Doctor resource matching method, system and equipment based on vital sign data

Abstract

The invention relates to the technical field of medical data information, and particularly discloses a doctor resource matching method, system and equipment based on vital sign data, and the method mainly comprises the steps: receiving patient data sent by a patient end, the patient data comprising the vital sign data of a patient; inputting the patient data into a deep learning-based model to obtain assessment data of the patient, the assessment data including disease type probability distribution; and acquiring doctor end information of the online doctor ends, matching the disease type probability distribution of the patient with the doctor end information to obtain a total matching degree of each online doctor end, and selecting a plurality of doctors with the highest total matching degree as a preliminary matching result. According to the method, the illness sign data are uploaded to the background, doctors are automatically matched, remote treatment intervention is carried out, remote communication and treatment between the doctors and the patients are achieved, the utilization efficiency of medical resources is improved, the waiting time of the patients is shortened, and meanwhile it is guaranteed that the patients obtain professional medical services in time.

Description

Technical Field

[0001] This application relates to the field of medical data information technology, and particularly to a doctor resource matching method based on vital sign data, a doctor resource matching system based on vital sign data, and a doctor resource matching device based on vital sign data. Background Art

[0002] Regular physical examinations, maintaining good living habits and eating habits, and actively dealing with physical discomfort are all manifestations of paying attention to one's own health. Seeking medical treatment in a timely manner means that when physical abnormalities or disease symptoms are found, actively seeking medical help without delay. This health awareness helps to detect and treat diseases at an early stage, reduce the risk of disease deterioration, and improve the treatment effect.

[0003] In the traditional medical system, patients usually need to judge their own symptoms by themselves and try to find a doctor in the corresponding specialty for treatment. However, this method has some challenges. First, patients lack medical expertise and clinical experience, and self-diagnosis may have the risk of misdiagnosis, resulting in incorrect doctor matching and treatment plans. Second, patients need to spend a lot of time and energy searching and screening doctors, which may lead to delays in treatment, especially in emergency situations. In addition, there may be inaccurate matching problems when patients choose doctors by themselves, because patients may not be able to fully understand the professional fields and experience backgrounds of all doctors. Therefore, self-diagnosis and self-selection of doctors have the disadvantages of low diagnostic accuracy, high consumption of time and energy, and inaccurate doctor matching. Summary of the Invention

[0004] By providing a doctor resource matching method, system and device based on vital sign data, this application solves the technical problems in the prior art that patients usually need to evaluate their own symptoms by themselves in order to find a doctor in the corresponding specialty for treatment, resulting in low diagnostic accuracy, high consumption of time and energy, and inaccurate doctor matching. It can improve the accuracy of medical diagnosis, optimize the doctor matching process and accelerate the doctor response time, so as to better meet the needs of patients for medical services.

[0005] In the first aspect, this application provides a doctor resource matching method based on vital sign data, which includes the following steps: receiving patient data sent by the patient terminal, where the patient data includes the vital sign data of the patient; inputting the patient data into a deep learning model to obtain the evaluation data of the patient, and the evaluation data includes the disease type probability distribution. Obtaining the doctor terminal information of the online doctor terminal, and the disease type probability distribution of the patient. Match with the doctor - side information to obtain the total matching degree of each online doctor - side, and select several doctors with the highest total matching degree as the preliminary matching result; send a diagnosis request to the doctor - side based on the preliminary matching result.

[0006] Further, receive the response information returned by the doctor - side according to the diagnosis request.

[0007] Further, the doctor - side information includes the professional field S c and the cross - field. S c =S 1 ,S 2 ,…,S m , where S m represents the m - th professional field; represents the probability of the n - th disease type D b ; Matching the probability distribution of the patient's disease type with the doctor - side information to obtain the total matching degree of each online doctor - side specifically includes: calculating the matching degree M(Doctor a and the disease type D b that the patient may have), M(Doctor a ,P(D b )) is: M(Doctor a ,P(D b ))=α*D a ,P(D b ))=α*D b *E(Doctor a ,S c )+β*C(Doctor a ,D b ); In the formula, M(Doctor a ,D b ) represents the matching degree between doctor Doctor a and the disease type D b that the patient may have, P(D b ) represents the probability that the patient has the disease D b , E(Doctor a ,S c ) represents the professional ability score of doctor Doctor a in the professional field S c , and C(Doctor a ,D b ) represents the cross - field cooperation score of doctor Doctor a in the disease type D b .

