Intelligent scheduling method and system for remote outpatient service

Abstract

The invention relates to an intelligent scheduling method and system for a remote outpatient service, and the method comprises the following steps: carrying out the information extraction of a remote outpatient service reservation request submitted by a patient, and obtaining a patient information set; performing time sequence correlation analysis on the patients based on the patient information set to obtain a patient illness state emergency degree scoring matrix; performing illness state trajectory analysis on the patient based on the illness state emergency degree scoring matrix of the patient to obtain an illness state evolution trajectory map; performing specialized mapping on a remote outpatient service based on the disease evolution trajectory map to obtain a department triage suggestion list; and performing medical resource matching based on the department triage suggestion list to obtain available online doctors, and seeing a doctor for the patient based on the available online doctors, thereby solving the problem that a simple appointment registration system can only perform preliminary triage according to the chief complaint of the patient or a selected department; and the emergency degree and the development trend of the illness state of the patient cannot be accurately judged.

Description

Technical Field

[0001] The present invention relates to the technical field of telemedicine, and in particular to an intelligent scheduling method and system for remote outpatient clinics. Background Art

[0002] Telemedicine, particularly teleclinics, has flourished in recent years with the rapid development of information technology and medical technology. It transcends the time and space constraints of traditional healthcare, providing patients with more convenient and efficient access to care. This has demonstrated significant advantages, particularly for patients in remote areas or with limited mobility. However, with the surge in demand for teleclinics, the efficient and intelligent scheduling of medical resources to ensure timely and accurate diagnosis and treatment has become a critical issue that needs to be addressed. Traditional teleclinic scheduling methods rely primarily on manual operations, which are inefficient and prone to human error, making them unable to meet the growing demand for telemedicine.

[0003] Most existing remote outpatient scheduling systems lack comprehensive, dynamic assessments of patients' conditions. Simple appointment registration systems only provide preliminary triage based on the patient's chief complaint or department selection, but are unable to accurately assess the urgency and progression of the patient's condition. This can result in critically ill patients not receiving timely treatment, while less seriously ill patients consume excessive medical resources, leading to uneven resource allocation and impacting overall medical efficiency. Furthermore, the lack of intelligent condition analysis and specialist mapping capabilities increases communication costs between doctors and patients and reduces the accuracy of diagnosis and treatment.

[0004] Therefore, developing an intelligent remote outpatient scheduling method that can accurately assess and dynamically track patients' conditions and optimize the allocation of medical resources is of great practical significance for improving the quality of telemedicine services, enhancing medical efficiency, and alleviating the current shortage of medical resources. This requires integrating multi-dimensional data such as patients' medical history, condition description, and expected visit time, and analyzing it with advanced algorithmic models to provide patients with more accurate, efficient, and personalized telemedicine services. Summary of the Invention

[0005] The main purpose of the present invention is to provide an intelligent scheduling method and system for remote outpatient clinics, which solves the technical problem that a simple appointment registration system can only perform preliminary triage based on the patient's chief complaint or selection of department, but cannot accurately judge the urgency and development trend of the patient's condition.

[0006] To achieve the above object, the present invention provides an intelligent scheduling method for remote outpatient clinics, comprising the following steps: Extract information from the remote outpatient appointment request submitted by the patient to obtain a patient information set, where the patient information set includes the patient's basic information, condition description, and expected consultation time; Performing a time series correlation analysis on the patient based on the patient information set to obtain a patient condition urgency score matrix; Performing a disease trajectory analysis on the patient based on the patient's disease urgency score matrix to obtain a disease evolution trajectory map; Based on the disease progression trajectory map, remote outpatient clinics are mapped to specialists to obtain a department triage recommendation list; Medical resources are matched based on the department triage suggestion list to obtain available online physicians, and the patient is treated based on the available online physicians.

[0007] Furthermore, the patient information set is used to perform a temporal correlation analysis on the patient to obtain a patient condition urgency score matrix, including: Extracting symptom features from the patient information set to obtain a condition description feature set, and performing semantic quantization processing on the condition description feature set to obtain a symptom severity vector; Performing time series feature mapping on the symptom severity vector through the medical knowledge graph to obtain a disease progression trend sequence, and performing urgency grading on the disease progression trend sequence to obtain a time series urgency feature; Classifying the patient's special needs based on the temporal urgency feature to obtain a special needs feature vector, and assigning weights to the special needs feature vector to obtain demand priority data; A hierarchical weighted calculation is performed on the demand priority data to obtain a spatiotemporal dimension scoring sequence, and an orthogonal transformation mapping is performed on the spatiotemporal dimension scoring sequence to obtain a patient condition urgency scoring matrix, wherein the patient condition urgency scoring matrix includes a main symptom urgency component, a condition aggravation trend component, a special demand response component, and a timeliness component for medical treatment.

[0008] Furthermore, the patient's condition trajectory analysis is performed based on the patient's condition urgency score matrix to obtain a condition evolution trajectory map, including: Performing multi-dimensional feature fusion processing on the patient's condition urgency score matrix to obtain a comprehensive condition feature vector, and performing time series association mining on the comprehensive condition feature vector based on preset patient historical medical data to obtain a patient condition evolution time series; Conducting network topology construction on the patient's disease progression time series through complex network analysis to obtain a disease progression network map, and performing community discovery calculation on the disease progression network map to obtain a disease progression trajectory clustering result; Performing a dynamic risk assessment on the patient's current condition based on the clustering results of the condition evolution trajectory to obtain the patient's current condition risk level, and performing risk factor decomposition on the patient's current condition risk level to obtain a set of risk contribution factors; The graph structure of the risk contribution factor set is learned through a graph neural network to obtain a disease evolution trajectory map.

[0009] Furthermore, the community discovery calculation is performed on the disease evolution network graph to obtain a disease evolution trajectory clustering result, including: Performing a topological structure analysis on the disease evolution network graph to obtain a node connectivity distribution graph, and performing density calculation on the node connectivity distribution graph to obtain a node aggregation feature matrix; Performing boundary recognition processing on the node cluster feature matrix to obtain a community boundary feature set, and dynamically segmenting the community boundary feature set to obtain a disease progression stage diagram; Performing time series association mining on the disease progression stage diagram to obtain a symptom evolution pattern sequence, and performing similarity measurement on the symptom evolution pattern sequence to obtain a disease trajectory similarity spectrum, wherein the disease trajectory similarity spectrum includes symptom evolution rules, disease progression rate, and treatment effect vector; Hierarchical clustering is performed on the disease trajectory similarity spectrum to obtain a disease evolution trajectory clustering result.

[0010] Furthermore, the topological structure analysis of the disease evolution network map is performed to obtain a node connectivity distribution map, including: Performing node metric statistics on the disease evolution network graph to obtain a symptom node correlation sequence, and performing centrality analysis on the symptom node correlation sequence to obtain a core symptom distribution matrix, wherein the core symptom distribution matrix includes symptom connection frequency, node influence range, and symptom propagation path; Performing hierarchical propagation calculation on the core symptom distribution matrix to obtain a symptom diffusion network diagram, and performing connectivity analysis on the symptom diffusion network diagram to obtain a node connection feature set, wherein the node connection feature set includes symptom association strength, propagation link weight, and node interaction frequency; Performing spatial distribution mapping on the node connection feature set to obtain a symptom spatial distribution spectrum, and performing density aggregation calculation on the symptom spatial distribution spectrum to obtain a regional connectivity map, wherein the regional connectivity map includes the degree of symptom clustering, regional interconnection strength, and node distribution density; A global topology optimization is performed on the regional connectivity graph to obtain a node connectivity distribution graph.

