A method, device and equipment for processing medical intention

By extracting and classifying the cases, identifying the intention to visit the doctor, the problem of difficulty in identifying the intention to visit the doctor is solved, and the reasonable allocation and management efficiency of medical resources is improved.

CN110428891BActive Publication Date: 2024-11-26TENCENT TECHNOLOGY (SHENZHEN) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN201910699975.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2019-07-31
Publication Date
2024-11-26
Estimated Expiration
2039-09-17

AI Technical Summary

Technical Problem

The existing technology lacks the intention to effectively identify patients who are present, resulting in unreasonable allocation of medical resources and affecting the efficiency of medical management.

Method used

By obtaining cases, performing feature extraction and classification processing, determining the case type, and identifying the intention to visit based on the classification model, including subclasses of diagnosis and treatment and non-diagnosis and treatment, combining the judgment results to identify the highest priority treatment intention.

Benefits of technology

It has realized automated and accurate identification of medical intentions, improved medical management efficiency, and promoted the rational allocation and use of medical resources.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN110428891B_ABST
    Figure CN110428891B_ABST
Patent Text Reader

Abstract

The present invention discloses a method, device, equipment and storage medium for processing medical intention. The method includes: obtaining a case; extracting features from the case to obtain corresponding case features; determining the type of the case based on the case features; when the type of the case is determined to be a non-diagnosis and treatment type, classifying the case features through a classification model to obtain a judgment result that the case belongs to each subclass included in the non-diagnosis and treatment type; based on the judgment results of each subclass, identifying the medical intention corresponding to the case. Through the present invention, the medical intention of the patient can be automatically and accurately identified.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to intelligent medical technology in the field of artificial intelligence (AI), and in particular to a method, device, equipment and storage medium for processing medical intention. Background Art

[0002] Smart medical technology is a branch of AI technology, which aims to use machines to learn based on prior knowledge, so as to have the logical ability to classify and judge data such as videos, images and texts.

[0003] Medical management technology in the medical field is a typical application of smart medical technology. Reasonable planning and full use of the hospital's existing resources and effective allocation of patient visits are of great significance in practical applications.

[0004] As the number of patients visiting the hospital continues to increase, analyzing the distribution of patients who have visited the hospital in the past and their intentions of coming to the hospital has a high application value for the hospital's rational use and allocation of resources. However, the relevant technology lacks an effective technical solution to identify the intentions of patients visiting the hospital. Summary of the invention

[0005] The embodiments of the present invention provide a method, device, equipment and storage medium for processing medical treatment intention, which can automatically and accurately identify the medical treatment intention of a medical treatment subject.

[0006] The technical solution of the embodiment of the present invention is achieved as follows:

[0007] The embodiment of the present invention provides a method for processing a medical consultation intention, including:

[0008] Obtain case records;

[0009] Extracting features of the case to obtain corresponding case features;

[0010] Determining the type of the case based on the characteristics of the case;

[0011] When it is determined that the type of the case is a non-diagnosis and treatment type, the case characteristics are classified by a classification model to obtain a judgment result that the case belongs to each subclass included in the non-diagnosis and treatment type;

[0012] Based on the judgment results of each subclass, the medical treatment intention corresponding to the case is identified.

[0013] In the above scheme, the identification of the medical treatment intention corresponding to the case based on the judgment results of each subclass includes:

[0014] Merging the judgment results of the subcategories to obtain a combined result;

[0015] Based on the merged result, the medical treatment intention corresponding to the case is identified.

[0016] In the above scheme, the identifying the medical treatment intention corresponding to the case based on the combined result includes:

[0017] Based on the merging result, each subclass for which the merging result is determined to be yes is output in order of priority;

[0018] Based on the output results, the medical treatment intention corresponding to the subclass with the highest priority is identified as the medical treatment intention corresponding to the case.

[0019] The embodiment of the present invention further provides a device for processing medical consultation intention, including:

[0020] An acquisition unit, used for acquiring cases;

[0021] An extraction unit, used for extracting features from the case to obtain corresponding case features;

[0022] A determination unit, configured to determine the type of the case based on the case characteristics;

[0023] A classification unit, for, when it is determined that the type of the case is a non-diagnosis and treatment type, classifying the case characteristics through a classification model to obtain a judgment result that the case belongs to each subclass included in the non-diagnosis and treatment type;

[0024] An identification unit is used to identify the medical treatment intention corresponding to the case based on the judgment results of each subclass.

[0025] In the above solution, the determining unit is specifically used to:

[0026] Classify the feature vector corresponding to the case feature through the classification model to obtain the probability that the case is a diagnosis and treatment class output by the classification model;

[0027] When the obtained probability is greater than a first probability threshold, determining the type of the case as diagnosis and treatment type;

[0028] When the obtained probability is less than or equal to the first probability threshold, the type of the case is determined to be non-diagnosis and treatment type.

[0029] In the above solution, the extraction unit is specifically used for at least one of the following:

[0030] Identifying the symptoms in the case, and counting the identified symptoms to obtain the corresponding number of symptoms;

[0031] Identify the part words in the case, and count the identified part words to obtain the corresponding number of part words;

[0032] Counting the number of words included in the case to obtain the corresponding case length;

[0033] The keywords or fragment hits with classification characteristics in the case are screened to obtain corresponding case keywords or case fragment hits.

[0034] In the above scheme, the classification unit is specifically used for:

[0035] Classify the feature vectors corresponding to the case features by at least two of the classification models respectively, and obtain the probability of the case belonging to each subclass output by the at least two classification models;

[0036] When the obtained probability is greater than a second probability threshold, determining that the case belongs to a subclass corresponding to the probability;

[0037] Among them, the subcategories include at least one of the following: follow-up visit category; examination category; prescription category; incomplete information category.

[0038] In the above scheme, the classification unit is also specifically used for:

[0039] According to the order of the follow-up visit class, examination class, prescription class and incomplete information class, the feature vectors corresponding to the case characteristics are classified and processed by the classification model respectively, and the corresponding probabilities of the case belonging to the follow-up visit class, examination class, prescription class and incomplete information class are obtained.

[0040] In the above solution, the identification unit is specifically used for:

[0041] Merging the judgment results of the subcategories to obtain a combined result;

[0042] Based on the merged result, the medical treatment intention corresponding to the case is identified.

[0043] In the above solution, the identification unit is further specifically used for:

[0044] Based on the merging result, each subclass for which the merging result is determined to be yes is output in order of priority;

[0045] Based on the output results, the medical treatment intention corresponding to the subclass with the highest priority is identified as the medical treatment intention corresponding to the case.

[0046] In the above solution, the identification unit is also used for:

[0047] When it is determined that the type of the case is diagnosis and treatment, the consultation intention corresponding to the case is identified as diagnosis and treatment.