[0008] Further, the patient data further includes disease-related characteristics, and the doctor information further includes professional levels; the disease-related characteristics are input into a trained support vector model to obtain the complexity of the disease; corresponding doctors with professional levels are matched based on the preliminary matching results and the complexity level of the disease.

[0009] Further, the complexity includes low-complexity diseases, medium-complexity diseases, and high-complexity diseases, and the professional levels include primary professionals, intermediate professionals, and senior professionals; the low-complexity diseases, medium-complexity diseases, and high-complexity diseases respectively preferentially match the doctor terminals of primary professionals, intermediate professionals, and senior professionals from the preliminary matching results and send diagnostic requests. If no response information is received within the specified time, doctor terminals of other professional levels are added.

[0010] Further, the evaluation data further includes the urgency level, and the doctor terminal information further includes emergency level information, and the emergency level information includes level A; after obtaining the preliminary matching results, doctor terminals of level A are preferentially matched from the preliminary matching results, and after a time threshold T, other doctor terminals are matched.

[0011] Further, T is: T = T 0 *e -kU ; where T 0 represents the initially set time threshold, T represents the actual time threshold, U represents the urgency level, 0 < U < 1, and k represents an adjustable parameter used to control the influence degree of the urgency level on the time threshold.

[0012] Further, the steps for obtaining the comprehensive score include:

[0013] The response time score is obtained by calculating the doctor's historical average response time. Specifically: the response times of the doctor's most recent N diagnostic requests are normalized and mapped to a 0 - 10 score system;

[0014] The treatment effect score is obtained by weighting the patient recovery rate and the satisfaction score. The patient recovery rate is the proportion of the physiological indicators of the cases handled by the doctor that return to normal, and the satisfaction score is the average score of the patient's evaluation of the diagnosis and treatment service; obtain the comprehensive score S em S em is:

[0015] S em = w 1 *S ir + w 2 *S ef

[0016] where w 1 is the response time weight coefficient, S ir represents the response time score, and S efRepresents the treatment effect score, w 2 is the treatment effect weight coefficient;

[0017] If the comprehensive score S em is greater than the preset value, the emergency level information at the doctor side is level A.

[0018] In a second aspect, the present application provides a doctor resource matching system based on vital sign data, which adopts the doctor resource matching method based on vital sign data as in the first aspect, and includes: a receiving module, an evaluation module, a matching module, and a sending module.

[0019] The receiving module is used to receive patient data sent by the patient side, and the patient data includes the vital sign data of the patient; the evaluation module is used to input the patient data into a deep learning model to obtain the evaluation data of the patient, and the evaluation data includes the disease type probability distribution The matching module is used to obtain the doctor side information of the online doctor side, and match the disease type probability distribution of the patient with the doctor side information to obtain the total matching degree of each online doctor side, and select several doctors with the highest total matching degree as the preliminary matching result; the sending module is used to send a diagnosis request to the doctor side based on the preliminary matching result.

[0020] In a third aspect, the present application provides a doctor resource matching device based on vital sign data, which includes: a memory and a processor.

[0021] The memory is used to store a computer program; the processor is used to implement the steps of the doctor resource matching method based on vital sign data as in the first aspect when executing the computer program.