[0011] Furthermore, the remote clinic is mapped to a specialist based on the disease evolution trajectory map to obtain a department triage recommendation list, including: Performing symptom spectrum analysis on the disease evolution trajectory map to obtain a key symptom feature set, and performing specialist knowledge mapping on the remote clinic based on the key symptom feature set to obtain a specialist relevance vector; Calculating the department matching degree of the specialty relevance vector through the department diagnosis and treatment standard library to obtain a department diagnosis and treatment capability evaluation table, and performing specialty resource allocation analysis on the department diagnosis and treatment capability evaluation table to obtain a department diagnosis and treatment load vector; Based on the department diagnosis and treatment load vector, the department diagnosis and treatment effect of historical diagnosis and treatment cases is evaluated to obtain a department diagnosis and treatment effect matrix, and the department diagnosis and treatment effect matrix is ​​subjected to multidimensional cross-validation to obtain a department optimization sequence; Performing a dynamic scheduling analysis of specialist resources on the department's preferred sequence to obtain a department scheduling state diagram, and performing time period matching optimization on the department scheduling state diagram to obtain a department time window sequence; A comprehensive evaluation of department triage plans is performed based on the department time window sequence to obtain a triage priority vector, and the triage priority vector is reconstructed in a list format to obtain a department triage recommendation list.

[0012] Furthermore, matching medical resources based on the department triage suggestion list to obtain available online physicians includes: Matching the department triage suggestion list with physician qualifications to obtain a professional qualification compliance table, and performing professional dimension analysis on the professional qualification compliance table to obtain a physician professional capability vector; Performing time window screening on the physician professional capability vector through the physician scheduling database to obtain a physician available time matrix, and performing workload calculation on the physician available time matrix to obtain a physician work saturation vector; Based on historical medical records, the physician's work saturation vector is evaluated for medical quality to obtain a physician's medical effect sequence, and the physician's medical effect sequence is quantified in multiple dimensions to obtain a physician's comprehensive evaluation vector; The comprehensive evaluation vectors of the physicians are prioritized to obtain a physician recommendation list, and the physician recommendation list is dynamically checked to obtain available online physicians, wherein the available online physicians include current online status, remaining consultation capacity, expected response time and remote consultation method.

[0013] The present invention also provides an intelligent scheduling system for remote outpatient clinics, comprising: An extraction module is used to extract information from the remote outpatient appointment request submitted by the patient to obtain a patient information set, wherein the patient information set includes the patient's basic information, condition description and expected consultation time; An association analysis module is used to perform a time series association analysis on the patient based on the patient information set to obtain a patient condition urgency score matrix; A trajectory analysis module is used to analyze the patient's condition trajectory based on the patient's condition urgency scoring matrix to obtain a condition evolution trajectory map; A mapping module, configured to map remote clinic specialists based on the disease progression trajectory map to obtain a department triage recommendation list; A matching module is used to match medical resources based on the department triage suggestion list, obtain available online physicians, and provide medical treatment for the patient based on the available online physicians.

[0014] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any one of the above methods when executing the computer program.

[0015] The present invention also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of any of the above methods are implemented.

[0016] The present invention provides an intelligent scheduling method for remote outpatient clinics, comprising the following steps: extracting information from remote outpatient clinic appointment requests submitted by patients to obtain a patient information set; performing temporal correlation analysis on the patients based on the patient information set to obtain a patient condition urgency score matrix; performing condition trajectory analysis on the patients based on the patient condition urgency score matrix to obtain a condition evolution trajectory map; mapping the remote outpatient clinic to specialists based on the condition evolution trajectory map to obtain a department triage recommendation list; matching medical resources based on the department triage recommendation list to obtain available online physicians; and treating the patient based on the available online physicians. This method solves the technical problem that simple appointment registration systems can only perform preliminary triage based on the patient's chief complaint or department selection, but cannot accurately determine the urgency and development trend of the patient's condition. Instead, it implements specialist mapping based on the condition evolution trajectory map, enabling more accurate patient triage to the appropriate department and avoiding delays in treatment due to patients selecting the wrong department. The generation of the department triage recommendation list provides patients and physicians with the beneficial effect of clearer medical guidance. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a schematic diagram of the steps of an intelligent scheduling method for remote outpatient clinics in one embodiment of the present invention; Figure 2 This is a structural block diagram of an intelligent scheduling system for remote outpatient clinics in one embodiment of the present invention; Figure 3 It is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0018] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0020] like Figure 1 As shown, Figure 1 This is a schematic diagram of the steps of an intelligent scheduling method for remote outpatient clinics in one embodiment of the present invention; In one embodiment of the present invention, a method for intelligent scheduling of remote outpatient services is provided, comprising the following steps: Step S1: extract information from the remote outpatient appointment request submitted by the patient to obtain a patient information set, wherein the patient information set includes the patient's basic information, condition description, and expected consultation time.

[0021] Specifically, to implement the step of extracting information from patient-submitted remote outpatient appointment requests to obtain a patient information set, which includes basic patient information, a description of their condition, and their desired appointment time, an information extraction module must be built to automatically extract the required information from these requests. First, patients can submit appointment requests in a variety of ways, such as through web forms, mobile apps, or WeChat mini-programs. The information extraction module needs to be compatible with these diverse submission methods and accurately identify and parse the information contained therein. Second, basic patient information typically includes name, gender, age, contact information, and medical record number, which can be obtained through structured form input. However, condition descriptions are typically in text form, and patients may describe their condition in their own words, such as, "I've had a headache for the past few days, accompanied by nausea and vomiting." The information extraction module needs to use natural language processing (NLP) technology to analyze the text, extract key symptom information such as "headache," "nausea," and "vomiting," and store it in a structured manner. Furthermore, the desired appointment time can be directly obtained through a date and time selector or inferred from the patient's text description. For example, if a patient says, "I hope to see a doctor tomorrow morning," the information extraction module needs to understand the meaning of this sentence and convert it into a specific date and time. For example, a patient submits a remote outpatient appointment request through a mobile app, entering their name as Zhang San, gender as male, age as 30, contact information as 138xxxxxxxx, and medical record number as 123456. In their medical description, they write, "I have experienced chest tightness for the past week, especially at night when sleeping. I hope to schedule a doctor's appointment as soon as possible, preferably tomorrow morning." They also select the desired appointment time as 9:00 AM on October 28, 2024. The information extraction module extracts this information to form a patient information set, which contains the patient's basic information (name: Zhang San, gender: male, age: 30, contact information: 138xxxxxxxx, medical record number: 123456), a description of their condition (chest tightness, one week, worse at night), and a desired appointment time (9:00 AM on October 28, 2024). This information set will serve as input for subsequent steps to assess the urgency of the condition and match medical resources.

[0022] Step S2: performing a time series correlation analysis on the patient based on the patient information set to obtain a patient condition urgency score matrix.

[0023] Specifically, to implement the step of "performing a temporal correlation analysis on the patient information set based on the patient information set to generate a patient condition urgency score matrix," a temporal correlation analysis module needs to be constructed. This module receives the patient information set obtained from the information extraction module as input and uses a temporal analysis algorithm to analyze the patient's condition. Information in the patient information set, such as the condition description and expected visit time, includes a time dimension. The temporal correlation analysis module arranges this information in chronological order and analyzes the correlations between them. For example, if a patient's condition description includes symptoms such as "worsening cough" or "dyspnea," and these symptoms show a trend of gradually worsening over time, the temporal correlation analysis module will identify this trend and use it as an important basis for determining the urgency of the condition. Furthermore, the patient's medical history needs to be incorporated into the temporal correlation analysis. For example, if a patient has a history of asthma and their cough has recently worsened, the system will consider the patient's condition to be more urgent. By performing temporal correlation analysis on the patient information set, the system generates a patient condition urgency score matrix. This matrix can be a multidimensional matrix, for example, containing information about a patient's various symptoms, signs, and medical history, as well as their severity scores at different time points. For example, if a patient's information set shows a worsening cough over the past week, accompanied by a low-grade fever, and a history of chronic bronchitis, the temporal association analysis module might generate a scoring matrix in which the cough severity score gradually increases over the past week, the low-grade fever score is also high, and the history of chronic bronchitis further increases the overall urgency score. This scoring matrix will serve as input for subsequent disease trajectory analysis. Suppose a patient's condition is described as "low-grade fever started three days ago, cough started yesterday, and chest tightness and difficulty breathing today," and they provide a history of asthma. The temporal association analysis module will analyze the chronological order (three days ago, yesterday, today) and symptom severity (low-grade fever, cough, chest tightness, difficulty breathing), combined with their asthma history, to produce a scoring matrix indicating the patient's high urgency and the need for prompt medical attention.

