response to a target patient's target option being selected, loading template content corresponding to the target option from a pre-built action library and displaying the template content on the doctor's operation page; and in response to the template content being confirmed, pushing the template content to the target patient's terminal.
[0006] In some embodiments, determining at least one recommended action for each managed patient based on their respective medical data includes: for each managed patient, performing the following steps: determining the patient's management category using a pre-trained artificial intelligence model based on the patient's medical data; determining the patient's management needs using an artificial intelligence model based on the patient's medical data and management category; and determining at least one recommended action for the patient using an artificial intelligence model based on the patient's management category and management needs.
[0007] In some embodiments, the patient's management category includes a primary classification, a secondary classification, and a tertiary subclassification; and the patient's management category is determined based on the patient's medical data using a pre-trained artificial intelligence model, including: determining the patient's core characteristics based on the patient's medical data using a pre-trained artificial intelligence model, wherein the core characteristics include at least one of the following: comprehensive risk coefficient characteristics, treatment response characteristics, and disease stability characteristics; determining the primary classification of the disease type dimension based on the patient's medical data using an artificial intelligence model; determining the secondary classification of the rehabilitation stage dimension based on the primary classification and the patient's medical data using an artificial intelligence model; and determining the tertiary subclassification of the risk and management intensity dimension based on the patient's core characteristics using an artificial intelligence model based on the secondary classification.
[0008] In some embodiments, the method further includes: in response to detecting an update to the patient's medical data, redetermining the patient's management category based on the updated medical data using an artificial intelligence model.
[0009] In some embodiments, the patient’s management needs include critical management needs, assisted rehabilitation management needs, and long-term rehabilitation management needs.
[0010] In some embodiments, the method further includes: verifying the patient's management needs based on historical effect feedback information and / or historical management records on the doctor's end; and correcting the patient's management needs in response to verification failure.
[0011] In some embodiments, the options corresponding to each recommended action of one or more managed patients are output on the doctor's operation page, including: in response to the number of recommended actions of the patient exceeding a predetermined number, determining the priority of each recommended action of the patient; and outputting the options of selecting a predetermined number of recommended actions of the patient in descending order of priority on the doctor's operation page.
[0012] In some embodiments, determining the priority of each recommended action for the patient includes: for each recommended action, calculating the semantic matching degree between the recommended action and the patient's management needs; determining a clinical effectiveness indicator based on the improvement rate of rehabilitation effect after the execution of the recommended action according to historical data statistics; determining a timeliness indicator based on the urgency of the recommended action; determining the doctor's preference based on the doctor's historical adoption rate of the recommended action; and determining the priority of the recommended action for the patient based on the weighted sum of the semantic matching degree, clinical effectiveness indicator, timeliness indicator, and doctor's preference.
[0013] In some embodiments, the method further includes adjusting the priority of each recommended action of the patient by at least one of the following methods: adjusting the priority of each recommended action of the patient based on the patient's feedback on each recommended action; adjusting the priority of each recommended action of the patient based on a change in the patient's management category; or adjusting the priority of each recommended action of the patient based on an operation that adjusts the order of the options.
[0014] In some embodiments, the method further includes: displaying an action management entry on the doctor's operation page; and, in response to detecting that the action management entry has been clicked, displaying a list of actions in the action library and custom configuration options for recommended actions.
[0015] In some embodiments, the method further includes: acquiring historical medical records of one or more managed patients; receiving post-diagnosis health status information input by each patient; cleaning the historical medical records and health status information; and standardizing the cleaned data to generate treatment data for each patient.
[0016] Secondly, embodiments of this disclosure provide a method for training an artificial intelligence model, comprising: acquiring sample data, wherein the sample data includes sample diagnosis and treatment data, sample management categories, sample management needs, and sample actions; determining a predicted management category based on the sample diagnosis and treatment data using an initial artificial intelligence model; determining a predicted management need based on the sample diagnosis and treatment data and the predicted management category using an artificial intelligence model; determining a predicted action based on the predicted management category and the predicted management need using an artificial intelligence model; and adjusting the network parameters of the artificial intelligence model based on the differences between the sample management category and the predicted management category, the differences between the sample management need and the predicted management need, and the differences between the sample action and the predicted action.
[0017] Thirdly, embodiments of this disclosure provide a patient service information push device, comprising: a recommendation unit configured to determine at least one recommended action for each patient based on the diagnosis and treatment data of one or more managed patients; an output unit configured to output options corresponding to each recommended action for one or more managed patients on a doctor's operation page; a loading unit configured to load template content corresponding to the target option from a pre-built action library in response to the selection of the target option by the target patient, and display the template content on the doctor's operation page; and a push unit configured to push the template content to the target patient's terminal in response to the confirmation of the template content.
[0018] Fourthly, embodiments of this disclosure provide an apparatus for training an artificial intelligence model, comprising: an acquisition unit configured to acquire sample data, wherein the sample data includes sample diagnosis and treatment data, sample management categories, sample management needs, and sample actions; a category determination unit configured to determine a predicted management category based on the sample diagnosis and treatment data and using an initial artificial intelligence model; a need determination unit configured to determine a predicted management need based on the sample diagnosis and treatment data and the predicted management category and using an artificial intelligence model; an action determination unit configured to determine a predicted action based on the predicted management category and the predicted management need and using an artificial intelligence model; and an adjustment unit configured to adjust the network parameters of the artificial intelligence model based on the differences between the sample management category and the predicted management category, the differences between the sample management need and the predicted management need, and the differences between the sample action and the predicted action.
[0019] Fifthly, embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more computer programs stored thereon, which, when executed by the one or more processors, cause the one or more processors to perform the method as described in any one of the first or second aspects.
[0020] In a sixth aspect, embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method as described in any one of the first or second aspects.