[0048] In the above scheme, the processing device for the intention to see a doctor also includes:

[0049] A construction unit, configured to construct a medical treatment distribution structure diagram for a medical treatment object based on the medical treatment intention corresponding to the case after the identification unit identifies the medical treatment intention corresponding to the case based on the judgment results of the subclasses;

[0050] An allocation unit is used to allocate corresponding medical resources to the patient based on the patient distribution structure diagram.

[0051] In the above scheme, the processing device for the intention to see a doctor also includes:

[0052] A deconstruction unit is used to deconstruct different types of cases based on the medical treatment intentions corresponding to the cases after the identification unit identifies the medical treatment intentions corresponding to the cases based on the judgment results of the subcategories, so as to assist in the prediction and diagnosis of the disease.

[0053] The embodiment of the present invention further provides a device for processing medical consultation intention, including:

[0054] A memory for storing executable instructions;

[0055] The processor is used to implement the method for processing the medical consultation intention provided by the embodiment of the present invention when executing the executable instructions stored in the memory.

[0056] An embodiment of the present invention further provides a storage medium storing executable instructions, which, when executed, are used to implement the method for processing medical consultation intention provided by the embodiment of the present invention.

[0057] The application of the above embodiments of the present invention has the following beneficial effects:

[0058] The method for processing the medical treatment intention provided in the embodiment of the present invention is applied to obtain the case, extract the features of the case, determine the type of the case based on the case features, and then further identify the medical treatment intention corresponding to the case based on the case type. In this way, by using the characteristics of the case text itself, the medical treatment intention of the patient corresponding to the case can be automatically and accurately identified, providing available data for subsequent case structuring, auxiliary diagnosis, etc., thereby improving the efficiency of medical management. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A schematic diagram of an optional application scenario of the system for processing medical consultation intention provided by an embodiment of the present invention;

[0060] Figure 2A An optional structural diagram of a device for processing medical consultation intention provided in an embodiment of the present invention;

[0061] Figure 2B An optional structural diagram of a device for processing medical consultation intention provided in an embodiment of the present invention;

[0062] Figure 3 An optional flowchart of a method for processing a medical consultation intention provided in an embodiment of the present invention;

[0063] Figure 4 Another optional flowchart of the method for processing the medical consultation intention provided in the embodiment of the present invention;

[0064] Figure 5 An optional schematic diagram of the distribution structure display of the distribution data of the patients who visited the hospital provided by the embodiment of the present invention;

[0065] Figure 6 Another optional flowchart of the method for processing the medical consultation intention provided in the embodiment of the present invention;

[0066] Figure 7 An optional schematic diagram of the principle structure of a method for processing medical consultation intention provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0067] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the present invention is further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present invention. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present invention.

[0068] In the following description, reference is made to “some embodiments”, which describe a subset of all possible embodiments. However, it can be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments, and the various technical solutions recorded in the embodiments of the present invention may be combined with each other without conflict.

[0069] Unless otherwise defined, all technical and scientific terms used in the embodiments of the present invention have the same meanings as those commonly understood by those skilled in the art of the technical field of the embodiments of the present invention. The terms used in the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0070] Before further describing the embodiments of the present invention in detail, the nouns and terms involved in the embodiments of the present invention are described first. The nouns and terms involved in the embodiments of the present invention are subject to the following explanations.

[0071] 1) A case refers to an example of a certain disease. For example, a person or organism that has suffered from a certain disease is a case of this disease. Usually, a case needs to be written in accordance with standard requirements.

[0072] 2) Treatment intention, which can be understood as the purpose of the patient, such as the patient, to visit the hospital for treatment. For example, a patient was diagnosed with kidney stones in Hospital A, but the patient was somewhat skeptical about the diagnosis of Hospital A. Therefore, the patient went to Hospital B for further examination to confirm whether he had kidney stones. At this time, the purpose of the patient's visit to Hospital B, that is, further examination, can be called the patient's treatment intention.

[0073] At present, the average daily outpatient volume of some hospitals in my country, especially some tertiary hospitals, usually exceeds 5,000 people, and the number of patients in different hospitals and departments is also extremely uneven. For example, a large proportion of patients visiting some departments are simply prescribed medicine or examined without the need for outpatient doctors to make a diagnosis. However, sometimes the allocation of such patients is not reasonable. This will lead to very tight medical resources. Therefore, reasonable planning and full use of medical resources, such as doctor resources and operating room resources, as well as effective allocation of patients for treatment, are of great significance in practical applications.

[0074] For the allocation of some hardware resources such as laboratory tests, estimates can be made in advance, which is of great reference value for increasing or reducing resource investment and making full use of existing resources. Therefore, analyzing the distribution of patients who have visited the hospital in the past and their intentions to visit the hospital has a high application value for the hospital's rational use and allocation of resources. However, there is a lack of effective solutions for how to quickly and accurately identify the intentions of patients to visit the hospital.

[0075] In order to at least solve the above-mentioned technical problems of the related art, the processing method, device, equipment and storage medium of the medical treatment intention provided by the embodiments of the present invention can automatically and accurately identify the medical treatment intention corresponding to the case according to the characteristics of the case text itself.

[0076] The following describes an exemplary application of a device for processing medical treatment intentions provided by an embodiment of the present invention. The device for processing medical treatment intentions provided by an embodiment of the present invention may be a server, such as a server deployed in the cloud, which provides the hospital with a function of remotely identifying the medical treatment intentions of patients corresponding to the cases according to various institutions with medical attributes, such as the patient's cases submitted by the hospital; it may also be a medical diagnostic device, such as a diagnostic device for one or more diseases (such as diabetes, hypertension, and cardiovascular and cerebrovascular diseases, etc.), which can identify the patient's medical treatment intentions based on the patient's cases, deconstruct different types of cases, and assist in the prediction and diagnosis of the disease; it may even be a handheld terminal or other device, such as a tablet computer, a laptop computer, and of course, a fixed terminal device such as a desktop computer.

[0077] The following will describe an exemplary application of the system for processing the intention of seeing a doctor according to an embodiment of the present invention with reference to the accompanying drawings. Figure 1 , Figure 1 A schematic diagram of an optional application scenario of the medical treatment intention processing system 10 provided in an embodiment of the present invention, wherein the terminal 100 can be located in various medical institutions (such as hospitals, medical research institutes), and can be used to collect patient medical records. For example, the patient medical records can be collected through the terminal 100's own collection device, or through other collection devices 400.

[0078] In some embodiments, the terminal 100 is used to locally execute the method for processing medical treatment intentions provided in an embodiment of the present invention to identify the medical treatment intentions corresponding to the patient's case, and output the results of the identified medical treatment intentions in a graphical manner, so that doctors and researchers can conduct research on disease prediction and diagnosis. For example, different types of cases can be deconstructed according to the medical treatment intentions corresponding to the identified cases, and the morphological manifestations of different types of cases can be determined, thereby assisting in predicting and diagnosing whether the patient is likely to suffer from a certain disease (such as diabetes) and its risks.