[0024] Step S3: analyzing the patient's disease trajectory based on the patient's disease urgency score matrix to obtain a disease evolution trajectory map.

[0025] Specifically, to implement the step of "analyzing the patient's condition trajectory based on the patient's condition urgency score matrix to generate a condition evolution trajectory map," a condition trajectory analysis module needs to be constructed. This module receives the patient's condition urgency score matrix generated by the time series correlation analysis module as input and uses data visualization techniques to display the patient's condition evolution trend in the form of a map. The condition trajectory analysis module analyzes the data in the score matrix. For example, it can connect the symptom severity scores at different time points to form a curve, thereby visually displaying the changing trend of the patient's condition. Furthermore, the condition trajectory analysis module can also mark different patient symptoms and signs with different colors or shapes to provide doctors with a clearer understanding of the patient's overall condition. Through condition trajectory analysis, the system ultimately generates a condition evolution trajectory map. This map can be a dynamic chart, for example, dynamically updating the patient's condition over time and predicting future disease trends. For example, suppose a patient's condition urgency score matrix shows that their cough symptoms have gradually worsened over the past week, accompanied by a low-grade fever, and that shortness of breath has suddenly appeared today. The disease trajectory analysis module generates a disease progression trajectory map based on this data. The cough severity score curve shows an upward trend, the low-grade fever score curve is relatively stable, and the dyspnea score shows a significant peak today. This map helps doctors quickly understand the patient's disease progression and determine whether the condition is worsening. For example, suppose a patient developed a low-grade fever three days ago, developed a cough yesterday, and today experiences chest tightness and dyspnea. They also have a history of asthma. The score matrix generated after time-series correlation analysis indicates that the patient's condition is highly urgent. The disease trajectory analysis module converts this information into a disease progression trajectory map, which clearly shows the temporal trends of symptoms such as low-grade fever, cough, chest tightness, and dyspnea, as well as markers of asthma history. This map allows doctors to intuitively understand the patient's disease progression, enabling more accurate diagnosis and treatment decisions.

[0026] Step S4: mapping the remote clinic to a specialist based on the disease progression trajectory map to obtain a department triage recommendation list.

[0027] Specifically, to implement the step of "mapping remote clinic specialists based on the disease progression trajectory map to generate a list of recommended department triage services," a specialist mapping module is required. This module receives the disease progression trajectory map generated by the disease progression trajectory analysis module as input and uses a pre-established specialist mapping rule base to match the patient's condition with the corresponding specialist. The specialist mapping rule base can contain mappings between various diseases, symptoms, and signs and corresponding departments. For example, if a patient's disease progression trajectory map indicates that their primary symptoms are respiratory symptoms such as cough, sputum production, and chest tightness, the specialist mapping module will map them to the respiratory department based on the rule base. The specialist mapping module also considers factors such as the patient's medical history, age, and gender. For example, if a patient has a history of heart disease, even if their primary symptoms are respiratory, they should be considered for mapping to the cardiology department for further evaluation. Through specialist mapping, the system generates a recommended department triage list, which can include multiple departments and is sorted by recommended priority. For example, for a patient who primarily presents with respiratory symptoms but also has a history of heart disease, the system may include both respiratory medicine and cardiology in the department triage recommendation list, and list one of the departments as the preferred department based on the patient's specific situation. Suppose a patient's disease evolution trajectory map shows that his cough has worsened in the past week, accompanied by a low-grade fever, and the symptoms of difficulty breathing suddenly appeared today. The patient also has a history of asthma. Based on this information and the specialist mapping rule base, the specialist mapping module will map the patient to the respiratory medicine department. Taking into account the history of asthma, it may also recommend a consultation with the allergy department or immunology department, and ultimately generate a department triage recommendation list, such as: respiratory medicine (preferred), allergy department, immunology department. This provides guidance for subsequent matching of medical resources, ensuring that patients can receive appropriate specialist diagnosis and treatment as soon as possible.

[0028] Step S5: Match medical resources based on the department triage suggestion list to obtain available online physicians, and then provide medical treatment to the patient based on the available online physicians.

[0029] Specifically, to implement the steps of "matching medical resources based on the department triage recommendation list, obtaining available online physicians, and matching the patient with these available online physicians," a medical resource matching module is required. This module receives the department triage recommendation list generated by the specialty mapping module as input and, based on information such as the patient's desired appointment time, the physician's specialty, availability, and current availability, searches for suitable online physicians within the recommended departments in the department triage recommendation list. For example, if the patient's desired appointment time is tomorrow morning, the system will prioritize matching with physicians who are available tomorrow morning. The system also considers the physician's specialty. For example, if the patient's condition is related to asthma, the system will prioritize matching with physicians with asthma diagnosis and treatment experience. Furthermore, the system needs to query the physician's current availability in real time to avoid matching patients with physicians who are already fully booked. Through medical resource matching, the system generates a list of available online physicians. The system then selects the most suitable physician from the available list based on specific criteria, such as the physician's experience, rating, and the urgency of the patient's condition. The system will push the patient's information to the selected doctor and arrange online diagnosis and treatment. For example, suppose a patient's department triage recommendation list is: respiratory medicine (preferred), allergy department, immunology department, and the patient hopes to see a doctor tomorrow morning. The medical resource matching module will query the doctors who will see patients in the respiratory medicine, allergy department, and immunology department tomorrow morning, and filter them based on the doctor's professional direction, experience, score, current reception status and other information, and finally obtain a list of available online physicians. Assuming that there are three respiratory physicians and one allergy physician in the list who meet the requirements, the system will give priority to a respiratory physician with asthma diagnosis and treatment experience for the patient for online diagnosis and treatment based on the patient's asthma history and respiratory symptoms. This completes the medical resource matching and online consultation arrangement based on the department triage recommendation list.

[0030] In a specific embodiment, the temporal correlation analysis of the patient based on the patient information set is performed to obtain the patient's condition urgency score matrix, including: Extracting symptom features from the patient information set to obtain a condition description feature set, and performing semantic quantization processing on the condition description feature set to obtain a symptom severity vector; Performing time series feature mapping on the symptom severity vector through the medical knowledge graph to obtain a disease progression trend sequence, and performing urgency grading on the disease progression trend sequence to obtain a time series urgency feature; Classifying the patient's special needs based on the temporal urgency feature to obtain a special needs feature vector, and assigning weights to the special needs feature vector to obtain demand priority data; A hierarchical weighted calculation is performed on the demand priority data to obtain a spatiotemporal dimension scoring sequence, and an orthogonal transformation mapping is performed on the spatiotemporal dimension scoring sequence to obtain a patient condition urgency scoring matrix, wherein the patient condition urgency scoring matrix includes a main symptom urgency component, a condition aggravation trend component, a special demand response component, and a timeliness component for medical treatment.