[0021] In a seventh aspect, embodiments of this disclosure provide a computer program product including a computer program that, when executed by a processor, implements the method as described in any one of the first or second aspects.
[0022] Current technologies cannot accurately analyze and classify data based on patient condition characteristics, recovery stage, and treatment history. They only provide general management functions, which are insufficient to meet the personalized management needs of different patients, resulting in inadequate targeted management measures and affecting rehabilitation outcomes. The doctor's operation process is cumbersome: current technologies have not optimized the post-treatment management interface and process. Doctors need to manually search for patients, switch function modules, and set management parameters, which involves many steps and is time-consuming. Especially when there are many patients, management efficiency is extremely low, increasing the doctor's workload. Current technologies lack intelligent recommendation mechanisms, requiring doctors to rely on personal experience to determine which management actions to take for patients (such as follow-up visits, follow-up questions, etc.). This can easily lead to inconsistent management measures due to differences in experience and cannot efficiently match the actual needs of patients.
[0023] To address the shortcomings of existing technologies, such as insufficient customization of post-diagnosis patient management, cumbersome doctor operations, and lack of intelligent management action recommendations, this application provides an AI-based method for pushing post-diagnosis patient service information. By leveraging AI technology to deeply analyze the condition data of patients reporting for treatment after diagnosis, it achieves accurate patient classification and identification of personalized management needs. The doctor's interface is optimized, supporting quick selection of follow-up visits and sending follow-up inquiry packages via left swiping, simplifying the operation process. Based on the patient's condition characteristics and dynamic changes, it intelligently recommends suitable management solutions, improving the efficiency and targeted nature of doctor management.
[0024] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description
[0025] Other features, objects, and advantages of this disclosure will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is an exemplary system architecture diagram to which one embodiment of this disclosure can be applied; Figure 2 This is a flowchart of an embodiment of the patient service information push method according to the present disclosure; Figure 3 This is a schematic diagram of an application scenario of the patient service information push method according to this disclosure; Figure 4 This is a flowchart of an embodiment of a method for training an artificial intelligence model according to the present disclosure; Figure 5 This is a schematic diagram of a structure of an embodiment of the patient service information push device according to the present disclosure; Figure 6 This is a schematic diagram of a structure for training an artificial intelligence model according to an embodiment of the present disclosure; Figure 7 This is a schematic diagram of the structure of a computer system suitable for implementing embodiments of the present disclosure. Detailed Implementation
[0026] The present disclosure will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.
[0027] It should be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other. This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.
[0028] The technical terms used in this article are explained below: (1) Patients who register after diagnosis: refers to patients who have completed outpatient or inpatient diagnosis and treatment and have registered and reported to the hospital as required, and have entered the post-diagnosis management stage.
[0029] (2) Action Library: A database that stores various post-diagnosis management related actions and corresponding standardized content templates, covering all scenarios of management actions such as follow-up, follow-up questioning, guidance, and reminders.
[0030] (3) Follow-up: Follow-up is a medical term that refers to the behavior of medical institutions to continuously track and investigate discharged patients. The purpose is to evaluate the treatment effect, monitor the rehabilitation process and optimize the quality of medical services by collecting patients' health information and feedback.
[0031] Figure 1 An exemplary system architecture 100 is shown, in which embodiments of the patient service information push method or patient service information push device of the present disclosure can be applied.
[0032] like Figure 1 As shown, system architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. Network 104 serves as the medium for providing communication links between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.
[0033] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, and 103, such as post-diagnosis management applications, 3D video players, web browser applications, shopping applications, search applications, instant messaging tools, email clients, social media platform software, etc.
[0034] Terminal devices 101, 102, and 103 can be either hardware or software. When terminal devices 101, 102, and 103 are hardware, they can be various electronic devices with displays and supporting page display, including but not limited to smartphones, tablets, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 players (Moving Picture Experts Group Audio Layer IV), laptops, and desktop computers, etc. When terminal devices 101, 102, and 103 are software, they can be installed in the aforementioned electronic devices. They can be implemented as multiple software programs or software modules (e.g., to provide distributed services) or as a single software program or software module. No specific limitations are imposed here.
[0035] Server 105 can be a server that provides various services, such as a backend medical server that supports the doctor's operation page displayed on terminal devices 101, 102, and 103. The backend medical server can store patient diagnosis and treatment data, make targeted post-diagnosis management recommendations based on the patient diagnosis and treatment data, and display the recommendations on the doctor's page for the doctor to choose from. After the doctor selects a recommended action, the associated content can be sent to the patient's terminal device.
[0036] It's important to note that a server can be either hardware or software. When a server is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or as a single server. When a server is software, it can be implemented as multiple software programs or software modules (e.g., multiple software programs or software modules used to provide distributed services), or as a single software program or software module. No specific limitations are made here. A server can also be a server for a distributed system, or a server integrated with blockchain technology. A server can also be a cloud server, or an intelligent cloud computing server or intelligent cloud host with artificial intelligence technology.
[0037] It should be noted that the patient service information push method provided in the embodiments of this disclosure is generally executed by server 105, and correspondingly, the patient service information push device is generally set in server 105.
[0038] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.
[0039] Continue to refer to Figure 2The diagram illustrates a flow 200 of an embodiment of a patient service information push method according to the present disclosure. The patient service information push method includes the following steps: Step 201: Based on the individual medical data of one or more managed patients, determine at least one recommended action for each patient using a pre-trained artificial intelligence model.