[0022] The technical solution provided by the present application has at least the following technical effects or advantages:

[0023] Since the vital sign data is uploaded to the background, automatic doctor matching and remote treatment intervention are realized, effectively solving the technical problems of inaccurate diagnosis, large consumption of time and energy, and inaccurate doctor matching caused by patients usually having to evaluate their own symptoms by themselves in order to find a corresponding professional doctor for medical treatment, realizing remote communication and treatment between doctors and patients, improving the utilization efficiency of medical resources, shortening the waiting time of patients, and at the same time ensuring that patients can obtain professional medical services in a timely manner. This method can also respond quickly in case of emergency, reducing the risk of disease deterioration caused by waiting for treatment and improving the quality and efficiency of medical services. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a schematic flow chart of the doctor resource matching method based on vital sign data in the present application;

[0025] Figure 2 This is a schematic flow diagram of the doctor's end that matches Class A in this application. Detailed implementation manners

[0026] To solve the technical problems of low diagnostic accuracy, large consumption of time and energy, and inaccurate doctor matching caused by patients usually having to self-evaluate their symptoms to find a corresponding professional doctor for medical treatment, this application provides a doctor resource matching method based on vital sign data, which can improve the accuracy of medical diagnosis, optimize the doctor matching process, and accelerate the doctor response time to better meet the needs of patients for medical services.

[0027] To better understand the above technical solution, the above technical solution will be described in detail below in combination with the accompanying drawings of the specification and specific implementation manners.

[0028] As Figure 1 - Figure 2 shown, this embodiment provides a doctor resource matching method based on vital sign data, which includes the following steps: receiving patient data sent by the patient end, where the patient data includes the vital sign data of the patient; inputting the patient data into a deep learning model to obtain the evaluation data of the patient, and the evaluation data includes the disease type probability distribution Obtaining the doctor end information of the online doctor end, and matching the disease type probability distribution of the patient with the doctor end information to obtain the total matching degree of each online doctor end, and selecting several doctors with the highest total matching degree as the preliminary matching result; sending a diagnosis request to the doctor end based on the preliminary matching result; receiving the response information returned by the doctor end according to the diagnosis request.

[0029] For some diseases, it may be necessary to use a prediction model and a matching mechanism to allocate doctor resources at appropriate levels, such as diseases like sudden heart attack cases, stroke cases, cancer patient treatment, and sudden allergic reactions. A life detector can real-time monitor the vital sign data of patients, such as heart rate, blood pressure, blood oxygen saturation, etc. By analyzing the vital sign data, the symptoms of patients can be initially judged to match the corresponding doctors. In addition, the medical history and symptom description of the patient can also be combined to comprehensively judge the symptoms of the patient.

[0030] Build a disease diagnosis model, which can use a machine learning model, such as a deep learning model or a rule-based expert system. The input of the model is the preprocessed vital sign data, and the output is the diseases or disease types that the patient may suffer from.

[0031] Taking a machine learning model for disease diagnosis based on a deep learning convolutional neural network as an example, after collecting a large amount of labeled vital sign data, the data is preprocessed, such as signal filtering, normalization, etc. Then, a disease diagnosis model is constructed, and a convolutional neural network model suitable for disease diagnosis is designed, including a convolutional layer, a pooling layer, a fully connected layer, etc. The convolutional neural network model can be expressed as: In the formula, X represents the input vital sign data, and W 1 and W 2 represent the weight matrices of two fully connected layers respectively, b 1 and b 2 represent the biases of two fully connected layers respectively, ReLU represents the activation function, and softmax is used for the output of multi-classification problems.

[0032] The doctor-side information includes the professional field S c and the cross-field. S c = S 1 , S 2 , …, S m . In the formula, S m represents the m-th professional field, represents the probability of the n-th disease type D b ; Matching the probability distribution of the patient's disease type with the doctor-side information, the total matching degree of each online doctor-side is obtained, specifically including: calculating the matching degree M(Doctor a and the possible disease type D b ) of the patient. M(Doctor a , P(D b )) is: M(Doctor a , P(D b )) = α * D a * E(Doctor b , S b ) + β * C(Doctor a , D c ); In the formula, M(Doctor a , D b ) represents the matching degree between Doctor a and the possible disease type D b of the patient, P(D a ) represents the probability that the patient has the disease D b , E(Doctor b , S b ) represents Doctor a , S c ) represents Doctora In the professional field S c The professional ability score, C(Doctor a , D b ) represents the cross-field cooperation score of doctor Doctor a in the disease type D b .