[0079] Of course, the terminal 100 can also send the patient's medical records to the server 300 through the network 200, and call the function of remotely identifying the patient's medical treatment intention provided by the server 300. The server 300 executes the method for processing medical treatment intention provided by the embodiment of the present invention. Specifically, after obtaining the patient's medical records, feature extraction is performed on the case to obtain corresponding case features, and then the type of the case is determined based on the case features. When the case type is determined to be a diagnosis and treatment type, the medical treatment intention corresponding to the case is identified as diagnosis and treatment; when the case type is determined to be a non-diagnosis and treatment type, the case features are classified and processed through a classification model to obtain a judgment result of whether the case belongs to each subclass included in the non-diagnosis and treatment type. Based on the judgment results of each subclass, the medical treatment intention corresponding to the case is identified, and finally the result of the identified medical treatment intention is returned to the terminal 100 for doctors and researchers to conduct research on disease prediction and diagnosis.

[0080] Here, the terminal 100 can be connected to the server 300 through the network 200 based on various wireless communication methods or wired communication methods. The network 200 can be a wide area network or a local area network, or a combination of the two, using a wireless link to achieve data transmission.

[0081] The terminal 100 may display various intermediate results and final results in the process of processing the medical consultation intention in the graphical interface 110 , for example, displaying the medical consultation intention corresponding to the identified case.

[0082] The structure of the processing device for the intention to see a doctor provided in the embodiment of the present invention is further described. The processing device for the intention to see a doctor may be various terminals, such as medical diagnostic equipment, desktop computers, etc., or may be a computer such as a computer. Figure 1 Server 300 is shown.

[0083] Now, a processing device for implementing a medical consultation intention according to an embodiment of the present invention will be described with reference to the accompanying drawings. Figure 2A This is an optional structural diagram of a device for processing medical consultation intention provided in an embodiment of the present invention. It can be understood that Figure 2A Only an exemplary structure of the processing device for the intention of seeing a doctor is shown, not all structures, and can be implemented as needed. Figure 2A The partial or entire structure shown shall not bring any limitation to the functions and application scope of the embodiments of the present invention.

[0084] See also Figure 2A, the treatment intention processing device 20 provided in the embodiment of the present invention includes: at least one processor 201, a memory 202, a user interface 203 and at least one network interface 204. The various components in the treatment intention processing device 20 are coupled together through a bus system 205. It can be understood that the bus system 205 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 205 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, in Figure 2A Various buses are labeled as bus system 205 .

[0085] The user interface 203 may include a display, a keyboard, a mouse, a trackball, a click wheel, keys, buttons, a touch pad or a touch screen.

[0086] It can be understood that the memory 202 can be a volatile memory or a non-volatile memory, and can also include both volatile and non-volatile memories.

[0087] The memory 202 in the embodiment of the present invention is used to store various types of data to support the operation of the treatment intention processing device 20. Examples of these data include: any executable instructions for operating on the treatment intention processing device 20, such as computer programs, including executable programs and operating systems, and the program for implementing the treatment intention processing method of the embodiment of the present invention can be included in the executable instructions.

[0088] The processor 201 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the method for processing the intention to see a doctor provided in the embodiment of the present invention can be completed by the integrated logic circuit of the hardware in the processor 201. The above-mentioned integrated logic circuit can be a general-purpose processor, a digital signal processor (DSP, Digital Signal Processor), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor 201 can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present invention.

[0089] Each step of the method for processing the intention to visit a doctor provided in the embodiment of the present invention can be completed by a software module. The software module can be located in a storage medium. The storage medium is located in the memory 202. The processor 201 executes the software module in the memory 202 and completes the steps of the method for processing the intention to visit a doctor provided in the embodiment of the present invention in combination with its hardware.

[0090] For example, as an example of a software module, the memory 202 may include a processing device 30 for the medical consultation intention provided in an embodiment of the present invention, which includes a series of software modules, such as an acquisition unit 31, an extraction unit 32, a determination unit 33, a classification unit 34 and an identification unit 35, see Figure 2B An optional structural diagram of a device 30 for processing medical consultation intention provided in an embodiment of the present invention is shown, and the functions of the above-mentioned various units will be explained below.

[0091] So far, the structure of the device for processing the medical consultation intention provided by the embodiment of the present invention and the exemplary application scenario of the system for processing the medical consultation intention have been described according to their functions. Next, the implementation of the method for processing the medical consultation intention provided by the embodiment of the present invention will be described.

[0092] See also Figure 3 , Figure 3 This is an optional flowchart of the method for processing the intention to see a doctor provided in an embodiment of the present invention. It can be understood from the above that the method for processing the intention to see a doctor provided in an embodiment of the present invention can be applied to various types of processing devices for processing the intention to see a doctor, such as medical diagnostic equipment, desktop computers, and servers. The following takes the processing device for processing the intention to see a doctor as an example, that is, the server executes the method for processing the intention to see a doctor provided in an embodiment of the present invention as an example. Figure 3 Steps shown.

[0093] Step 301: Obtain case data.

[0094] The following is an explanation of the method for obtaining the case. In some embodiments, the server may obtain the case in the following manner: the receiving terminal obtains the case by calling the collection device.

[0095] Specifically, the medical records obtained by the server can be obtained by the terminal by calling its own collection device, such as a camera installed on the terminal to collect the patient's medical records, and then sending the collected patient's medical records to the server; or it can be obtained by calling other collection devices deployed in the environment where the above-mentioned terminal is located, such as a camera in the same environment as the above-mentioned terminal, and the patient's medical records are collected by the camera, and then the collected patient's medical records are directly sent to the server. Of course, the camera can also send the collected patient's medical records to the above-mentioned terminal indirectly (such as through a wireless router), and then the above-mentioned terminal sends it to the server. The embodiment of the present invention does not limit the method used by the server to obtain the medical records.

[0096] In the embodiment of the present invention, the case obtained by the server can be a case written by the doctor based on the patient's consultation during the consultation, or it can be a case generated by the patient through a pre-consultation, which is not limited here. It should be noted that in the collection of patient consultations, the chief complaint and current medical history are the most direct descriptions of the patient's self-reported condition by the doctor. Therefore, in the embodiment of the present invention, the chief complaint and current medical history are used as the data that best reflects the case.

[0097] Step 302: Extract features from the case to obtain corresponding case features.

[0098] In an embodiment of the present invention, case features may include at least one of the following: number of symptoms; number of part words; case length; case keywords or case fragment hit words.