[0040] In this embodiment, the executing entity of the patient service information push method (e.g. Figure 1 The server shown can receive post-diagnosis consultation requests from patient terminals via wired or wireless connections. Patients authorize the hospital to manage their medical data after sending a post-diagnosis consultation request. Patients whose requests are confirmed by the hospital are referred to as "managed patients," and all patients mentioned in this article are referred to as "patients." The server can retrieve pre-stored medical data based on the patient ID and input the patient's medical data into a pre-trained artificial intelligence model (e.g., a large language model). The model then uses pre-set prompts (e.g., recommending post-diagnosis management actions to the doctor based on the patient's medical data). The artificial intelligence model can extract patient characteristics from the patient data, combine them with medical knowledge, and recommend multiple actions to the doctor from a pre-built action library.
[0041] A standardized management action library (hereinafter referred to as "action library") covering all scenarios of post-diagnosis management is established in advance, including follow-up (routine follow-up, key monitoring follow-up, rehabilitation progress follow-up), guidance (medication guidance, rehabilitation training guidance, dietary guidance), reminder (return visit reminder, medication reminder, examination item reminder), and plan adjustment (drug dosage adjustment suggestions, treatment plan optimization suggestions), etc. Each management action includes a standardized content template (which can be customized and modified).
[0042] Step 202: On the doctor's operation page, output the options corresponding to each recommended action for one or more managed patients.
[0043] In this embodiment, the option can be a shortcut key. An example of the interface layout design for the doctor's operation page is as follows: The patient list page on the doctor's side adopts a display format of "basic patient information + core disease tags + post-visit duration," clearly presenting the key information of each patient; for each patient entry, a left-swipe operation area is set up, displaying the 3-4 highest priority management actions recommended by AI (such as sending follow-up visits, sending follow-up questions, etc.) as shortcut operation options, intuitively displayed in the left-swipe panel. The final left-swipe effect of the doctor's interface is as follows. Figure 3 As shown.
[0044] Step 203: In response to the target patient's target option being selected, load the template content corresponding to the target option from the pre-built action library and display the template content on the doctor's operation page.
[0045] In this embodiment, for example, when a doctor finds a target patient in the patient list, swiping left on the patient entry will display quick operation options. Clicking on any option will quickly initiate the corresponding management action: (1) Send follow-up: The follow-up template corresponding to this action (including AI-generated personalized follow-up questions) will be automatically loaded. Doctors can send it directly or manually modify and supplement it before sending. (2) Send follow-up questions: Doctors can select and send the questions to the patient's mobile device with one click; (3) Other quick actions (such as rehabilitation guidance, follow-up visit reminders, etc.): all have corresponding standardized templates loaded, supporting one-click execution or quick editing before execution.
[0046] Step 204: In response to the confirmation of the template content, push the template content to the target patient's terminal.
[0047] In this embodiment, doctors can directly use the template content or re-edit the template content, and then click confirm to push the template content to the target patient's terminal.
[0048] The method provided in the above embodiments of this disclosure can achieve the following technical effects: (1) Realize personalized management of patients after diagnosis: Through AI-based accurate classification and demand identification, customized management solutions are provided for patients with different disease types, rehabilitation stages, and risk levels, avoiding the blindness of generalized management, improving the pertinence of management measures, and effectively promoting patient recovery.
[0049] (2) Significantly simplify doctors' operation process: The combination of quick operation design and AI intelligent recommendation enables doctors to quickly initiate follow-up visits, send follow-up questions and other management actions without having to search and switch modules. This reduces the management operation time for a single patient to less than 1 minute, significantly improving management efficiency and reducing the workload of doctors.
[0050] (3) Improve the timeliness and effectiveness of management: AI analyzes patient data in real time and recommends management actions. Combined with the early warning and reminder function, it ensures that doctors can promptly capture changes in the patient's condition and take intervention measures to avoid management delays and reduce the risk of relapse or aggravation of the patient's condition.
[0051] In some optional implementations of this embodiment, determining at least one recommended action for each managed patient based on their respective medical data includes: for each managed patient, performing the following steps: determining the patient's management category using a pre-trained artificial intelligence model based on the patient's medical data; determining the patient's management needs using an artificial intelligence model based on the medical data and management category; and determining at least one recommended action for the patient using an artificial intelligence model based on the patient's management category and management needs.
[0052] We employ a deep learning model (i.e., an artificial intelligence model, or simply an "AI model") specifically designed for the medical field. This model is trained using large-scale post-diagnosis management data (covering patient data across multiple disease types and different rehabilitation stages) to optimize its ability to identify disease characteristics, rehabilitation patterns, and management needs, ensuring the accuracy of classification and needs identification.
[0053] AI models extract core features from patient data (disease type, disease stage, treatment method, age, underlying health status, etc.) and classify patients into different management categories. For example, by disease type, patients can be classified as hypertension, diabetes, postoperative rehabilitation, etc.; by rehabilitation stage, patients can be classified as acute recovery, stable management, long-term maintenance, etc.; and by risk level, patients can be classified as low-risk routine management, medium-risk key attention, and high-risk close monitoring, etc.
[0054] Based on management categories and combined with individual patient data, the AI model further identifies each patient's specific management needs, including: whether regular follow-up monitoring is needed, whether specific details of the condition need to be inquired about (such as postoperative wound healing and drug side effects), whether the treatment plan needs to be adjusted, whether supplementary rehabilitation guidance is needed, and whether reminders for follow-up visits are needed, thus forming a personalized management needs list for each patient.
[0055] The AI model intelligently recommends appropriate management actions based on the patient's management category and personalized needs list, combined with medical guidelines, clinical pathways, and historical management experience data.
[0056] In some optional implementations of this embodiment, the patient's management category includes a primary classification, a secondary classification, and a tertiary subclassification; and the patient's management category is determined based on the patient's medical data using a pre-trained artificial intelligence model, including: determining the patient's core characteristics based on the patient's medical data using a pre-trained artificial intelligence model, wherein the core characteristics include at least one of the following: comprehensive risk coefficient characteristics, treatment response characteristics, and disease stability characteristics; determining the primary classification of the disease type dimension based on the patient's medical data using an artificial intelligence model; determining the secondary classification of the rehabilitation stage dimension based on the primary classification and the patient's medical data using an artificial intelligence model; and determining the tertiary subclassification of the risk and management intensity dimension based on the patient's core characteristics using an artificial intelligence model.