[0031] Specifically, to achieve "performing temporal correlation analysis on the patient information set based on the patient information set to obtain a patient urgency score matrix," a more sophisticated process is required, consisting of four main steps: symptom feature extraction and semantic quantification, temporal feature mapping and urgency grading, special needs classification and weight assignment, and hierarchical weighted calculation and orthogonal transformation mapping. First, symptom feature extraction is performed on the patient information set, identifying and extracting key symptom information from the patient's description of their condition to form a condition description feature set. For example, if a patient describes "My cough has worsened over the past week, accompanied by a low-grade fever, and today I feel chest tightness and difficulty breathing," the extracted feature set includes cough, low-grade fever, chest tightness, and difficulty breathing. This feature set is then semantically quantified, representing these symptoms numerically to form a symptom severity vector. This vector contains multiple dimensions, such as the intensity of the primary symptom (e.g., cough severity), duration of onset (e.g., cough lasting a week), symptom count (e.g., four symptoms present), and symptom correlation coefficients (e.g., cough and difficulty breathing may be correlated). In this way, we convert the patient's condition description into a numerical vector that can be calculated. Next, we use the medical knowledge graph to perform time-series feature mapping on the symptom severity vector. This involves linking this symptom information with knowledge in the medical knowledge graph, analyzing the possible development trends of the condition, and generating a sequence of disease progression trends. For example, based on symptoms such as coughing and dyspnea, as well as their duration and severity, we can infer that the patient's condition may be developing pneumonia or asthma. This trend sequence is then graded for urgency, resulting in a time-series urgency feature. This feature also includes multiple dimensions, such as symptom exacerbation rate (e.g., cough has gradually worsened over the past week), complication risk index (e.g., dyspnea may lead to other complications), treatment timeliness score (e.g., certain conditions require treatment within a certain timeframe), and disease stability parameter (e.g., whether the condition is stable or continues to worsen). These features reflect the urgency and development trends of the patient's condition. Based on the time-series urgency feature, we then classify patients for special needs. This involves adjusting the urgency level based on individual patient differences, such as age and underlying medical conditions, to generate a special needs feature vector. For example, elderly patients or those with underlying medical conditions deserve priority treatment, even if their symptoms are not severe. This vector is then weighted to generate demand priority data. This data includes multiple dimensions, such as age group weight (e.g., elderly patients have a higher weight), underlying disease coefficient (e.g., patients with heart disease have a higher coefficient), medical convenience factor (e.g., whether the patient can easily access a hospital), and remote diagnosis and treatment adaptability score (e.g., whether the patient is suitable for remote diagnosis and treatment). This data reflects the patient's specific needs and treatment priorities.Finally, a hierarchical weighted calculation is performed on the need priority data. This involves weighted summation of the scores across each dimension to produce a sequence of spatiotemporal scores. This sequence is then subjected to an orthogonal transformation, integrating the scores across the different dimensions to ultimately produce a patient urgency score matrix. This matrix contains multiple components, such as the urgency of the primary symptom (e.g., the urgency of the cough), the tendency of worsening (e.g., whether the condition is rapidly worsening), the response to special needs (e.g., whether the patient's special needs are being met), and the timeliness of medical attention (e.g., whether the patient needs to see a doctor as soon as possible). This matrix comprehensively reflects information on multiple aspects of the patient's condition, including urgency, developmental trends, special needs, and timeliness of medical attention. For example, a patient's condition may be described as "a low-grade fever started three days ago, a cough began yesterday, and today I feel chest tightness and difficulty breathing." The patient also has a history of asthma and is 70 years old. First, symptom features are extracted: low-grade fever, cough, chest tightness, and difficulty breathing. Semantic quantification is then performed. For example, a patient with a low-grade fever of 37.5°C, a moderate cough, severe chest tightness, significant difficulty breathing, a three-day onset duration, four symptoms, and a high correlation between cough and difficulty breathing. Next, a time series feature mapping is performed using the medical knowledge graph. Due to the gradual worsening of symptoms and a history of asthma, the patient may develop a serious respiratory disease. Therefore, the symptom exacerbation rate is high, the complication risk index is high, the treatment timeliness score is high, and the condition stability parameter is low. Special needs classification is then performed. Since the patient is 70 years old and has a history of asthma, the age level weight is high, and the coefficient for underlying diseases is high. Assuming the patient chooses remote diagnosis and treatment, the medical convenience factor is high, and the remote diagnosis and treatment adaptability score is high. Finally, hierarchical weighted calculation and orthogonal transformation mapping are performed to obtain the patient's condition urgency score matrix. The main symptom urgency component is high, the condition exacerbation trend component is high, the special needs response component is high, and the treatment timeliness component is high. This indicates that the patient's condition is urgent and requires timely treatment.

[0032] In a specific embodiment, the patient's condition trajectory analysis is performed based on the patient's condition urgency score matrix to obtain a condition evolution trajectory map, including: Performing multi-dimensional feature fusion processing on the patient's condition urgency score matrix to obtain a comprehensive condition feature vector, and performing time series association mining on the comprehensive condition feature vector based on preset patient historical medical data to obtain a patient condition evolution time series; Conducting network topology construction on the patient's disease progression time series through complex network analysis to obtain a disease progression network map, and performing community discovery calculation on the disease progression network map to obtain a disease progression trajectory clustering result; Performing a dynamic risk assessment on the patient's current condition based on the clustering results of the condition evolution trajectory to obtain the patient's current condition risk level, and performing risk factor decomposition on the patient's current condition risk level to obtain a set of risk contribution factors; The graph structure of the risk contribution factor set is learned through a graph neural network to obtain a disease evolution trajectory map.

[0033] Specifically, to achieve "analyzing the patient's disease trajectory based on the patient's disease urgency score matrix and generating a disease evolution trajectory map," a more sophisticated process is required, consisting of four main steps: multidimensional feature fusion and temporal correlation mining, network topology construction and community discovery calculation, dynamic risk assessment and risk factor decomposition, and graph structure learning and map generation. First, multidimensional feature fusion is performed on the patient's disease urgency score matrix. This involves integrating the various components of the score matrix, such as the primary symptom urgency component, the disease exacerbation trend component, the special needs response component, and the timeliness of the treatment, into a comprehensive disease feature vector. This vector represents the patient's overall condition. Then, temporal correlation mining is performed on this feature vector based on pre-set historical patient visit data. This involves comparing and analyzing the patient's current condition with historical visit data to identify correlations and development trends, thereby generating a time series of the patient's disease evolution. This time series contains information on various aspects, such as disease change trends (e.g., whether the condition is improving or worsening), symptom evolution patterns (e.g., how symptoms change), and treatment response patterns (e.g., how the patient responded to previous treatments). Next, complex network analysis is used to construct a network topology based on the patient's disease progression time series. This represents the patient's disease progression as a network graph, where nodes represent different disease states and edges represent transitions between states. Once the disease progression network graph is generated, community discovery is performed on this graph. This clustering of closely connected nodes in the graph forms distinct communities, each representing a typical disease progression trajectory. Once the disease progression trajectory clustering results are obtained, the patient's current disease state can be compared with these clustering results to determine which evolution trajectory the patient falls into and predict future development trends. Based on the disease progression trajectory clustering results, a dynamic risk assessment is then performed on the patient's current disease state. This involves assessing the patient's current disease risk level based on the disease progression trajectory they fall into. For example, if a patient's disease falls into a high-risk evolution trajectory, their risk level is relatively high. The patient's current disease risk level is then decomposed into risk factors to identify the primary factors contributing to the patient's increased risk and form a set of risk contribution factors. For example, a high risk level may be due to a rapid worsening of symptoms or an underlying condition. Finally, graph neural networks are used to learn the graph structure of the risk contribution factor set. This involves using graph neural networks to learn the relationships between risk factors and their impact on disease progression. The learned graph structure can be used to construct a disease evolution trajectory map, which includes the risk evolution path (e.g., how the disease risk changes), disease development trends (e.g., how the disease may progress), and treatment intervention recommendations (e.g., what treatment measures should be taken to reduce the risk).For example, a patient's condition description is "a low-grade fever started three days ago, a cough developed yesterday, and today I feel chest tightness and difficulty breathing." They also have a history of asthma and are 70 years old. Through the previous steps, we obtained a urgency score matrix for the patient's condition. Now, we perform multi-dimensional feature fusion on this matrix to obtain a comprehensive condition feature vector. Then, combined with the patient's historical medical records, such as a previous hospitalization for an asthma attack, we perform time series association mining to obtain a time series of the patient's condition evolution. This shows that his condition was gradually worsening, with symptoms progressing from a low-grade fever to cough, chest tightness, and difficulty breathing, and that he had been responding poorly to previous asthma treatment. Next, we construct a disease progression network graph through complex network analysis and perform community discovery to cluster the patient's disease trajectory into the "acute asthma exacerbation" community. Based on this clustering result, we perform a dynamic risk assessment on the patient. Due to the rapid worsening of his condition and his history of asthma, he is at a higher risk. We then perform risk factor decomposition to obtain a set of risk contributing factors, such as a history of asthma, difficulty breathing, and advanced age. Finally, the relationship between risk factors is learned through graph neural networks to construct a disease evolution trajectory map, which shows that the patient's disease risk is increasing and may develop severe respiratory failure. Treatment intervention recommendations are given, such as the immediate use of bronchodilators and glucocorticoids, and close monitoring of disease changes.