[0099] In some embodiments, the server may extract features from the case in at least one of the following ways to obtain corresponding case features:

[0100] Identify the symptoms in the case, and count the identified symptoms to obtain the corresponding number of symptoms;

[0101] Identify the part words in the case, and count the identified part words to obtain the corresponding number of part words;

[0102] Count the number of words included in the case to get the corresponding case length;

[0103] The keywords or fragment hits with classification characteristics in the cases are screened to obtain corresponding case keywords or case fragment hits.

[0104] It should be noted that the server can also obtain case features through the doctor's diagnosis in the case, that is, whether the doctor's diagnosis result contains symptoms or diseases. For example, if there is no doctor's diagnosis result in a case, the corresponding case feature can be set to zero.

[0105] Step 303: Determine the type of the case based on the case characteristics.

[0106] In practical applications, since there is no overlap between the diagnosis and treatment types in the current medical field and other types of medical treatment, in order to further improve the server's recognition efficiency of the patient's medical treatment intention, the embodiments of the present invention can first divide the acquired cases into diagnosis and treatment types and non-diagnosis and treatment types, and then for cases of the diagnosis and treatment type, it can be directly identified that the medical treatment intention corresponding to the case is diagnosis and treatment; for cases of the non-diagnosis and treatment type, by checking other labels of the disease respectively, continue to determine which subclass of the non-diagnosis and treatment type the case belongs to, so as to identify the medical treatment intention corresponding to the case according to the obtained subclass results.

[0107] The following is an explanation of the method for determining the case type. In some embodiments, the server may determine the type of the case in the following manner: classify the feature vector corresponding to the case feature through a classification model to obtain the probability that the case output by the classification model is a diagnosis and treatment type; when the obtained probability is greater than a first probability threshold, the type of the case is determined to be a diagnosis and treatment type; when the obtained probability is less than or equal to the first probability threshold, the type of the case is determined to be a non-diagnosis and treatment type.

[0108] In an embodiment of the present invention, the classification model may be a recurrent neural network (RNN) model in a deep learning model, or a logistic regression model; that is, the server may use an RNN model to classify and learn case features to determine whether the case type is a diagnosis and treatment class or a non-diagnosis and treatment class; of course, a logistic regression model, such as a binary logistic regression model, may also be used to classify and learn case features to determine whether the case type is a diagnosis and treatment class or a non-diagnosis and treatment class. The embodiment of the present invention does not limit which classification model the server chooses to use for classification processing.

[0109] Step 304: When it is determined that the type of the case is a non-diagnosis and treatment type, the case characteristics are classified using a classification model to obtain a judgment result of whether the case belongs to each subclass included in the non-diagnosis and treatment type.

[0110] The non-diagnosis and treatment category in the embodiment of the present invention includes multiple subcategories, and the subcategories include at least one of the following: follow-up category; examination category; prescription category; incomplete information category.

[0111] In the embodiment of the present invention, when the server determines that the type of the case is diagnosis and treatment, it can directly identify that the consultation intention corresponding to the case is diagnosis and treatment.

[0112] In some embodiments, the server may classify the case features in the following manner to obtain a judgment result of whether the case belongs to each subclass included in the non-diagnosis and treatment class: classify the feature vectors corresponding to the case features through at least two classification models respectively, and obtain the probability of the case belonging to each subclass output by at least two classification models; when the obtained probability is greater than a second probability threshold, it is determined that the case belongs to the subclass corresponding to the probability.

[0113] It should be noted that the first probability threshold and the second probability threshold may be the same or different. In practical applications, the second probability threshold may be appropriately adjusted according to the type of case, such as the differences among the subcategories of the non-diagnosis and treatment category.

[0114] In actual implementation, the applicant found that when the server determines that the type of case is non-diagnosis and treatment, the execution order of the judgment of each subclass does not need to be considered for further judging which subclass of the non-diagnosis and treatment class the case belongs to; that is, the four subclasses included in the non-diagnosis and treatment class can be classified and processed in any order. However, by analyzing a large amount of patient visit collection data, it can be found that a large proportion of patients' intention to come to the hospital for treatment is for follow-up visits, followed by examinations and prescriptions, and the number of incomplete information classes is almost small. Therefore, in order to improve the efficiency of the classification model in classifying and processing case features, the embodiment of the present invention uses the order of follow-up visits, examinations, prescriptions, and incomplete information to perform the classification process in sequence.

[0115] Exemplarily, the server can classify the case characteristics in the following manner: classify the feature vectors corresponding to the case characteristics through the classification model in the order of follow-up visit class, examination class, prescription class and incomplete information class, and obtain the corresponding probabilities of the case belonging to the follow-up visit class, examination class, prescription class and incomplete information class.

[0116] Step 305: Based on the judgment results of each subclass, identify the medical treatment intention corresponding to the case.

[0117] In some embodiments, the server may identify the medical treatment intention corresponding to the case based on the judgment results of each subclass in the following manner: merging the judgment results of each subclass to obtain a merged result; and then, identifying the medical treatment intention corresponding to the case based on the merged result.

[0118] Exemplarily, the server can identify the medical intention corresponding to the case based on the merging result in the following manner: based on the merging result, output each subclass for which the merging result is determined to be yes in order of priority; based on the output result, identify the medical intention corresponding to the subclass with the highest priority as the medical intention corresponding to the case. In an embodiment of the present invention, the priority order of each subclass output is: follow-up class, examination class, prescription class, and incomplete information class.

[0119] In some embodiments, the method for processing the intention to see a doctor may further include:

[0120] After identifying the medical treatment intention corresponding to the case based on the judgment results of each subclass, a medical treatment distribution structure diagram for the medical treatment object is constructed based on the medical treatment intention corresponding to the case; based on the medical treatment distribution structure diagram, corresponding medical resources are allocated to the medical treatment object.

[0121] In other embodiments, the method for processing the intention to see a doctor may further include:

[0122] After identifying the medical treatment intention corresponding to the case based on the judgment results of each subclass, different types of cases are deconstructed based on the medical treatment intention corresponding to the case to assist in the prediction and diagnosis of the disease.

[0123] By adopting the technical solution provided in the embodiment of the present invention and utilizing the characteristics of the case text itself, the medical treatment intention of the patient corresponding to the case can be automatically and accurately identified, providing available data for subsequent case structuring, auxiliary diagnosis and case association, thereby improving the efficiency of medical management.

[0124] based on Figure 3 The method for processing the intention to see a doctor is shown. The specific implementation of the method for processing the intention to see a doctor provided in the embodiment of the present invention will be described below.

[0125] See also Figure 4 , Figure 4 Another optional flowchart of the method for processing the intention to see a doctor provided in the embodiment of the present invention is provided. The method for processing the intention to see a doctor provided in the embodiment of the present invention can be applied to various types of processing devices for processing the intention to see a doctor, such as medical diagnostic equipment, desktop computers, and servers. The following still takes the processing device for processing the intention to see a doctor as an example, that is, the server executes the method for processing the intention to see a doctor provided in the embodiment of the present invention as an example. Figure 4 For the details not included in the description of each step below, you can refer to the above for understanding.