[0033] Select an appropriate loss function to measure the difference between the model output and the true disease label. Given that each patient only needs to match a doctor in a single professional field at a time, that is, the main disease of each patient corresponds to a unique disease type, the loss function uses the multi-class cross-entropy loss function.

[0034] If there are n possible disease types, the output of the model is an n-dimensional probability distribution vector representing the predicted probability of the model for each disease type. For a sample, the true disease type label is represented as an n-dimensional encoded vector y, y = (y 1 , y 2 , …, y n ), where only one element is 1, representing the true disease type. Then the loss function can be expressed as: In the formula, y i is the i-th group of encoded vectors, is the i-th group of probability distribution vectors, i represents the number of groups, represents the model output and the cross-entropy loss between the true label. By minimizing the multi-class cross-entropy loss function, the output probability distribution vector of the model can be made as close as possible to the true disease type label y. Since the severity and impact of different disease types may vary, the loss function can also be adjusted according to specific needs, such as the weighted cross-entropy loss function, to better reflect the importance of different disease types. After obtaining the predicted probability distribution of the disease types that the patient may have, match the online doctor information from the information in the pre-established doctor database, so as to achieve automatic selection and matching of doctors.

[0035] In the solution of this embodiment, the matching degree of doctors is evaluated by considering multiple factors such as the probability distribution of the patient's disease type and the doctor's professional field and cross-field cooperation, so as to achieve preliminary matching. For example, the doctor's professional field is S c , S c = S 1 , S 2 , …, S m , for each doctor Doctor a and the disease types that the patient may have The matching degree formula between them can be calculated as: M(Doctor a ,P(D b ))=α*D b* E(Doctor a ,S c )+β*C(Doctor a ,D b ), where M(Doctor a ,D b ) represents the matching degree between doctor Doctor a and the disease type D b that the patient may have, P(D b ) represents the probability that the patient has the disease D b , E(Doctor a ,S c ) represents the professional ability score of doctor Doctor a in the professional field S c , C(Doctor a ,D b ) represents the cross-field cooperation score of doctor Doctor a in the disease type D b . If the doctor has cooperation experience in other fields related to the disease type D b , additional matching degree points can be given. α and β respectively represent the weight coefficients used to balance the matching degree of the patient's disease type probability distribution and the doctor's professional field as well as the matching degree of cross-field cooperation. The final matching degree matrix can be represented as an m×n matrix, where m represents the number of doctors and n represents the number of disease types that the patient may have. After calculating the matching degree matrix according to the matching degree evaluation formula, the matching degrees of various disease types can be comprehensively considered, the total matching degree can be calculated for each doctor, and several doctors with the highest matching degree can be selected as the preliminary matching results and wait for the doctors' responses.

[0036] By uploading the sick sign data to the background, automatic doctor matching and remote treatment intervention are realized. The remote communication and treatment between doctors and patients improve the utilization efficiency of medical resources, shorten the waiting time of patients, and at the same time ensure that patients can obtain professional medical services in a timely manner. This method can also respond quickly in case of emergencies, reduce the risk of the patient's condition worsening due to waiting for treatment, and improve the quality and efficiency of medical services.

[0037] Patient data also includes disease-related characteristics, and doctor information also includes professional levels; input the disease-related characteristics into the trained support vector model to obtain the complexity of the disease; match doctors with corresponding professional levels based on the preliminary matching results and the complexity level of the disease. The complexity includes low-complexity diseases, medium-complexity diseases, and high-complexity diseases, and the professional levels include primary professionals, intermediate professionals, and senior professionals; low-complexity diseases, medium-complexity diseases, and high-complexity diseases respectively preferentially match doctors' terminals of primary professionals, intermediate professionals, and senior professionals from the preliminary matching results and send diagnosis requests. If no response information is received within the specified time, add doctors' terminals of other professional levels.