[0057] Feature engineering optimizations can be performed: for example, a "comprehensive risk coefficient" feature can be constructed based on "disease type + age + underlying disease", a "treatment response" feature can be constructed based on "treatment plan + medication adherence", and a "disease stability" feature can be constructed based on "frequency of post-diagnosis symptom changes + amplitude of indicator fluctuations", thereby enhancing the richness of classification dimensions.
[0058] A three-tiered classification structure of "main category - secondary category - subcategory" is adopted to ensure the accuracy and usability of the classification results. Primary Classification (Disease Type Dimension): Patients can be classified into 20+ core categories according to the ICD-11 disease classification standard, including circulatory system diseases, respiratory system diseases, digestive system diseases, orthopedic postoperative rehabilitation, and obstetrics and gynecology diseases, covering mainstream clinical post-diagnosis management scenarios.
[0059] Secondary subcategories (rehabilitation stage dimension): Based on the main category, and combined with the duration of the disease, treatment effect, and symptom improvement, it is further divided into four subcategories: acute recovery (1-2 weeks after surgery, within 1 month after acute disease treatment), stable management (3-6 months after surgery, stable control of chronic disease), long-term maintenance (more than 1 year after surgery, long-term management of chronic disease), and rehabilitation fluctuation (recurrence of disease, abnormal fluctuation of indicators).
[0060] Three-level subcategories (risk and management intensity dimension): Based on comprehensive risk coefficients, disease stability and other characteristics, the categories are further subdivided into low-risk routine management category (stable condition, no underlying diseases, and smooth recovery progress), medium-risk key attention category (underlying diseases, slow recovery speed, and slight fluctuations in indicators), high-risk close monitoring category (multiple organ dysfunction, high risk of postoperative complications, and easy deterioration of condition), and special needs category (pregnant women, the elderly, children, disabled patients and other special populations), so as to achieve precise matching between classification results and management intensity.
[0061] In some optional implementations of this embodiment, the method further includes: in response to detecting an update to the patient's medical data, redetermining the patient's management category based on the updated medical data using an artificial intelligence model.
[0062] Establish a real-time update mechanism for classification results, automatically collect the latest post-diagnosis data of patients (such as symptom feedback, physiological indicators, and medication status), and re-trigger the AI model for classification and evaluation.
[0063] In some optional implementations of this embodiment, the patient's management needs include critical management needs, assisted rehabilitation management needs, and long-term rehabilitation management needs.
[0064] Personalized needs are categorized into three levels: core needs (key management needs), important needs (auxiliary rehabilitation management needs), and potential needs (long-term rehabilitation management needs), forming a structured needs list: Core needs (key management actions that must be met): such as "daily follow-up monitoring of physiological indicators" for high-risk patients, "follow-up inquiry on wound recovery" for patients in the acute postoperative period, and "medication adherence reminders" for patients with chronic diseases.
[0065] Key needs (key measures to assist rehabilitation): such as "personalized rehabilitation training guidance" for patients in the rehabilitation period, "dietary plan adjustment suggestions" for diabetic patients, and "drug side effect monitoring" for patients on long-term medication.
[0066] Potential needs (needs that are easily overlooked but affect long-term rehabilitation): such as "home care operation guidance" for elderly patients, "psychological counseling" for patients with chronic diseases, and "suggestions for social reintegration" for postoperative patients.
[0067] In some optional implementations of this embodiment, the method further includes: verifying the patient's management needs based on historical effect feedback information and / or historical management records on the doctor's end; and correcting the patient's management needs in response to verification failure.
[0068] After the list of requirements is generated, it is verified in two dimensions: first, it is compared with the effective management requirements of similar patients in the past to ensure the universality and feasibility of the requirements; second, it is combined with the doctor's historical management records to screen out the types of requirements that are frequently adopted, so as to improve the matching degree between the requirements and clinical practice.
[0069] In some optional implementations of this embodiment, the options corresponding to each recommended action of one or more managed patients are output on the doctor's operation page, including: in response to the number of recommended actions of the patient exceeding a predetermined number, determining the priority of each recommended action of the patient; and outputting the options of selecting a predetermined number of recommended actions in descending order of priority on the doctor's operation page.
[0070] For each patient, the AI model outputs 3-5 priority-ranked recommended management actions. For example, for a hypertensive patient 3 days after diagnosis, it recommends "send follow-up (monitor blood pressure)", "send follow-up question package", and "set follow-up visit reminder (1 month later)". For a lumbar disc herniation patient 2 weeks after surgery, it recommends "send follow-up (rehabilitation training progress)", "send follow-up question package (wound pain and activity limitation)" and "rehabilitation guidance (lumbar muscle strengthening exercise plan)". The recommendation results are synchronized to the doctor's operation interface.
[0071] A quantitative scoring system can also be established to prioritize recommended actions (Top 3-5), placing the most accurate recommended actions at the top.
[0072] In some optional implementations of this embodiment, determining the priority of each recommended action for the patient includes: for each recommended action of the patient, calculating the semantic matching degree between the recommended action and the patient's management needs; determining a clinical effectiveness indicator based on the improvement rate of rehabilitation effect after the execution of the recommended action according to historical data statistics; determining a timeliness indicator based on the urgency of the recommended action; determining the doctor's preference based on the doctor's historical adoption rate of the recommended action; and determining the priority of the patient's recommended action based on the weighted sum of the semantic matching degree, clinical effectiveness indicator, timeliness indicator, and doctor's preference.