[0126] Step 401: Obtain case data.

[0127] In some embodiments, the server may obtain the case in the following manner: the receiving terminal obtains the case by calling the collection device. The case involved in the embodiments of the present invention may refer to the case written by the doctor based on the consultation of the patient during the consultation, or may refer to the case generated by the patient through the pre-consultation method, which is not limited here.

[0128] Step 402: Extract features from the acquired cases to obtain corresponding case features.

[0129] In an embodiment of the present invention, case features may include at least one of the following: number of symptoms; number of part words; case length; case keywords or case fragment hit words.

[0130] In some embodiments, the server may extract features from the case in at least one of the following ways to obtain corresponding case features: identifying symptoms in the case and counting the identified symptoms to obtain the corresponding number of symptoms; identifying part words in the case and counting the identified part words to obtain the corresponding number of part words; counting the number of words included in the case to obtain the corresponding case length; screening keywords or fragment hit words with classification characteristics in the case to obtain corresponding case keywords or case fragment hit words.

[0131] The following is an explanation of the feature extraction process when the case feature is the number of symptoms. Specifically, the server first uses a pre-generated recognition model to identify the symptoms in the case, and then counts the number of identified symptoms to obtain the corresponding number of symptoms. Here, as far as the generation of the recognition model is concerned, it can be implemented in the following way: using prior knowledge and corresponding annotations to generate a symptom vocabulary, and then using the symptom vocabulary to generate pseudo-annotated data, and then training the pseudo-annotated data through a conditional random field (CRF, Conditional Random Fields) to obtain a recognition model. The pseudo-annotated data here refers to all symptom data to be identified in the case. It should be pointed out that in the embodiment of the present invention, no distinction is made between negative symptoms and positive symptoms included in the case. In addition, symptoms written as "none" in the case can be understood as negative symptoms.

[0132] When the case feature is the number of part words, the server can use a feature extraction process similar to the above-mentioned symptom number to process. Specifically, the server first uses the recognition model to identify the part words in the case, and then counts the number of the identified part words to obtain the corresponding number of part words. Here, the part words in the case are the parts of the disease in the human body, such as the kidney, liver blood vessels, right 1 rib, etc.

[0133] When the case feature is the case length, the server may use a word count model to directly count the number of words included in the case to obtain the corresponding case length.

[0134] When the case feature is a case keyword or a case fragment hit word, the server can filter the keywords or fragment hit words with classification characteristics in the case (such as "same medical history as before", "prescription", "examination", etc.) based on different types of target models to obtain the corresponding case keywords or case fragment hit words. It should be noted that different types of target models, such as the target models for judging whether it is a diagnosis and treatment type or a follow-up type, use different keywords or case fragment hit words.

[0135] It should be noted that the server can also obtain case features through the doctor's diagnosis in the case, that is, whether the doctor's diagnosis results contain symptoms or diseases. For example, if there is no doctor's diagnosis result in a case, the corresponding case feature can be set to zero. The method for extracting symptoms or diseases in the diagnosis results here can be understood by referring to the method for extracting the number of symptoms above.

[0136] Step 403: Determine whether the case type is diagnosis and treatment based on the case characteristics. When the case type is determined to be diagnosis and treatment, execute step 404; when the case type is determined to be non-diagnosis and treatment, execute step 405.

[0137] In order to further improve the server's efficiency in identifying the patient's intention to see a doctor, the embodiment of the present invention divides the acquired cases into treatment types and non-treatment types, and then identifies the corresponding intention to see a doctor for different types of cases. The non-treatment category in the embodiment of the present invention includes multiple subcategories, and the subcategories include at least one of the following: follow-up category; examination category; prescription category; incomplete information category. It should be noted that the four subcategories included in the non-treatment category can overlap.

[0138] When the cases in the embodiments of the present invention are diagnosis and treatment cases, it can actually be understood that the purpose of the patient coming to the hospital is mainly to ask the doctor to diagnose and treat the disease. Usually, the information in the diagnosis and treatment cases is more comprehensive, and the current medical history or chief complaint in the case contains relevant symptom descriptions for the doctor's diagnosis. Diagnosis and treatment cases, for example, "repeated syncope for 4 years, each lasting about half an hour, without incontinence"; "menstrual period and menstrual cycle extended for half a year, see the test report"; "right femoral plate removal surgery, limited knee flexion for more than 10 months" and so on. It should be noted that for the judgment of diagnosis and treatment cases, it is necessary to first determine whether the patient's intention to come to the hospital requires a doctor's diagnosis, and whether there is information description of relevant symptoms in the case to provide data support for the doctor's diagnosis.

[0139] When the case in the embodiment of the present invention is a follow-up visit, it can actually be understood that the patient is not coming to the hospital for the first time for the current situation. It may be that the patient comes to the hospital for a follow-up examination after the condition has improved or worsened, or it may be a regular examination or prescription for a long-term chronic disease. The case mainly includes the situation of "incomplete medical history" and "no diagnosis required". Follow-up cases, such as "kidney stones and liver hemangioma were found during the examination"; "medical history is the same as before, no fever, cough and asthma have improved, and runny nose has decreased"; "hypertension medication"; "re-examination two months after right 1st rib fracture". It should be noted that the characteristic of the follow-up class is that the patient has a certain understanding of the condition of the disease he suffers from, and the doctor can use the previous diagnosis results. If new symptoms appear in the current case, it is necessary to consider that the case is actually a diagnosis and treatment type, not a follow-up type, such as "right femoral plate removal surgery, limited knee flexion for more than 10 months"; "half a month after liver tumor resection, abdominal pain recently, more obvious in the afternoon. Defecation once a day, rotten"; "removal of stitches"; "dressing change", etc.

[0140] When the case in the embodiment of the present invention is of the inspection type, it can actually be understood that the purpose of the patient coming to the hospital is mainly to ask the doctor to issue an inspection form. The inspection here may be a regular inspection, a health check-up, or it may be self-distrust, or further inspection and diagnosis of a certain type of disease diagnosed by other hospitals. Inspection cases, such as "request for laboratory tests. LMP2016.4.28"; "request blood lead test, no fever, no cough, good appetite, normal urination and defecation"; "high blood pressure is found, come to make an appointment for examination"; "physical examination". It should be noted that the inspection category may overlap with the follow-up category. Only when the patient actively requests an examination will the case be classified as an inspection category; when the patient does not actively request an examination, but the doctor actively prescribes an inspection for a confirmed disease, it does not belong to the inspection category. In addition, if the examination is prescribed by the doctor this time, such as "prescribing medicine after gastroscopy", it does not belong to the inspection category.