[0038] The process for obtaining professional levels includes:

[0039] Extract four basic indicators of educational background, qualification certificates, years of clinical experience, and scientific research achievements from the doctor database;

[0040] Perform standardization processing on each indicator and obtain a score range through weighted average;

[0041] Divide professional levels according to the score range:

[0042] Primary: S < 60 points;

[0043] Intermediate: 60 ≤ S < 85 points;

[0044] Senior: S ≥ 85 points.

[0045] Since the above matching method may cause some complex cases to be assigned to doctors with relatively low professional levels for treatment, which may affect the quality of medical services and the treatment effect of patients. High-level doctors handle some relatively simple cases, which will waste the professional skills and time of high-level doctors, reduce the utilization efficiency of medical resources, and may also cause patients to be dissatisfied with medical services, affecting the hospital's reputation and patients' trust. In the solution of this embodiment, on the above basis, by analyzing the patient's vital sign data and symptoms through a support vector machine, the cases can be evaluated and their complexity levels can be determined, and then the doctors are classified according to their professional levels to ensure that each case can receive the services of doctors at the corresponding level.

[0046] When training a support vector model, features related to the patient's condition are used as input. Patient disease-related features include, but are not limited to, physiological indicators, text information of symptom descriptions, and the patient's basic information. Physiological indicators such as blood pressure, heart rate, body temperature, blood test results, etc. Symptom descriptions such as the symptoms described by the patient, pain level, duration, etc. The patient's basic information can include age, gender, past medical history, etc. These features can be numerical features, text features, or even categorical features, which are processed and feature engineered according to specific situations. The output is the corresponding complexity label. For example, -1 represents low complexity, 0 represents medium complexity, and 1 represents high complexity. After completing the training of the model, the complexity level of the disease is evaluated by inputting features.

[0047] Doctors can be rated according to multiple indicators, such as educational background and professional qualifications, clinical experience and technical level, patient satisfaction and medical quality, and professional development and academic contributions. Educational background and professional qualifications include the doctor's academic qualifications, professional qualification certificates, continuing education, etc. Clinical experience and technical level refer to the doctor's clinical work experience, technical level, research achievements, etc. Patient satisfaction and medical quality include patient satisfaction, medical quality evaluation, medical accident records, etc. Professional development and academic contributions can be the doctor's academic research achievements, academic exchange activities, contributions to the medical cause, etc. By performing a weighted sum of different indicators, the doctor's rating score can be obtained. By setting intervals of different rating scores, the doctor's professional level can be divided into junior, intermediate, and senior levels from low to high.

[0048] Among the initially matched doctors, doctors with corresponding professional levels can be matched according to the complexity level of the disease. For low-complexity diseases, doctors with a junior professional level are preferentially matched and a diagnosis request is sent to doctors with a junior professional level. If there is no response from the doctor side within the specified time, doctors with an intermediate professional level are simultaneously matched and a diagnosis request is sent. If there is still no response within the specified time, doctors with a senior professional level are finally simultaneously matched. For medium-complexity diseases, doctors with an intermediate professional level are preferentially matched, followed by doctors with a senior professional level. When doctors with a senior professional level do not respond, doctors with a junior professional level are simultaneously matched. For high-complexity diseases, doctors with a senior professional level are preferentially matched, followed by doctors with an intermediate professional level, and finally doctors with a junior professional level are simultaneously matched. Classifying doctors according to their professional levels is conducive to better matching the patient's disease needs, reasonably allocating medical resources, and improving the efficiency and quality of medical services.

[0049] The evaluation data also includes the urgency level. The doctor-side information also includes emergency level information, and the emergency level information includes level A. After obtaining the initial matching results, doctors with level A on the doctor side are preferentially matched from the initial matching results. After a time threshold T, other doctor sides are matched.