[0073] The scoring dimensions and weights can be as follows (the weights and scores below are merely examples and are not limited to the examples; the weights and scores can be adjusted according to needs): Needs matching degree (weight 0.35): The degree of semantic fit between the action and the patient's core needs. The score for matching the action with the core needs is 1, the score for matching the important needs is 0.7, and the score for matching the potential needs is 0.4.
[0074] Clinical efficacy (weight 0.3): Based on the improvement rate of rehabilitation effect after the implementation of the exercise according to historical data statistics. ≥70% improvement rate gets 1 point, 50%-69% gets 0.8 points, 30%-49% gets 0.5 points, and <30% gets 0.2 points.
[0075] Timeliness requirement (weight 0.2): The urgency of the action. 1 point for execution within 24 hours, 0.7 points for execution within 3 days, 0.4 points for execution within 1 week, and 0.2 points for no specific time limit.
[0076] Doctor preference (weight 0.15): The doctor's historical adoption rate of this action. ≥80% adoption rate gets 1 point, 60%-79% gets 0.8 points, 40%-59% gets 0.5 points, and <40% gets 0.2 points.
[0077] In some optional implementations of this embodiment, the method further includes adjusting the priority of each recommended action by at least one of the following methods: adjusting the priority of each recommended action of the patient based on the patient's feedback on each recommended action; adjusting the priority of each recommended action of the patient based on changes in the patient's management category; adjusting the priority of each recommended action of the patient based on the operation of adjusting the sorting of options.
[0078] The recommendation results can be dynamically adjusted in at least one of the following ways: Adjustments based on patient feedback: If a patient does not respond to a follow-up action, or explicitly states that a certain type of guidance is not applicable (e.g., a vegetarian patient refuses meat diet guidance), the system automatically lowers the recommendation priority of that action, or replaces it with an alternative action (e.g., replacing meat diet guidance with plant protein supplementation guidance).
[0079] Adjustments based on changes in the patient's condition: When the patient's classification result is updated (e.g., from the stable phase to the recovery fluctuation phase), the recommendation algorithm is immediately triggered to recalculate, replace the original set of recommended actions, and add management actions that are suitable for the fluctuation phase (e.g., follow up on the reasons for the fluctuation in the condition and suggestions for adjusting the treatment plan).
[0080] Manual intervention adaptation: It supports doctors to manually adjust the priority of recommended actions or add custom actions. The system records the doctor's intervention behavior and feeds it back to the recommendation algorithm. Subsequently, it automatically adapts the intervention preference to similar patients, realizing a closed-loop iteration of "AI recommendation + manual optimization".
[0081] In some optional implementations of this embodiment, the method further includes: displaying an action management entry on the doctor's operation page; and, in response to detecting that the action management entry has been clicked, displaying a list of actions in the action library and custom configuration options for recommended actions.
[0082] The doctor's operation page can be set up with an "More Operations" entry, through which doctors can view all management actions (including other actions besides the recommended ones) to meet special management needs; it also supports doctors to customize the recommended shortcut operation options (such as adding frequently used actions and deleting infrequently used actions).
[0083] In some optional implementations of this embodiment, the method further includes: acquiring historical medical data of one or more managed patients; receiving post-diagnosis health status information input by each patient; cleaning the historical medical data and health status information; and standardizing the cleaned data to generate treatment data for each patient.
[0084] It can collect and preprocess post-diagnosis patient data. It comprehensively collects relevant data from patients reporting for treatment after their visit, and performs cleaning and standardization processes to provide high-quality data support for subsequent AI analysis. Specific implementation steps: 1. Data Collection: The system automatically collects basic patient information (name, age, gender, contact information), core diagnosis and treatment data (diagnosed disease, course of disease, treatment plan, surgical history, medication records), and post-diagnosis information (diagnosis time, self-reported current physical condition, symptom feedback), among other multi-dimensional data. It also supports patients to submit post-diagnosis health data (such as blood pressure, blood sugar levels, exercise status, dietary records, etc.) independently.
[0085] 2. Data Cleaning: Using a combination of rule engines and AI algorithms, invalid information (such as duplicate records, incorrectly formatted data, and meaningless characters) is removed from the data, data deviations are corrected (such as standardizing the units of physiological indicators such as blood pressure and blood sugar), and irrelevant data (such as past medical history unrelated to the current condition) is eliminated to ensure data accuracy.
[0086] 3. Data Standardization: The cleaned data undergoes format standardization and structuring, including standardization of medical terminology (e.g., unifying "high blood pressure" as "hypertension"), standardization of data fields (e.g., standardizing disease course in "year / month / day" units), and data classification and storage (classifying and archiving data according to dimensions such as basic information, diagnosis and treatment data, and post-diagnosis data), providing a structured data foundation for subsequent AI analysis.
[0087] This application can achieve the following technical effects: (1) Realize personalized management of patients after diagnosis: Through AI-based accurate classification and demand identification, customized management solutions are provided for patients with different disease types, rehabilitation stages, and risk levels, avoiding the blindness of generalized management, improving the pertinence of management measures, and effectively promoting patient recovery.
[0088] (2) Significantly simplify doctors' operation process: The combination of quick operation design and AI intelligent recommendation enables doctors to quickly initiate follow-up visits, send follow-up questions and other management actions without having to search and switch modules. This reduces the management operation time for a single patient to less than 1 minute, significantly improving management efficiency and reducing the workload of doctors.
[0089] (3) Improve the timeliness and effectiveness of management: AI analyzes patient data in real time and recommends management actions. Combined with the early warning and reminder function, it ensures that doctors can promptly capture changes in the patient's condition and take intervention measures to avoid management delays and reduce the risk of relapse or aggravation of the patient's condition.