[0141] When the case in the embodiment of the present invention is a prescription drug case, it can actually be understood that the patient's purpose of coming to the hospital is mainly to ask the doctor to prescribe a medicine. The patient may also have a certain understanding of the disease or often take certain medicines, and usually does not need the doctor to make further diagnosis. Cases of prescription drugs, such as "come to get the medicine"; "the patient is not here, and the family will get the medicine on behalf of the patient"; "myasthenia gravis medicine"; "medication after gastroscopy". It should be noted that the prescription drug category may overlap with other non-diagnosis and treatment categories, especially when the patient has a more direct need for prescription drugs when coming to the hospital, then the case belongs to the prescription drug category.

[0142] When the cases in the embodiments of the present invention are of the incomplete information type, it actually means that there is no detailed description of the patient's disease in the case. For example, there is no current condition of the patient (such as symptoms or diseases) in the current medical history. It may be a simple prescription or examination, or it may be other statements that are not related to the disease or the case is too short and has no value for further processing. Cases of the incomplete information type include, for example, "vaccination"; "requires B-ultrasound examination and follicle monitoring. LMP2014.5.24"; "suture removal"; "come to make an appointment for examination"; "the patient's condition is currently stable and the patient requires medication". It should be noted that the incomplete information type may overlap with some prescriptions, examinations, and follow-up visits. The key point is that there is no specific description of the condition in the incomplete information type cases (general descriptions such as stable condition also belong to incomplete information).

[0143] In an embodiment of the present invention, the server can determine the type of case in the following manner: classify the feature vector corresponding to the case characteristics through a classification model to obtain the probability that the case output by the classification model is a diagnosis and treatment class; when the obtained probability is greater than a first probability threshold, the type of the case is determined to be a diagnosis and treatment class; when the obtained probability is less than or equal to the first probability threshold, the type of the case is determined to be a non-diagnosis and treatment class.

[0144] The classification model here can be an RNN model or a logistic regression model. The following takes the classification model as a logistic regression model as an example to illustrate the method for determining the type of the case in the embodiment of the present invention.

[0145] Exemplarily, the feature vector corresponding to the case feature is input into the logistic regression model, and the type of the case is determined by the logistic regression model. For example, the feature vector corresponding to the input case feature is classified by the following formula (1), and the probability of the case output by the logistic regression model being a diagnosis and treatment type is obtained:

[0146]

[0147] Among them, P(Y=diagnosis and treatment category) represents the probability that the case belongs to the diagnosis and treatment category; x represents the feature vector corresponding to the case characteristics; w and b represent the weights of the logistic regression model respectively.

[0148] Assuming that the first probability threshold is 0.5, when the probability P(Y=diagnosis and treatment category) obtained by the above formula (1) is greater than 0.5, the type of the case is determined to be the diagnosis and treatment category; when the probability P(Y=diagnosis and treatment category) obtained by the above formula (1) is less than or equal to 0.5, the type of the case is determined to be the non-diagnosis and treatment category. It should be pointed out that the first probability threshold can be adjusted according to actual needs and is not limited here.

[0149] Step 404: Identify that the consultation intention corresponding to the case is diagnosis and treatment, and end the current processing flow.

[0150] Step 405: Classify the case characteristics through the classification model to obtain the judgment result of whether the case belongs to each subclass included in the non-diagnosis and treatment class.

[0151] In an embodiment of the present invention, the server can classify the case characteristics in the following manner to obtain a judgment result of whether the case belongs to each subclass included in the non-diagnosis and treatment class: classify the feature vectors corresponding to the case characteristics through at least two classification models respectively, and obtain the probability of the case belonging to each subclass output by at least two classification models; when the obtained probability is greater than the second probability threshold, it is determined that the case belongs to the subclass corresponding to the probability.

[0152] The at least two classification models here can both be RNN models, or both be logistic regression models, or of course any combination of RNN models and logistic regression models. The following takes at least two classification models as logistic regression models as an example to illustrate the process of determining which subclass of the non-diagnosis and treatment class a case belongs to in an embodiment of the present invention. The above-mentioned method similar to the method for determining the diagnosis and treatment class can be used to determine each subclass.

[0153] Exemplarily, the feature vector corresponding to the case feature is input into the logistic regression model, and the type of the case is determined by the logistic regression model. For example, the feature vector corresponding to the input case feature is classified by the following formula (2), and the probability of the case output by the logistic regression model being a return visit class is obtained:

[0154]

[0155] Wherein, P(Y = follow-up class) represents the probability that the case belongs to the follow-up class; x represents the feature vector corresponding to the case feature; w and b represent the weights of the logistic regression model respectively. The first probability threshold and the second probability threshold may be the same or different. In practical applications, the second probability threshold may be appropriately adjusted according to the type of case, such as the differences among the subclasses in the non-diagnosis and treatment class.

[0156] Assuming that the second probability threshold is 0.6, similarly, when the probability P(Y=follow-up visit category) obtained by the above formula (2) is greater than 0.6, the case type is determined to be a follow-up visit category; when the probability P(Y=follow-up visit category) obtained by the above formula (2) is less than or equal to 0.6, the case type is determined to be a non-follow-up visit category. Similar judgment methods can be used to determine whether a case belongs to the examination category, the prescription category, and the incomplete information category.

[0157] In actual implementation, the applicant found that when the server determines that the type of case is non-diagnosis and treatment, the execution order of the judgment of each subclass does not need to be considered for further judging which subclass of the non-diagnosis and treatment class the case belongs to; that is, the four subclasses included in the non-diagnosis and treatment class can be classified and processed in any order. However, by analyzing a large amount of patient visit collection data, it can be found that a large proportion of patients' intention to come to the hospital for treatment is for follow-up visits, followed by examinations and prescriptions, and the number of incomplete information classes is almost small. Therefore, in order to improve the efficiency of the classification model in classifying and processing case features, the embodiment of the present invention uses the order of follow-up visits, examinations, prescriptions, and incomplete information to perform the classification process in sequence.

[0158] Exemplarily, the server can classify the case characteristics in the following manner: classify the feature vectors corresponding to the case characteristics through the classification model in the order of follow-up visit class, examination class, prescription class and incomplete information class, and obtain the corresponding probabilities of the case belonging to the follow-up visit class, examination class, prescription class and incomplete information class.

[0159] Step 406: Based on the judgment results of each subclass, identify the medical treatment intention corresponding to the case.

[0160] In some embodiments, the server may identify the medical intention corresponding to the case based on the judgment results of each subclass in the following manner: merge the judgment results of each subclass to obtain a merged result; then, identify the medical intention corresponding to the case based on the merged result. Here, in the process of the server merging the judgment results of each subclass, when the judgment results of two or more subclasses in the follow-up visit class, the examination class, the prescription class, and the incomplete information class are yes, the merged result is yes; when the judgment result of only one subclass in the follow-up visit class, the examination class, the prescription class, and the incomplete information class is yes, the merged result is no.