[0050] In emergency situations such as stroke, sudden heart attack, severe trauma, respiratory distress, severe bleeding, etc., urgent medical intervention is required to ensure that patients can receive appropriate medical services in a timely manner. When using a convolutional neural network based on deep learning to analyze vital sign data to initially judge the symptoms of patients, an additional output layer is added to the model to evaluate and predict the urgency of the patient's condition. The additional output layer can be designed as a regression layer, outputting a value ranging from 0 to 1, representing the urgency of the patient, which represents the probability or degree of the patient's need for urgent medical intervention. When training the deep learning model, a training data set with vital sign data and corresponding urgency level labels needs to be prepared. These labels can be the urgency level evaluations given by doctors based on the patient's condition, or calibrated according to factors such as the severity of the patient's illness. A fully connected layer is added after the last layer of the convolutional neural network as an additional output layer to output a value representing the urgency. At the same time, an appropriate loss function, such as mean squared error, needs to be defined to measure the gap between the predicted value of the urgency percentage and the true label.

[0051] After the model inputs the patient's vital sign data, while obtaining the possible diseases or types of diseases the patient may have, it can also obtain the evaluation result of the urgency percentage, so as to be able to match doctor resources according to the urgency of the patient. When evaluating the matching degree of doctors based on the probability distribution of the patient's disease type and the cooperation between the doctor's professional field and cross-field, by focusing on doctors who can meet the medical needs in emergency situations, rapid and accurate doctor matching can be achieved in emergency situations, improving the survival rate and treatment effect of patients.

[0052] The steps to obtain the comprehensive score include:

[0053] The response time score is obtained by calculating the doctor's historical average response time. Specifically: normalize the response times of the doctor's most recent N diagnostic requests and map them to a 0-10 score system;

[0054] The treatment effect score is obtained by weighting the patient's recovery rate and satisfaction score. The patient's recovery rate is the proportion of the physiological indicators of the cases handled by the doctor that return to normal, and the satisfaction score is the average score of the patient's evaluation of the diagnosis and treatment service; obtain the comprehensive score S em , S em is:

[0055] S em = w 1 * S ir + w 2 * S ef

[0056] In the formula, w 1is the response time weight coefficient, S ir represents the response time score, S ef represents the treatment effect score, w 2 is the treatment effect weight coefficient;

[0057] If the comprehensive score S em is greater than the preset value, the emergency level information at the doctor's end is level A.

[0058] Specifically, information such as response time and treatment effect can be introduced to match doctors. During the interaction process of matching doctors with patients, collect the response time of the matching doctor, the treatment plan provided and its effect. Start timing when the matching doctor receives the alarm and starts to handle the emergency, record the response time of the matching doctor, that is, the time from receiving the response request to starting to provide medical support. After the matching doctor handles the emergency, record the treatment result of the patient, including changes in the patient's vital sign data, symptom relief, etc., to evaluate the doctor's treatment effect.

[0059] When recording the response time, timing can start from the time when the matching doctor responds to the diagnosis, record the timestamp, the matching doctor starts to provide medical support until the emergency is properly handled, and after the matching doctor provides medical support, stop timing and record the timestamp of the end time. Use the end time minus the start time to calculate the response time of the matching doctor.

[0060] When recording the treatment effect, the collected treatment result data can be analyzed to evaluate the doctor's treatment effect. For example, whether the vital sign data is effectively controlled, whether the symptoms are relieved, etc. Record the evaluated treatment effect in the patient's medical record or a dedicated treatment effect database. At the same time, feedback from the patient or family members can also be collected to understand their evaluation and opinions on the treatment effect provided by the doctor. Considering the changes in vital sign data, symptom relief, and patient feedback comprehensively, comprehensively evaluate the doctor's treatment effect.

[0061] T is: T = T 0 *e -kU ; In the formula, T 0 represents the initially set time threshold, T represents the actual time threshold, U represents the emergency level, 0 < U < 1, and k represents an adjustable parameter used to control the influence degree of the emergency level on the time threshold.