[0090] Further reference Figure 4 This illustrates a flow 400 of an embodiment of a method for training an artificial intelligence model. The flow 400 of the method for training an artificial intelligence model includes the following steps: Step 401: Obtain sample data.
[0091] In this embodiment, the sample data includes sample diagnosis and treatment data, sample management categories, sample management requirements, and sample actions. The sample data utilizes large-scale post-diagnosis management data, covering multiple disease types (including common conditions such as cardiovascular diseases, endocrine diseases, and orthopedic diseases) and different stages of patient rehabilitation (post-operative rehabilitation, long-term rehabilitation for chronic diseases, and post-acute phase rehabilitation). After data anonymization, deduplication, and completion preprocessing, a valid sample set is formed. Each sample data entry includes sample diagnosis and treatment data, sample management category, sample management requirements, and sample actions. The sample diagnosis and treatment data primarily covers post-diagnosis related information such as post-diagnosis follow-up results, rehabilitation status assessment, and medication usage.
[0092] Step 402: Based on the sample medical data, determine the predictive management category through the initial artificial intelligence model.
[0093] In this embodiment, the initial artificial intelligence model adopts a deep learning model specifically designed for the medical field. This model is optimized for post-diagnosis management scenarios and can better adapt to the feature extraction requirements of post-diagnosis data. Preprocessed sample medical data is input into this initial model, which identifies post-diagnosis-related disease characteristics, recovery status, and other information, outputting a predicted management category to achieve a preliminary classification of the post-diagnosis management scenario.
[0094] Step 403: Based on sample diagnosis and treatment data and predictive management categories, determine predictive management needs through an artificial intelligence model.
[0095] In this embodiment, the aforementioned medical-specific deep learning model is used. Combining sample diagnosis and treatment data with the predicted management category output in step 402, the model further mines the rehabilitation patterns in the data, identifies the specific post-diagnosis management needs of patients with different diseases and at different rehabilitation stages, and outputs predicted management needs to ensure that the identified needs are consistent with the actual post-diagnosis rehabilitation scenario.
[0096] Step 404: Based on the predictive management category and predictive management needs, determine the predictive action using an artificial intelligence model.
[0097] In this embodiment, the dedicated deep learning model continues to be used. Based on the predicted management category obtained in step 402 and the predicted management requirements obtained in step 403, and combined with the conventional processes and standards of post-diagnosis management, the model outputs appropriate predicted actions to provide specific and executable operational guidance for post-diagnosis management.
[0098] Step 405: Adjust the network parameters of the artificial intelligence model based on the differences between the sample management category and the prediction management category, the differences between the sample management requirements and the prediction management requirements, and the differences between the sample actions and the prediction actions.
[0099] In this embodiment, the actual management categories, management needs, and actions in the samples are compared with the predicted results output by the model. The differences between the two are analyzed, and the network parameters of the medical-specific deep learning model are adjusted according to the differences. Through repeated iterative training with a large number of post-diagnosis management samples, the model's ability to identify disease characteristics, rehabilitation patterns, and management needs is continuously optimized, gradually improving the accuracy of model classification and need identification until the model performance reaches the preset standard, thus completing model training.
[0100] In this embodiment, the specific structure of the medical-specific deep learning model and the specific scale of the sample data are illustrative examples. In actual applications, they can be adjusted according to the specific scenario of post-diagnosis management. As long as the functions of each step can be achieved, they all fall within the protection scope of this application.
[0101] Further reference Figure 5 As an implementation of the methods shown in the above figures, this disclosure provides an embodiment of a patient service information push device, which is similar to... Figure 2 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.
[0102] like Figure 5 As shown, the patient service information push device 500 of this embodiment includes: a recommendation unit 501, an output unit 502, a loading unit 503, and a push unit 504. The recommendation unit 501 is configured to determine at least one recommended action for each of the one or more managed patients based on their respective medical data using a pre-trained artificial intelligence model. The output unit 502 is configured to output the options corresponding to each recommended action for the one or more managed patients on the doctor's operation page. The loading unit 503 is configured to load template content corresponding to the selected target option from a pre-built action library and display the template content on the doctor's operation page in response to the selection of the target option by the target patient. The push unit 504 is configured to push the template content to the target patient's terminal in response to confirmation of the template content.
[0103] In this embodiment, the specific processing of the recommendation unit 501, output unit 502, loading unit 503, and push unit 504 of the patient service information push device 500 can be referred to Figure 2 The corresponding steps are 201, 202, 203 and 204 in the embodiment.
[0104] In some optional implementations of this embodiment, the recommendation unit 501 is further configured to perform the following steps for each managed patient: Based on the patient's medical data, determine the patient's management category using a pre-trained artificial intelligence model. Based on the patient's medical data and management category, determine the patient's management needs using an artificial intelligence model. Based on the patient's management category and management needs, determine at least one recommended action for the patient using an artificial intelligence model.
[0105] In some optional implementations of this embodiment, the patient's management category includes a primary classification, a secondary classification, and a tertiary subclassification. The recommendation unit 501 is further configured to: determine the patient's core characteristics based on the patient's medical data using a pre-trained artificial intelligence model, wherein the core characteristics include at least one of the following: comprehensive risk coefficient characteristics, treatment response characteristics, and disease stability characteristics. Based on the patient's medical data, determine the primary classification for the disease type dimension using the artificial intelligence model. Based on the primary classification, determine the secondary classification for the rehabilitation stage dimension using the artificial intelligence model, based on the patient's medical data. Based on the secondary classification, determine the tertiary subclassification for the risk and management intensity dimension using the artificial intelligence model, based on the patient's core characteristics.
[0106] In some optional implementations of this embodiment, the device 500 further includes an updating unit (not shown in the figures): in response to detecting an update to the patient's medical data, the device redetermines the patient's management category based on the updated medical data and through an artificial intelligence model.