[0034] In a specific embodiment, performing community discovery calculation on the disease progression network graph to obtain a disease progression trajectory clustering result includes: Performing a topological structure analysis on the disease evolution network graph to obtain a node connectivity distribution graph, and performing density calculation on the node connectivity distribution graph to obtain a node aggregation feature matrix; Performing boundary recognition processing on the node cluster feature matrix to obtain a community boundary feature set, and dynamically segmenting the community boundary feature set to obtain a disease progression stage diagram; Performing time series association mining on the disease progression stage diagram to obtain a symptom evolution pattern sequence, and performing similarity measurement on the symptom evolution pattern sequence to obtain a disease trajectory similarity spectrum, wherein the disease trajectory similarity spectrum includes symptom evolution rules, disease progression rate, and treatment effect vector; Hierarchical clustering is performed on the disease trajectory similarity spectrum to obtain a disease evolution trajectory clustering result.

[0035] Specifically, to perform community discovery on the disease evolution network graph and obtain clustering results for the disease evolution trajectory, we need to have a deep understanding of four key steps: topological structure analysis and density calculation, boundary identification and dynamic segmentation, temporal association mining and similarity measurement, and hierarchical clustering. First, topological structure analysis of the disease evolution network graph involves analyzing the connectivity and distribution between nodes (representing different disease states) in the network graph to obtain a node connectivity distribution map. This map shows which nodes are most closely connected and which nodes are centrally located in the network. Then, density calculation is performed on the node connectivity distribution map to obtain a node clustering feature matrix. This matrix includes multiple dimensions, such as symptom association density (e.g., which symptoms frequently occur together), pathology transmission intensity (e.g., how the disease spreads between different symptoms), and symptom cluster characteristics (e.g., which symptoms form a cluster). These characteristics reflect the correlation and degree of clustering between different disease states. Next, boundary identification is performed on the node clustering feature matrix to identify the boundaries between different communities in the network graph. A community is a collection of nodes in the network graph that are closely connected and have more internal connections than external connections, representing a typical disease evolution trajectory. Boundary identification involves identifying features that distinguish different communities, generating a set of community boundary features. Dynamic segmentation of the community boundary features involves segmenting the network graph into different communities based on boundary features, generating a disease progression diagram. This diagram illustrates the different stages of disease progression and the transition relationships between them. This diagram includes disease progression (e.g., the progression from mild to severe), symptom transition nodes (e.g., which symptoms signal the onset of the next stage), and treatment response cutoffs (e.g., which treatments can alter the disease trajectory). Temporal association mining is then performed on the disease progression diagram, analyzing the temporal relationships between different stages and the patterns of symptom evolution to generate a symptom evolution pattern sequence. This sequence describes how symptoms change over time. Similarity is then measured across the symptom evolution pattern sequences to generate a disease trajectory similarity spectrum. This spectrum displays the degree of similarity between different disease progression trajectories, including symptom evolution patterns (e.g., which trajectories have similar symptom evolution patterns), disease progression rates (e.g., which trajectories progress more rapidly), and treatment effect vectors (e.g., which trajectories respond best to treatment). Finally, we perform hierarchical clustering on the similarity spectrum of disease trajectories. This involves grouping different disease evolution trajectories based on their similarity, resulting in a clustering result. Trajectories with high similarity are grouped together, forming a cluster. Each cluster represents a typical disease evolution pattern. For example, a patient's condition may be described as "a low-grade fever started three days ago, a cough began yesterday, and today I feel chest tightness and difficulty breathing" and have a history of asthma. We have constructed a network map of the patient's disease evolution.Now, we perform a topological analysis on this graph and find that the nodes for "cough," "dyspnea," and "chest tightness" have high connectivity, indicating a strong correlation between these symptoms. Density calculations are then performed to obtain a node cluster feature matrix, showing that these symptoms have high correlation density, high pathological transmission intensity, and form a symptom cluster. Next, boundary identification reveals that the connections between the "low-grade fever" node and the "cough," "dyspnea," and "chest tightness" nodes are relatively weak, forming community boundaries. Dynamic segmentation is then performed to divide the disease progression into two stages: the initial stage is dominated by low-grade fever, while the later stage is dominated by cough, dyspnea, and chest tightness. Time series association mining is performed on the disease progression stage graph, revealing a symptom evolution pattern sequence: low-grade fever -> cough -> chest tightness -> dyspnea. Similarity metrics are then used to compare the disease trajectory of other patients, revealing a high degree of similarity with the "acute asthma exacerbation" trajectory. Finally, hierarchical clustering is performed to classify this patient's disease trajectory into the "acute asthma exacerbation" cluster.

[0036] In a specific embodiment, the topological structure analysis of the disease evolution network graph to obtain a node connectivity distribution graph includes: Performing node metric statistics on the disease evolution network graph to obtain a symptom node correlation sequence, and performing centrality analysis on the symptom node correlation sequence to obtain a core symptom distribution matrix, wherein the core symptom distribution matrix includes symptom connection frequency, node influence range, and symptom propagation path; Performing hierarchical propagation calculation on the core symptom distribution matrix to obtain a symptom diffusion network diagram, and performing connectivity analysis on the symptom diffusion network diagram to obtain a node connection feature set, wherein the node connection feature set includes symptom association strength, propagation link weight, and node interaction frequency; Performing spatial distribution mapping on the node connection feature set to obtain a symptom spatial distribution spectrum, and performing density aggregation calculation on the symptom spatial distribution spectrum to obtain a regional connectivity map, wherein the regional connectivity map includes the degree of symptom clustering, regional interconnection strength, and node distribution density; A global topology optimization is performed on the regional connectivity graph to obtain a node connectivity distribution graph.

[0037] Specifically, to perform topological analysis on the disease evolution network graph and obtain a node connectivity distribution map, we need to break it down into four key steps: node metric statistics and centrality analysis, hierarchical propagation calculation and connectivity analysis, spatial distribution mapping and density clustering calculation, and global topology optimization. First, node metric statistics are performed on the disease evolution network graph. This involves counting the number of occurrences and connectivity relationships of each symptom node (representing a symptom) in the network to obtain a symptom node correlation sequence. This sequence reflects the importance of each symptom in the disease evolution process. Next, centrality analysis is performed on the symptom node correlation sequence to identify which symptom nodes are central to the network and have a greater impact on other symptoms, resulting in a core symptom distribution matrix. This matrix includes multiple dimensions, such as symptom connection frequency (e.g., the number of times a symptom is connected to other symptoms), node influence range (e.g., how many other symptoms a symptom can affect), and symptom propagation path (e.g., how symptoms propagate). These features reveal which symptoms play a key role in disease evolution. Next, hierarchical propagation calculations are performed on the core symptom distribution matrix, simulating the propagation of symptoms in the network and calculating the impact of each symptom on other symptoms, resulting in a symptom diffusion network map. This diagram illustrates how symptoms influence and spread. Connectivity analysis is then performed on the symptom diffusion network diagram. This involves analyzing the connection strength and connectivity between nodes in the network diagram to obtain a node connectivity feature set. This feature set includes multiple dimensions, such as symptom association strength (e.g., the strength of the association between two symptoms), propagation link weight (e.g., the weight of the propagation path between symptoms), and node interaction frequency (e.g., the frequency of interaction between two symptoms). These features reflect the closeness of the interaction between symptoms. Spatial distribution mapping is then performed on the node connectivity feature set. This involves mapping the symptom nodes into a space and arranging them based on their connectivity features to obtain a symptom spatial distribution spectrum. This spectrum visually demonstrates the spatial relationships and clustering between symptoms. Density aggregation calculation is then performed on the symptom spatial distribution spectrum. This involves calculating the node density and connection strength of different regions in the space to obtain a regional connectivity map. This map includes multiple dimensions, such as the degree of symptom clustering (e.g., which symptoms cluster together), regional interconnectivity (e.g., the strength of connections between different regions), and node distribution density (e.g., the node density in different regions). These features reflect the spatial distribution and interaction relationships of symptoms. Finally, we perform global topology optimization on the regional connectivity graph. This involves adjusting the network topology based on the regional connectivity graph and optimizing the connections between nodes, ultimately generating a node connectivity distribution map. This map displays the connectivity of each node and reflects the importance and influence of each symptom within the overall disease evolution network.