[0062] By evaluating the doctor's response time and treatment effect, the comprehensive score S of the doctor's emergency treatment can be obtained em , S em = w 1 *S ir + w 2 *S ef , in the formula, Sir Represents the response time score, S ef Represents the treatment effect score. Through a preset threshold, based on the comprehensive score S em Doctors are divided into several levels. For those with a high comprehensive score, they can be classified as Class A doctors. After obtaining the preliminary matching results, the evaluation results of the patient's disease type and the percentage of urgency are obtained. First, match Class A doctors from the preliminarily matched doctors. After passing the time threshold T, then match Class B doctors from the preliminarily matched doctors at the same time.

[0063] The time threshold can be determined according to the urgency of the patient. Specifically, the calculation formula for the time threshold T is: T = T 0 *e -kU , where T 0 Represents the initially set time threshold, T represents the actual time threshold, k represents an adjustable parameter used to control the influence degree of urgency on the time threshold, represents the urgency, and the value range is between 0 and 1.

[0064] By preferentially matching preferred doctors, the effective utilization rate of doctor resources and the patient's medical treatment efficiency can be significantly improved. On the premise of ensuring the quality of medical services, the professional ability and experience of preferred doctors can better meet the needs of patients, thus directly improving the overall quality of medical services and patient satisfaction. Introducing the patient's urgency as an adjustment factor for the matching time has strong flexibility and adaptability. When the patient's condition is urgent, the system can automatically shorten the time to switch and match all doctors, ensuring that the patient can receive corresponding medical services in the shortest time, which plays an important role in improving the first aid efficiency and reducing the risks that the patient may face due to waiting for treatment.

[0065] Embodiment 2

[0066] This embodiment provides a doctor resource matching system based on vital sign data, which adopts the doctor resource matching method based on vital sign data in Embodiment 1, and includes: a receiving module, an evaluation module, a matching module, and a sending module.

[0067] The receiving module is used to receive patient data sent by the patient terminal, and the patient data includes the patient's vital sign data; the evaluation module is used to input the patient data into the deep learning model to obtain the patient's evaluation data, and the evaluation data includes the disease type probability distribution The matching module is used to obtain the doctor terminal information of the online doctor terminal, and the disease type probability distribution of the patient Is matched with the doctor terminal information to obtain the total matching degree of each online doctor terminal, and select several doctors with the highest total matching degree as the preliminary matching results; the sending module is used to send a diagnosis request to the doctor terminal based on the preliminary matching results.

[0068] This doctor resource matching system based on vital sign data has all the advantages of the doctor resource matching method based on vital sign data and can automatically implement all the steps of the doctor resource matching method based on vital sign data.

[0069] Embodiment III

[0070] This application provides a doctor resource matching device based on vital sign data, which includes: a memory and a processor.

[0071] The memory is used to store computer programs; the processor is used to implement the steps of the doctor resource matching method based on vital sign data as in Embodiment I when executing the computer programs.

[0072] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0073] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate computer-implemented processing, and thus the instructions executed on the computer or other programmable device provide steps for realizing the functions specified in one Figure 1 one process or multiple processes and / or blocks Figure 1 or multiple blocks. Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications to these embodiments once they know the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present invention.

Claims

1. A doctor resource matching method based on vital signs data, characterized in that: It includes the following steps: Receiving patient data sent by the patient terminal, where the patient data includes the patient's vital sign data; The patient data is input into a deep learning model to obtain the patient's evaluation data, which includes the probability distribution of disease types. Obtain the doctor's side information of the online doctor side and distribute the probability of the patient's disease type Matching with the doctor-side information, obtaining the total matching degree of each online doctor-side, and selecting the doctors with the highest total matching degree as the preliminary matching results; Sending a diagnosis request to the doctor terminal based on the preliminary matching result.

2. The doctor resource matching method based on vital sign data according to claim 1, characterized in that: Receiving the response information returned by the doctor terminal according to the diagnosis request.