[0107] In some optional implementations of this embodiment, the patient's management needs include critical management needs, assisted rehabilitation management needs, and long-term rehabilitation management needs.
[0108] In some optional implementations of this embodiment, the device 500 further includes a verification unit (not shown in the figures): verifying the patient's management needs based on historical effect feedback information and / or historical management records from the doctor's end. In response to verification failure, the patient's management needs are corrected.
[0109] In some optional implementations of this embodiment, the output unit 502 is further configured to: determine the priority of each recommended action of the patient in response to the number of recommended actions exceeding a predetermined number; and output the option to select a predetermined number of recommended actions of the patient in descending order of priority on the doctor's operation page.
[0110] In some optional implementations of this embodiment, the output unit 502 is further configured to: calculate the semantic matching degree between the recommended action and the patient's management needs for each recommended action; determine a clinical effectiveness indicator based on the improvement rate of rehabilitation effect after the execution of the recommended action according to historical data statistics; determine a timeliness indicator based on the urgency of the recommended action; determine the doctor's preference based on the doctor's historical adoption rate of the recommended action; and determine the priority of the recommended action for the patient based on the weighted sum of the semantic matching degree, clinical effectiveness indicator, timeliness indicator, and doctor's preference.
[0111] In some optional implementations of this embodiment, the device 500 further includes an adjustment unit (not shown in the figures), configured to: adjust the priority of each recommended action based on the patient's feedback on each recommended action; adjust the priority of each recommended action based on changes in the patient's management category; and adjust the priority of each recommended action based on the operation of adjusting the order of the adjustment options.
[0112] In some optional implementations of this embodiment, the output unit 502 is further configured to: display an action management entry on the doctor's operation page. In response to detecting that the action management entry has been clicked, a list of actions in the action library and custom configuration options for recommended actions are displayed.
[0113] In some optional implementations of this embodiment, the device 500 further includes a data processing unit (not shown in the accompanying drawings): acquiring historical medical data of one or more managed patients; receiving post-visit health status information input by each patient; cleaning the historical medical data and health status information; and standardizing the cleaned data to generate treatment data for each patient.
[0114] Further reference Figure 6 As an implementation of the methods shown in the above figures, this disclosure provides an embodiment of an apparatus for training an artificial intelligence model, which is similar to... Figure 4 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.
[0115] like Figure 6As shown, the apparatus 600 for training an artificial intelligence model in this embodiment includes: an acquisition unit 601, a category determination unit 602, a requirement determination unit 603, an action determination unit 604, and an adjustment unit 605. The acquisition unit 601 is configured to acquire sample data, which includes sample diagnosis and treatment data, sample management categories, sample management requirements, and sample actions. The category determination unit 602 is configured to determine the predicted management category based on the sample diagnosis and treatment data using an initial artificial intelligence model. The requirement determination unit 603 is configured to determine the predicted management requirement based on the sample diagnosis and treatment data and the predicted management category using an artificial intelligence model. The action determination unit 604 is configured to determine the predicted action based on the predicted management category and the predicted management requirement using an artificial intelligence model. The adjustment unit 605 is configured to adjust the network parameters of the artificial intelligence model based on the differences between the sample management category and the predicted management category, the differences between the sample management requirement and the predicted management requirement, and the differences between the sample action and the predicted action.
[0116] It should be noted that the collection, gathering, updating, analysis, processing, use, transmission, and storage of user personal information involved in this disclosed technical solution all comply with relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. Necessary measures are taken to prevent unauthorized access to user personal information data and to safeguard user personal information security and network security.
[0117] According to embodiments of this disclosure, this disclosure also provides an electronic device and a readable storage medium.
[0118] An electronic device includes: one or more processors; and a storage device having one or more computer programs stored thereon, which, when executed by the one or more processors, cause the one or more processors to implement the method described in process 200 or 400.
[0119] A computer-readable medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method described in process 200 or 400.
[0120] Figure 7A schematic block diagram of an example electronic device 700 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.
[0121] like Figure 7 As shown, device 700 includes a computing unit 701, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 702 or a computer program loaded into random access memory (RAM) 703 from storage unit 708. The RAM 703 may also store various programs and data required for the operation of device 700. The computing unit 701, ROM 702, and RAM 703 are interconnected via bus 704. Input / output (I / O) interface 705 is also connected to bus 704.
[0122] Multiple components in device 700 are connected to I / O interface 705, including: input unit 706, such as keyboard, mouse, etc.; output unit 707, such as various types of monitors, speakers, etc.; storage unit 708, such as disk, optical disk, etc.; and communication unit 709, such as network card, modem, wireless transceiver, etc. Communication unit 709 allows device 700 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.
[0123] The computing unit 701 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 701 performs the various methods and processes described above, such as road planning methods. For example, in some embodiments, the road planning method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 708. In some embodiments, part or all of the computer program may be loaded and / or installed on device 700 via ROM 702 and / or communication unit 709. When the computer program is loaded into RAM 703 and executed by the computing unit 701, one or more steps of the road planning method described above may be performed. Alternatively, in other embodiments, the computing unit 701 may be configured to perform the road planning method by any other suitable means (e.g., by means of firmware).
[0124] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.
[0125] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.
[0126] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
[0127] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).
[0128] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.
[0129] Computer systems can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be servers in distributed systems or servers incorporating blockchain technology. Servers can also be cloud servers, or intelligent cloud computing servers or intelligent cloud hosts with artificial intelligence technology.
[0130] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.
[0131] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.
Claims
1. A method for pushing patient service information, comprising: Based on the individual medical data of one or more managed patients, determine at least one recommended action for each patient; The doctor's operation page displays the options corresponding to each recommended action for one or more managed patients; In response to the selection of a target option by a target patient, the template content corresponding to the target option is loaded from a pre-built action library and displayed on the doctor's operation page; In response to the confirmation of the template content, the template content is pushed to the target patient's terminal.