[0038] For example, a patient's condition description is "a low-grade fever started three days ago, a cough began yesterday, and today I feel chest tightness and difficulty breathing." They also have a history of asthma. We have constructed a network diagram of the patient's condition evolution. First, we perform node metric statistics to obtain a sequence of symptom node associations. For example, "cough" appears the most frequently and has the highest association. We then perform centrality analysis to obtain a core symptom distribution matrix. For example, "cough" has the highest connection frequency and the widest impact, making it a core symptom. Next, we perform hierarchical propagation calculations to obtain a symptom diffusion network diagram, showing that "cough" can spread to "chest tightness" and "dyspnea." We then perform connectivity analysis to obtain a node connection feature set. For example, "cough" and "dyspnea" have a high correlation strength. We then perform spatial distribution mapping to obtain a symptom spatial distribution spectrum, showing that "cough," "chest tightness," and "dyspnea" are clustered together. We then perform density clustering to obtain a regional connectivity map, showing that the regions containing these three symptoms have the highest connectivity. Finally, we perform global topology optimization to obtain a node connectivity distribution map, showing that the "cough" node has the highest connectivity and is at the center of the network.

[0039] In a specific embodiment, the remote clinic is mapped to a specialist based on the disease evolution trajectory map to obtain a department triage recommendation list, including: Performing symptom spectrum analysis on the disease evolution trajectory map to obtain a key symptom feature set, and performing specialist knowledge mapping on the remote clinic based on the key symptom feature set to obtain a specialist relevance vector; Calculating the department matching degree of the specialty relevance vector through the department diagnosis and treatment standard library to obtain a department diagnosis and treatment capability evaluation table, and performing specialty resource allocation analysis on the department diagnosis and treatment capability evaluation table to obtain a department diagnosis and treatment load vector; Based on the department diagnosis and treatment load vector, the department diagnosis and treatment effect of historical diagnosis and treatment cases is evaluated to obtain a department diagnosis and treatment effect matrix, and the department diagnosis and treatment effect matrix is ​​subjected to multidimensional cross-validation to obtain a department optimization sequence; Performing a dynamic scheduling analysis of specialist resources on the department's preferred sequence to obtain a department scheduling state diagram, and performing time period matching optimization on the department scheduling state diagram to obtain a department time window sequence; A comprehensive evaluation of department triage plans is performed based on the department time window sequence to obtain a triage priority vector, and the triage priority vector is reconstructed in a list format to obtain a department triage recommendation list.

[0040] Specifically, to achieve "specialty mapping for remote clinics based on the disease progression trajectory map and obtaining a list of department triage recommendations," we need to analyze five key steps: symptom spectrum analysis and specialty knowledge mapping, department matching calculation and specialty resource allocation analysis, department diagnosis and treatment effectiveness evaluation and multi-dimensional cross-validation, dynamic specialty resource scheduling analysis and time period matching optimization, and comprehensive evaluation and list reconstruction of triage plans. First, symptom spectrum analysis of the disease progression trajectory map involves analyzing the relationships and evolution patterns between symptoms in the map, extracting key symptom features, and forming a key symptom feature set. For example, if a patient's primary symptoms are cough, dyspnea, and chest tightness, and their condition progresses rapidly, these symptoms constitute the key symptom feature set. Then, specialty knowledge mapping for remote clinics based on the key symptom feature set involves matching these key symptoms with different specialty knowledge, calculating the relevance of each specialty to the patient's condition, and obtaining a specialty relevance vector. This vector includes multiple dimensions, such as the strength of primary symptom attribution (e.g., cough primarily belongs to respiratory medicine), the distribution of complications by specialty (e.g., asthma may be associated with allergy medicine), disease transmission pathways (e.g., the path from respiratory infection to pneumonia), and cross-departmental connections (e.g., the connection between respiratory medicine and cardiovascular medicine). Next, the specialty correlation vector is used to calculate departmental matching using the departmental diagnosis and treatment standard database. This involves evaluating each department's ability to diagnose and treat the patient's condition based on its own diagnosis and treatment standards, resulting in a departmental diagnosis and treatment capacity evaluation table. This departmental resource allocation analysis is then performed on this departmental diagnosis and treatment capacity evaluation table. This involves analyzing each department's allocation of physician and equipment resources to assess its ability to effectively treat the patient's condition, resulting in a departmental diagnosis and treatment load vector. This vector includes multiple dimensions, such as the specialist treatment success rate (e.g., the success rate of asthma treatment in respiratory medicine), remote consultation suitability (e.g., whether the disease is suitable for remote consultation), expert resource utilization (e.g., appointment availability of respiratory medicine specialists), and diagnosis and treatment cycle matching (e.g., whether the disease's diagnosis and treatment cycle is consistent with the department's scheduled work schedule). Next, based on the department's treatment load vector, the department's treatment effectiveness is evaluated for historical treatment cases. This involves analyzing the treatment effectiveness of different departments for similar conditions in historical treatment cases to obtain a department treatment effectiveness matrix. This department treatment effectiveness matrix is ​​then subjected to multidimensional cross-validation, which verifies the department's treatment effectiveness from multiple dimensions, such as cure rate and patient satisfaction, to obtain a department optimization sequence. This sequence is sorted by department treatment effectiveness, with the best departments ranked first. The department optimization sequence includes treatment plan completeness (e.g., whether the department can provide a complete treatment plan), cure rate statistics (e.g., the department's cure rate for similar cases), patient feedback scores (e.g., patients' evaluation of the department), and follow-up effect coefficients (e.g., the department's follow-up effect).Next, the departmental resource dynamic scheduling analysis is performed on the preferred department sequence. This involves analyzing each department's real-time capacity based on information such as physician schedules and equipment availability, resulting in a departmental scheduling statemap. This time-period matching optimization is then performed on the departmental scheduling statemap. This involves selecting appropriate departments and time periods based on the patient's desired visit time, resulting in a departmental time window sequence. This sequence includes multiple dimensions, such as specialist availability (e.g., the time periods during which respiratory specialists are available), remote consultation slots (e.g., the number of slots reserved for remote consultations), emergency slots (e.g., slots reserved for emergency situations), and time-period load rates (e.g., the patient load during different time periods). Finally, a comprehensive evaluation of departmental triage plans is conducted based on the departmental time window sequence. This involves comprehensively evaluating different triage plans based on factors such as the department's diagnostic and treatment capabilities, resource availability, and the patient's desired visit time, resulting in a triage priority vector. This triage priority vector is then reconstructed into a list, sorting the triage plans in order of priority to form a recommended departmental triage list. This list includes multiple dimensions, such as recommended departments (for example, patients are advised to visit the respiratory department), recommended expert list (for example, several respiratory department experts are recommended), appointment time recommendations (for example, patients are advised to make an appointment during a certain time period), and remote consultation methods (for example, video consultation or text consultation can be selected).