[0161] In an embodiment of the present invention, the server can identify the medical intention corresponding to the case based on the merging result in the following manner: based on the merging result, output each subclass for which the merging result is determined to be yes in order of priority; based on the output result, identify the medical intention corresponding to the subclass with the highest priority as the medical intention corresponding to the case. In an embodiment of the present invention, the priority order of each subclass output is: follow-up class, examination class, prescription class, and incomplete information class.

[0162] For example, when the merged result is yes, the judgment results of the follow-up visit, examination and prescription categories in each subclass are all yes, and the judgment result of the incomplete information category is no, then the follow-up visit, examination and prescription categories are output in order of priority, and the subclass with the highest priority, that is, the medical intention corresponding to the follow-up visit category, is identified as the medical intention corresponding to the case; for another example, when the merged result is yes, the judgment results of the examination and prescription categories in each subclass are yes, and the judgment results of the follow-up visit and incomplete information category are no, then the examination and prescription categories are output in order of priority, and the medical intention corresponding to the subclass with the highest priority, that is, the examination category, is identified as the medical intention corresponding to the case.

[0163] In some embodiments, the method for processing the intention of medical treatment may further include: after identifying the medical treatment intention corresponding to the case based on the judgment results of each subclass, constructing a medical treatment distribution structure diagram for the medical treatment object based on the medical treatment intention corresponding to the case; and allocating corresponding medical resources to the medical treatment object based on the medical treatment distribution structure diagram. In other embodiments, the method for processing the intention of medical treatment may further include: after identifying the medical treatment intention corresponding to the case based on the judgment results of each subclass, deconstructing different types of cases based on the medical treatment intention corresponding to the case to assist in the prediction and diagnosis of the disease.

[0164] In the embodiment of the present invention, after the patient's intention to see a doctor is identified according to the patient's case, the distribution data of the patient can be provided to the hospital according to the intention to see a doctor, and then a distribution structure diagram of the patients seeing a doctor is constructed according to the distribution data, such as Figure 5 An optional distribution structure display diagram of the distribution data of the patients provided by the embodiment of the present invention is shown, see Figure 5 It can intuitively display the proportion of patients with different types of medical treatment intentions, so as to assist hospitals in understanding the distribution of patients' medical treatment needs. When guiding patients, it can allocate corresponding numbers to simple examination and prescription cases, as well as cases that require doctor's consultation and examination. In addition, it can also allocate reasonable medical resources to each patient to achieve the rational allocation and use of medical resources.

[0165] In addition, after the patient's medical treatment intention is identified based on the patient's case, it can also be used by doctors and researchers to conduct research on disease prediction and diagnosis. For example, different types of cases can be deconstructed based on the medical treatment intention corresponding to the identified case, and the morphological manifestations of different types of cases can be determined, thereby assisting in predicting and diagnosing whether the patient is likely to have a certain disease (such as diabetes) and its risks.

[0166] By adopting the technical solution provided by the embodiment of the present invention, by extracting features from the acquired cases, the corresponding case features are obtained, and the type of the case is determined based on the case features. When the type of the case is determined to be a diagnosis and treatment type, the corresponding medical intention of the case is identified as diagnosis and treatment; when the type of the case is determined to be a non-diagnosis and treatment type, the case features are classified and processed through a classification model to obtain the judgment results of the subclasses included in the non-diagnosis and treatment type, and then the corresponding medical intention of the case is identified based on the judgment results of each subclass. In this way, not only can the medical intention of the patient corresponding to the case be automatically and accurately identified, but also available data can be provided for subsequent case structuring, auxiliary diagnosis and case association, thereby improving the efficiency of medical management; and, after the cases are divided into diagnosis and treatment types and non-diagnosis and treatment types, the medical intention is identified, which can further improve the recognition efficiency of medical intention.

[0167] The following is an example of a specific application scenario to illustrate the method for processing the medical treatment intention of the embodiment of the present invention. The specific application scenario is that a hospital wants to understand the distribution of patients' medical treatment needs. The method of the embodiment of the present invention can be used to distribute and count the patients' medical treatment intentions to the hospital, so as to allocate reasonable medical resources to the patients.

[0168] See also Figure 6 , Figure 6 Another optional flow chart of the method for processing the intention of seeing a doctor provided in the embodiment of the present invention. The method for processing the intention of seeing a doctor provided in the embodiment of the present invention can be applied to various types of processing devices for processing the intention of seeing a doctor, such as medical diagnostic equipment, desktop computers and servers. The following takes the processing device for processing the intention of seeing a doctor as a server as an example, that is, the method for processing the intention of seeing a doctor provided in the embodiment of the present invention is executed by the server as an example. Figure 6 For the details not included in the description of each step below, you can refer to the above for understanding.

[0169] Step 601: Obtain case data.

[0170] In some embodiments, the server may obtain the case in the following manner: the receiving terminal obtains the case by calling the collection device. The case involved in the embodiments of the present invention may refer to the case written by the doctor based on the consultation of the patient during the consultation, or may refer to the case generated by the patient through the pre-consultation method, which is not limited here.

[0171] Step 602: Extract features from the acquired cases to obtain corresponding case features.

[0172] In an embodiment of the present invention, case features may include at least one of the following: number of symptoms; number of part words; case length; case keywords or case fragment hit words.

[0173] Step 603: Determine whether the case type is diagnosis and treatment based on the case characteristics. When the case type is determined to be diagnosis and treatment, execute step 604; when the case type is determined to be non-diagnosis and treatment, execute step 605.

[0174] In an embodiment of the present invention, the server can determine the type of case in the following manner: classify the feature vector corresponding to the case characteristics through a classification model to obtain the probability that the case output by the classification model is a diagnosis and treatment class; when the obtained probability is greater than a first probability threshold, the type of the case is determined to be a diagnosis and treatment class; when the obtained probability is less than or equal to the first probability threshold, the type of the case is determined to be a non-diagnosis and treatment class.

[0175] Step 604: Identify that the consultation intention corresponding to the case is diagnosis and treatment, and end the current processing flow.

[0176] Step 605: Classify the feature vectors corresponding to the case features using at least two classification models respectively, and obtain the probability of the case belonging to each subclass output by at least two classification models.

[0177] Step 606: When the obtained probability is greater than the second probability threshold, it is determined that the case belongs to the subclass of the corresponding probability.

[0178] The at least two classification models here may both be RNN models, or both be logistic regression models, or of course any combination of RNN models and logistic regression models.

[0179] In order to improve the efficiency of the classification model in classifying case features, the embodiment of the present invention adopts the order of follow-up visit, examination, prescription and incomplete information to perform the classification process in sequence. Specifically, the case features can be classified in the following manner: in the order of follow-up visit, examination, prescription and incomplete information, the feature vectors corresponding to the case features are classified by the classification model respectively, and the corresponding probabilities of the case belonging to the follow-up visit, examination, prescription and incomplete information categories are obtained.