2. The method according to claim 1, wherein, The determination of at least one recommended action for each patient based on their individual medical data (one or more managed patients) includes: For each managed patient, perform the following steps: Based on the patient's medical data, the management category of the patient is determined through a pre-trained artificial intelligence model; Based on the patient's medical data and management category, the patient's management needs are determined using the aforementioned artificial intelligence model; Based on the patient's management category and management needs, the artificial intelligence model determines at least one recommended action for the patient.
3. The method according to claim 2, wherein, The patient's management categories include primary classification, secondary classification, and tertiary classification. as well as Based on the patient's medical data, a pre-trained artificial intelligence model is used to determine the patient's management category, including: Based on the patient's medical data, the patient's core characteristics are determined through a pre-trained artificial intelligence model. The core characteristics include at least one of the following: comprehensive risk coefficient characteristics, treatment response characteristics, and disease stability characteristics. Based on the patient's medical data, the primary classification of the disease type dimension is determined using the aforementioned artificial intelligence model; Based on the primary classification, and using the patient's medical data, the artificial intelligence model determines the secondary classification of the rehabilitation stage dimension. Based on the secondary subclassification, and using the patient's core characteristics, the artificial intelligence model determines the tertiary subclassification of the risk and management intensity dimension.
4. The method according to claim 2, wherein, The method further includes: In response to the detection of an update to the patient's medical data, the patient's management category is redefined based on the updated medical data and the artificial intelligence model.
5. The method according to claim 2, wherein, The patient's management needs include critical management needs, auxiliary rehabilitation management needs, and long-term rehabilitation management needs.
6. The method according to claim 2, wherein, The method further includes: Based on the patient's historical performance feedback information and / or the doctor's historical management records, the patient's management needs are verified; In response to the validation failure, the patient's management requirements were revised.
7. The method according to claim 2, wherein, The option to output the recommended actions for one or more managed patients on the doctor's operation page includes: In response to the patient having more than a predetermined number of recommended actions, the priority of each recommended action for that patient is determined; The options for selecting a predetermined number of recommended actions for the patient, in descending order of priority, are displayed on the doctor's operation page.
8. The method according to claim 7, wherein, Determining the priority of each recommended action for the patient includes: For each recommended action for the patient, the semantic matching degree between the recommended action and the patient's management needs is calculated; the clinical effectiveness indicator is determined based on the improvement rate of rehabilitation effect after the implementation of the recommended action according to historical data statistics; the timeliness indicator is determined based on the urgency of the recommended action; and the doctor's preference is determined based on the doctor's historical adoption rate of the recommended action. The priority of the recommended action for the patient is determined by a weighted sum of the semantic matching degree, the clinical effectiveness index, the timeliness index, and the physician preference.
9. The method according to claim 7, wherein, The method also includes adjusting the priority of the patient's recommended actions in at least one of the following ways: The priority of each recommended action is adjusted based on the patient's feedback on each action; Adjust the priority of each recommended action for this patient based on the change in the patient's management category; The priority of each recommended action for this patient is adjusted based on the sorting of the adjustment options.
10. The method according to claim 1, wherein, The method further includes: The action management entry is displayed on the doctor's operation page; In response to the detection that the action management entry has been clicked, the list of actions in the action library and the custom configuration options for the recommended actions are displayed.
11. The method according to any one of claims 1-10, wherein, The method further includes: Obtain historical medical records of one or more managed patients; Receive post-diagnosis health status information input by each patient; The historical medical records and health status information are cleaned. The cleaned data is then standardized to generate treatment data for each patient.
12. A method for training an artificial intelligence model, comprising: Acquire sample data, wherein the sample data includes sample diagnosis and treatment data, sample management categories, sample management requirements, and sample actions; Based on the sample medical data, the predictive management category is determined through an initial artificial intelligence model; Based on the sample diagnosis and treatment data and the predictive management category, the predictive management needs are determined through the artificial intelligence model; Based on the predicted management category and the predicted management requirements, the predicted action is determined through the artificial intelligence model; Based on the differences between the sample management category and the prediction management category, the differences between the sample management requirements and the prediction management requirements, and the differences between the sample actions and the prediction actions, the network parameters of the artificial intelligence model are adjusted.
13. A patient service information push device, comprising: The recommendation unit is configured to determine at least one recommended action for each patient based on the medical data of one or more managed patients; The output unit is configured to output the options corresponding to each recommended action for one or more managed patients on the doctor's operation page; The loading unit is configured to load the template content corresponding to the target option from a pre-built action library in response to the selection of the target patient's target option, and display the template content on the doctor's operation page; The push unit is configured to push the template content to the target patient's terminal in response to the confirmation of the template content.
14. An apparatus for training an artificial intelligence model, comprising: The acquisition unit is configured to acquire sample data, wherein the sample data includes sample diagnosis and treatment data, sample management categories, sample management requirements, and sample actions; The category determination unit is configured to determine the predictive management category based on the sample diagnosis and treatment data, using an initial artificial intelligence model. The unit needs to be determined and configured to determine predictive management needs based on the sample diagnosis and treatment data and the predictive management category, using the artificial intelligence model. The action determination unit is configured to determine the predicted action based on the predicted management category and the predicted management requirements, using the artificial intelligence model. The adjustment unit is configured to adjust the network parameters of the artificial intelligence model based on the differences between the sample management category and the prediction management category, the differences between the sample management requirements and the prediction management requirements, and the differences between the sample actions and the prediction actions.
15. An electronic device comprising: One or more processors; Storage device, on which one or more computer programs are stored, When the one or more computer programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-12.
16. A computer-readable medium having a computer program stored thereon, wherein, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-12.
17. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-12.