[0041] For example, a patient's condition description is "a low-grade fever started three days ago, a cough developed yesterday, and today I feel chest tightness and difficulty breathing." They also have a history of asthma. We have already generated a map of the patient's condition evolution. First, we perform symptom spectrum analysis to obtain the key symptom feature set: cough, difficulty breathing, chest tightness, and asthma. We then perform specialty knowledge mapping to obtain a specialty relevance vector, showing that respiratory medicine has the highest relevance. Next, we calculate department matching using the department's diagnosis and treatment standard database, finding respiratory medicine to be the highest match. We then perform specialty resource allocation analysis to obtain a departmental diagnosis and treatment load vector, showing that respiratory medicine has a high treatment success rate and is highly suitable for remote consultations. We then evaluate departmental treatment effectiveness based on historical diagnosis and treatment cases and perform multi-dimensional cross-validation to determine a preferred department sequence, with respiratory medicine ranking first. We then perform dynamic specialty resource scheduling analysis to obtain a departmental schedule status diagram, showing how many specialists will be available for consultations in the respiratory department tomorrow morning. We then perform time-slot matching optimization to obtain a departmental time window sequence, recommending a remote consultation for the patient tomorrow morning. Finally, a comprehensive evaluation of the triage plan was conducted, and the results were tabulated and reconstructed to obtain a list of department triage recommendations: Recommended department: Respiratory Medicine Department; Recommended experts: Dr. Zhang, Dr. Li; Recommended appointment time: Tomorrow morning; Remote consultation method: Video consultation.

[0042] In a specific embodiment, matching medical resources based on the department triage suggestion list to obtain available online physicians includes: Matching the department triage suggestion list with physician qualifications to obtain a professional qualification compliance table, and performing professional dimension analysis on the professional qualification compliance table to obtain a physician professional capability vector; Performing time window screening on the physician professional capability vector through the physician scheduling database to obtain a physician available time matrix, and performing workload calculation on the physician available time matrix to obtain a physician work saturation vector; Based on historical medical records, the physician's work saturation vector is evaluated for medical quality to obtain a physician's medical effect sequence, and the physician's medical effect sequence is quantified in multiple dimensions to obtain a physician's comprehensive evaluation vector; The comprehensive evaluation vectors of the physicians are prioritized to obtain a physician recommendation list, and the physician recommendation list is dynamically checked to obtain available online physicians, wherein the available online physicians include current online status, remaining consultation capacity, expected response time and remote consultation method.

[0043] Specifically, to achieve "matching medical resources based on the recommended department triage list to obtain available online physicians," we need to break down into four key steps: physician qualification matching and professional dimension analysis, time window screening and workload calculation, diagnosis and treatment quality assessment and multi-dimensional quantification, and priority sorting and dynamic status detection. First, physician qualification matching is performed on the recommended department triage list. This involves matching the recommended departments with the physician's qualifications to generate a professional qualification match table. This table records the degree of match between each physician and the recommended department. Next, the professional qualification match table is analyzed in professional dimensions. This converts the physician's qualification information into quantifiable indicators to generate a physician professional competency vector. This vector includes multiple dimensions, such as specialty qualification level (e.g., chief physician, deputy chief physician, attending physician), remote consultation experience (e.g., number of completed remote consultations), professional skill score (e.g., a score based on the physician's skill level), and diagnosis and treatment coverage (e.g., the disease areas in which the physician specializes). Next, the physician professional capacity vectors are filtered through a time window using the physician scheduling database. This involves selecting physicians available during the recommended consultation time periods based on their schedule information, resulting in a physician availability matrix. This matrix records each physician's available time periods. Workload calculation is then performed on the physician availability matrix, calculating each physician's workload saturation during the recommended consultation time periods, resulting in a physician workload saturation vector. This vector includes multiple dimensions, such as online consultation capacity (e.g., the number of online consultations available simultaneously), appointment queue length (e.g., the number of patients currently waiting for consultation), average consultation duration (e.g., the average duration of each consultation), and work period distribution (e.g., the physician's work schedule at different times of the day). The physician workload saturation vector is then evaluated for treatment quality based on historical medical records. This evaluation evaluates the physician's treatment quality based on their historical medical records, resulting in a physician performance sequence. This sequence is then ranked according to the physician's treatment performance. This physician performance sequence is then quantified across multiple dimensions, converting the physician's treatment performance into quantifiable indicators to produce a comprehensive physician evaluation vector. This vector includes multiple dimensions, such as patient satisfaction (e.g., patient evaluation of the physician), diagnostic accuracy (e.g., the accuracy of the physician's diagnosis), prescription rationality (e.g., the rationality of the physician's prescription), and follow-up completion rate (e.g., the percentage of follow-up visits completed by the physician). The physician's comprehensive evaluation vector is then prioritized. This means that physicians are ranked based on their comprehensive evaluations to create a recommended physician list. Physicians ranked higher in the recommended list have higher comprehensive evaluations. Finally, the recommended physician list undergoes dynamic status monitoring, which means that the physician's online status and ability to receive consultations are monitored in real time to identify available online physicians.Information about available online physicians includes their current online status (e.g., online or offline), remaining capacity (e.g., how many patients can still be seen), expected response time (e.g., how long patients will have to wait for a physician's response), and the remote consultation method (e.g., video, text, or voice). For example, a patient's medical condition may include "a low-grade fever started three days ago, a cough began yesterday, and today chest tightness and difficulty breathing" with a history of asthma. The recommended department triage list recommends respiratory medicine, with a recommended consultation time of tomorrow morning. First, physician qualifications are matched to identify respiratory physicians. Professional dimension analysis is then performed to obtain a physician's professional competency vector. Next, a time window is filtered using the physician scheduling database to identify respiratory physicians available tomorrow morning. Workload calculation is then performed to obtain a physician workload saturation vector. Next, treatment quality is assessed based on historical medical records and quantified across multiple dimensions to obtain a comprehensive physician evaluation vector. Priority sorting is then performed to obtain a recommended physician list. Finally, dynamic status detection is performed to obtain available online doctors. For example, Dr. Zhang is currently online, the remaining consultation capacity is 5 people, the expected response time is 10 minutes, and the remote consultation method is video consultation.

[0044] The above describes the intelligent scheduling method of the remote clinic in the embodiment of the present invention. The following describes the intelligent scheduling system of the remote clinic in the embodiment of the present invention. Figure 2 In one embodiment of the present invention, an intelligent scheduling system for remote outpatient services includes: Extraction module 21, for extracting information from the remote outpatient appointment request submitted by the patient to obtain a patient information set, wherein the patient information set includes the patient's basic information, condition description and expected consultation time; The correlation analysis module 22 is used to perform a time series correlation analysis on the patient based on the patient information set to obtain a patient condition urgency score matrix; A trajectory analysis module 23 is configured to perform a trajectory analysis on the patient's condition based on the patient's condition urgency score matrix to obtain a condition evolution trajectory map; A mapping module 24 is configured to map remote clinic specialists based on the disease progression trajectory map to obtain a department triage recommendation list; The matching module 25 is configured to match medical resources based on the department triage suggestion list, obtain available online physicians, and provide medical treatment to the patient based on the available online physicians.

[0045] In this embodiment, for the specific implementation of each unit in the above system embodiment, please refer to the above method embodiment, which will not be repeated here.

[0046] Reference Figure 3In an embodiment of the present invention, a computer device is also provided, wherein the internal structure of the computer device can be as follows: Figure 3 As shown. The computer device includes a processor, memory, display screen, input device, network interface and database connected via a system bus. The processor of the computer design is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store the corresponding data in this embodiment. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the above method is implemented.

[0047] Those skilled in the art will understand that Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present invention and does not constitute a limitation on the computer device to which the solution of the present invention is applied.

[0048] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which implements the above-described method when executed by a processor. It is understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0049] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware using a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the above-described method embodiments. Any reference to memory, storage, database, or other media provided herein and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM.

[0050] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, apparatus, article, or method comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or method. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, apparatus, article, or method comprising the element.