3. The doctor resource matching method based on vital sign data according to claim 1, characterized in that: The doctor-side information includes professional field S c and cross-field, S c =S1,S2,…,S m , where S m represents the mth professional field; Represents the nth disease type D b probability; The probability distribution of the disease type of the patient Matching with the doctor's end information to obtain the total matching degree of each online doctor's end specifically includes: Calculate each doctor a and the type of disease the patient may have b The matching degree between M(Doctor a ,P(D b )), M (Doctor a ,P(D b ))for: M(Doctor a ,P(D b ))=α*D b *E(Doctor a ,S c )+β*C(Doctor a ,D b ) In the formula, M(Doctor a ,D b ) stands for Doctor a The type of disease the patient may have b The matching degree, P(D b ) represents the patient suffering from disease D b The probability of E(Doctor a ,S c ) stands for Doctor a In the professional field c Professional ability score, C (Doctor a ,D b ) stands for Doctor a In disease type D b Cross-field cooperation scores.

4. The doctor resource matching method based on vital sign data according to claim 1, characterized in that: The patient data further includes disease-related characteristics, and the doctor information further includes professional levels; inputting the disease-related characteristics into the trained support vector model to obtain the complexity of the disease; Matching doctors with corresponding professional levels based on the preliminary matching result and the complexity level of the disease.

5. The doctor resource matching method based on vital sign data according to claim 4, characterized in that: The complexity includes low-complexity diseases, medium-complexity diseases, and high-complexity diseases, and the professional levels include primary professionals, intermediate professionals, and senior professionals; For the low-complexity diseases, medium-complexity diseases, and high-complexity diseases, doctors' terminals of primary professionals, intermediate professionals, and senior professionals are respectively preferentially matched from the preliminary matching result and a diagnosis request is sent. If no response information is received within the specified time, doctors' terminals of other professional levels are added.

6. The doctor resource matching method based on vital sign data according to claim 1, characterized in that: The evaluation data further includes the urgency level, and the doctor terminal information further includes emergency level information, and the emergency level information includes level A; After obtaining the preliminary matching result, doctors' terminals of level A are preferentially matched from the preliminary matching result. After a time threshold T has passed, other doctors' terminals are matched.

7. The doctor resource matching method based on vital sign data according to claim 6, characterized in that: T is: T=T0*e -kU In the formula, T0 represents the initially set time threshold, T represents the actual time threshold, U represents the urgency level, 0 < U < 1, and k represents an adjustable parameter used to control the influence degree of the urgency level on the time threshold.

8. The doctor resource matching method based on vital sign data according to claim 6, characterized in that: The steps of obtaining the comprehensive score include: The response time score is obtained by calculating the doctor's historical average response time. Specifically: normalizing the response times of the doctor's most recent N diagnosis requests and then mapping them to a 0-10 score system; The treatment effect score is obtained by weighting the patient recovery rate and satisfaction score. The patient recovery rate is the proportion of physiological indicators of the cases treated by the doctor that return to normal, and the satisfaction score is the average score of the patients on the diagnosis and treatment services; the comprehensive score S is obtained. em , S em for: S em =w1*S ir +w2*S ef Where w1 is the response time weight coefficient, S ir Represents the response time score, S ef represents the treatment effect score, w2 is the treatment effect weight coefficient; If the comprehensive score S em If it is greater than the preset value, the emergency level information on the doctor's side is A.

9. A doctor resource matching system based on vital signs data, which adopts the doctor resource matching method based on vital signs data as claimed in any one of claims 1 to 8, characterized in that: It includes: A receiving module for receiving patient data sent by the patient terminal, where the patient data includes the patient's vital sign data; An evaluation module is used to input the patient data into a deep learning model to obtain the patient's evaluation data, wherein the evaluation data includes a probability distribution of disease types. The matching module is used to obtain the doctor-side information of the online doctor side and distribute the disease type probability of the patient Matching with the doctor-side information, obtaining the total matching degree of each online doctor-side, and selecting the doctors with the highest total matching degree as the preliminary matching results; A sending module for sending a diagnosis request to the doctor terminal based on the preliminary matching result.

10. A doctor resource matching device based on vital signs data, characterized in that: It includes: A memory for storing computer programs; A processor for implementing the steps of the doctor resource matching method based on vital sign data as described in any one of claims 1-8 when executing the computer programs.

Images