[0051] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. An intelligent scheduling method for remote outpatient clinics, characterized in that: The following steps are involved: Extract information from the remote outpatient appointment request submitted by the patient to obtain a patient information set, where the patient information set includes the patient's basic information, condition description, and expected consultation time; Performing a time series correlation analysis on the patient based on the patient information set to obtain a patient condition urgency score matrix; Performing a disease trajectory analysis on the patient based on the patient's disease urgency score matrix to obtain a disease evolution trajectory map; Based on the disease progression trajectory map, remote outpatient clinics are mapped to specialists to obtain a department triage recommendation list; Medical resources are matched based on the department triage suggestion list to obtain available online physicians, and the patient is treated based on the available online physicians.

2. The intelligent scheduling method for remote outpatient clinic according to claim 1, characterized in that: The step of performing temporal correlation analysis on the patient based on the patient information set to obtain a patient condition urgency score matrix includes: Extracting symptom features from the patient information set to obtain a condition description feature set, and performing semantic quantization processing on the condition description feature set to obtain a symptom severity vector; Performing time series feature mapping on the symptom severity vector through the medical knowledge graph to obtain a disease progression trend sequence, and performing urgency grading on the disease progression trend sequence to obtain a time series urgency feature; Classifying the patient's special needs based on the temporal urgency feature to obtain a special needs feature vector, and assigning weights to the special needs feature vector to obtain demand priority data; A hierarchical weighted calculation is performed on the demand priority data to obtain a spatiotemporal dimension scoring sequence, and an orthogonal transformation mapping is performed on the spatiotemporal dimension scoring sequence to obtain a patient condition urgency scoring matrix, wherein the patient condition urgency scoring matrix includes a main symptom urgency component, a condition aggravation trend component, a special demand response component, and a timeliness component for medical treatment.

3. The intelligent scheduling method for remote outpatient clinic according to claim 1, characterized in that: The step of performing a disease trajectory analysis on the patient based on the patient disease urgency score matrix to obtain a disease evolution trajectory map includes: Performing multi-dimensional feature fusion processing on the patient's condition urgency score matrix to obtain a comprehensive condition feature vector, and performing time series association mining on the comprehensive condition feature vector based on preset patient historical medical data to obtain a patient condition evolution time series; Conducting network topology construction on the patient's disease progression time series through complex network analysis to obtain a disease progression network map, and performing community discovery calculation on the disease progression network map to obtain a disease progression trajectory clustering result; Performing a dynamic risk assessment on the patient's current condition based on the clustering results of the condition evolution trajectory to obtain the patient's current condition risk level, and performing risk factor decomposition on the patient's current condition risk level to obtain a set of risk contribution factors; The graph structure of the risk contribution factor set is learned through a graph neural network to obtain a disease evolution trajectory map.

4. The intelligent scheduling method for remote outpatient clinic according to claim 3, characterized in that: The community discovery calculation is performed on the disease evolution network graph to obtain a disease evolution trajectory clustering result, including: Performing a topological structure analysis on the disease evolution network graph to obtain a node connectivity distribution graph, and performing density calculation on the node connectivity distribution graph to obtain a node aggregation feature matrix; Performing boundary recognition processing on the node cluster feature matrix to obtain a community boundary feature set, and dynamically segmenting the community boundary feature set to obtain a disease progression stage diagram; Performing time series association mining on the disease progression stage diagram to obtain a symptom evolution pattern sequence, and performing similarity measurement on the symptom evolution pattern sequence to obtain a disease trajectory similarity spectrum, wherein the disease trajectory similarity spectrum includes symptom evolution rules, disease progression rate, and treatment effect vector; Hierarchical clustering is performed on the disease trajectory similarity spectrum to obtain a disease evolution trajectory clustering result.

5. The intelligent scheduling method for remote outpatient clinic according to claim 4, characterized in that: The topological structure analysis of the disease evolution network graph is performed to obtain a node connectivity distribution graph, including: Performing node metric statistics on the disease evolution network graph to obtain a symptom node correlation sequence, and performing centrality analysis on the symptom node correlation sequence to obtain a core symptom distribution matrix, wherein the core symptom distribution matrix includes symptom connection frequency, node influence range, and symptom propagation path; Performing hierarchical propagation calculation on the core symptom distribution matrix to obtain a symptom diffusion network diagram, and performing connectivity analysis on the symptom diffusion network diagram to obtain a node connection feature set, wherein the node connection feature set includes symptom association strength, propagation link weight, and node interaction frequency; Performing spatial distribution mapping on the node connection feature set to obtain a symptom spatial distribution spectrum, and performing density aggregation calculation on the symptom spatial distribution spectrum to obtain a regional connectivity map, wherein the regional connectivity map includes the degree of symptom clustering, regional interconnection strength, and node distribution density; A global topology optimization is performed on the regional connectivity graph to obtain a node connectivity distribution graph.

6. The intelligent scheduling method for remote outpatient clinic according to claim 1, characterized in that: The remote clinic is mapped to a specialist based on the disease evolution trajectory map to obtain a department triage recommendation list, including: Performing symptom spectrum analysis on the disease evolution trajectory map to obtain a key symptom feature set, and performing specialist knowledge mapping on the remote clinic based on the key symptom feature set to obtain a specialist relevance vector; Calculating the department matching degree of the specialty relevance vector through the department diagnosis and treatment standard library to obtain a department diagnosis and treatment capability evaluation table, and performing specialty resource allocation analysis on the department diagnosis and treatment capability evaluation table to obtain a department diagnosis and treatment load vector; Based on the department diagnosis and treatment load vector, the department diagnosis and treatment effect of historical diagnosis and treatment cases is evaluated to obtain a department diagnosis and treatment effect matrix, and the department diagnosis and treatment effect matrix is ​​subjected to multidimensional cross-validation to obtain a department optimization sequence; Performing a dynamic scheduling analysis of specialist resources on the department's preferred sequence to obtain a department scheduling state diagram, and performing time period matching optimization on the department scheduling state diagram to obtain a department time window sequence; A comprehensive evaluation of department triage plans is performed based on the department time window sequence to obtain a triage priority vector, and the triage priority vector is reconstructed in a list format to obtain a department triage recommendation list.

7. The intelligent scheduling method for remote outpatient clinic according to claim 1, characterized in that: The matching of medical resources based on the department triage suggestion list to obtain available online physicians includes: Matching the department triage suggestion list with physician qualifications to obtain a professional qualification compliance table, and performing professional dimension analysis on the professional qualification compliance table to obtain a physician professional capability vector; Performing time window screening on the physician professional capability vector through the physician scheduling database to obtain a physician available time matrix, and performing workload calculation on the physician available time matrix to obtain a physician work saturation vector; Based on historical medical records, the physician's work saturation vector is evaluated for medical quality to obtain a physician's medical effect sequence, and the physician's medical effect sequence is quantified in multiple dimensions to obtain a physician's comprehensive evaluation vector; The comprehensive evaluation vectors of the physicians are prioritized to obtain a physician recommendation list, and the physician recommendation list is dynamically checked to obtain available online physicians, wherein the available online physicians include current online status, remaining consultation capacity, expected response time and remote consultation method.

8. An intelligent scheduling system for remote outpatient clinics, characterized in that: include: An extraction module is used to extract information from the remote outpatient appointment request submitted by the patient to obtain a patient information set, wherein the patient information set includes the patient's basic information, condition description and expected consultation time; An association analysis module is used to perform a time series association analysis on the patient based on the patient information set to obtain a patient condition urgency score matrix; A trajectory analysis module is used to analyze the patient's condition trajectory based on the patient's condition urgency scoring matrix to obtain a condition evolution trajectory map; A mapping module, configured to map remote clinic specialists based on the disease progression trajectory map to obtain a department triage recommendation list; A matching module is used to match medical resources based on the department triage suggestion list, obtain available online physicians, and provide medical treatment for the patient based on the available online physicians.

9. A computer device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images