[0180] Step 607: Merge the judgment results of each subclass to obtain a merged result.

[0181] Step 608: Based on the merging result, each subclass for which the merging result is yes is output in order of priority; based on the output result, the medical treatment intention corresponding to the subclass with the highest priority is identified as the medical treatment intention corresponding to the case.

[0182] In the embodiment of the present invention, the priority order of each sub-category output is: follow-up visit category, examination category, prescription category and incomplete information category.

[0183] Step 609: Based on the medical treatment intention corresponding to the case, a medical treatment distribution structure diagram for the medical treatment object is constructed, and corresponding medical resources are allocated to the medical treatment object based on the medical treatment distribution structure diagram.

[0184] See also Figure 7 , Figure 7 An optional principle structure diagram of a method for processing medical treatment intention provided in an embodiment of the present invention is provided, with the medical treatment intention processing device being a server, and the patient's case being the patient's chief complaint or current medical history as an example for explanation. The server first obtains the patient's chief complaint or current medical history, and then performs feature extraction on the obtained patient's chief complaint or current medical history to obtain corresponding case features; next, a classification model, such as a logistic regression model, is used to classify the feature vector corresponding to the case feature to determine whether the patient's case type is a diagnosis and treatment type. When the case type is determined to be a diagnosis and treatment type, the medical treatment intention corresponding to the case is identified as diagnosis and treatment, and the medical treatment intention corresponding to the identified case is type-marked, i.e., marked as diagnosis and treatment; next, the classification model is used to classify the patient's chief complaint or current medical history in the order of follow-up visit, examination, prescription, and incomplete information. For example, the logistic regression model classifies the feature vectors corresponding to the case characteristics, that is, it determines whether the type of the case is a follow-up visit, examination, prescription, and incomplete information, and obtains the judgment results of the corresponding subclasses. Next, the judgment results of each subclass are merged to obtain a merged result, and the subclasses with a merged result of yes are output in order of priority. Finally, based on the output results, the medical treatment intention corresponding to the subclass with the highest priority is identified as the medical treatment intention corresponding to the case, and the medical treatment intention corresponding to the identified case is marked by type to form a list of medical treatment intention types.

[0185] By adopting the technical solution provided by the embodiment of the present invention and utilizing the characteristics of the case text itself, the medical treatment intention of the patient corresponding to the case can be automatically and accurately identified, and available data can be provided for subsequent case structuring, auxiliary diagnosis and case association, thereby improving the efficiency of medical management; and, after dividing the case into diagnosis and treatment types and non-diagnosis and treatment types, the medical treatment intention is identified, which can further improve the recognition efficiency of the medical treatment intention; the embodiment of the present invention can also allocate reasonable medical resources to each patient according to the identified medical treatment intention, so as to achieve the reasonable allocation and use of medical resources.

[0186] Next, the software implementation of the device 30 for processing the intention of seeing a doctor provided in the embodiment of the present invention is described. The software module included in the memory 202 in the processing device 20 for processing the intention of seeing a doctor is used as an example for description. For the details not covered in the functional description of the module below, the above can be understood by referring to the above.

[0187] An acquisition unit 31 is used to acquire cases; an extraction unit 32 is used to extract features from the cases to obtain corresponding case features; a determination unit 33 is used to determine the type of the case based on the case features; a classification unit 34 is used to classify the case features through a classification model when the type of the case is determined to be a non-diagnosis and treatment type, to obtain a judgment result of whether the case belongs to each subclass included in the non-diagnosis and treatment type; an identification unit 35 is used to identify the medical treatment intention corresponding to the case based on the judgment results of each subclass.

[0188] In some embodiments, the identification unit is further used for: when it is determined that the type of the case is diagnosis and treatment, identifying that the medical consultation intention corresponding to the case is diagnosis and treatment.

[0189] In some embodiments, the extraction unit extracts features from the case to obtain corresponding case features, which can be achieved by at least one of the following methods:

[0190] The symptoms in the case are identified, and the identified symptoms are counted to obtain the corresponding number of symptoms; the part words in the case are identified, and the identified part words are counted to obtain the corresponding number of part words; the number of words included in the case is counted to obtain the corresponding case length; the keywords or fragment hit words with classification characteristics in the case are screened to obtain the corresponding case keywords or case fragment hit words.

[0191] In some embodiments, the determination unit determines the type of the case based on the case characteristics, which can be implemented in the following manner: classifying the feature vector corresponding to the case characteristics through the classification model to obtain the probability of the case being a diagnosis and treatment type output by the classification model;

[0192] When the obtained probability is greater than a first probability threshold, determining the type of the case as diagnosis and treatment type;

[0193] When the obtained probability is less than or equal to the first probability threshold, the type of the case is determined to be non-diagnosis and treatment type.

[0194] In some embodiments, the classification unit classifies the case features through a classification model to obtain a judgment result that the case belongs to each subclass included in the non-diagnosis and treatment class, which can be achieved in the following manner: classify the feature vectors corresponding to the case features through at least two of the classification models respectively to obtain the probability of the case belonging to each subclass output by each of the at least two classification models;

[0195] When the obtained probability is greater than the second probability threshold, it is determined that the case belongs to the subclass corresponding to the probability; wherein the subclass includes at least one of the following: follow-up visit class; examination class; prescription class; incomplete information class.

[0196] Here, as for the classification unit classifying and processing the feature vectors corresponding to the case characteristics through at least two of the classification models respectively, and obtaining the probability that the case belongs to each subclass output by the at least two classification models respectively, it can be implemented in the following way: in the order of the follow-up visit class, examination class, prescription class and incomplete information class, the feature vectors corresponding to the case characteristics are classified and processed through the classification models respectively, and the corresponding probabilities of the case belonging to the follow-up visit class, examination class, prescription class and incomplete information class are obtained.

[0197] In some embodiments, as for the identification unit to identify the medical treatment intention corresponding to the case based on the judgment results of the subcategories, it can be implemented in the following way: merging the judgment results of the subcategories to obtain a merged result; based on the merged result, identifying the medical treatment intention corresponding to the case.

[0198] Here, based on the merged result, identifying the medical consultation intention corresponding to the case can be achieved in the following manner:

[0199] Based on the merged result, each subclass for which the merged result is yes is output in order of priority; based on the output result, the medical treatment intention corresponding to the subclass with the highest priority is identified as the medical treatment intention corresponding to the case.

[0200] In some embodiments, the device for processing the intention to see a doctor may further include:

[0201] A construction unit, configured to construct a medical treatment distribution structure diagram for a medical treatment object based on the medical treatment intention corresponding to the case after the identification unit identifies the medical treatment intention corresponding to the case based on the judgment results of the subclasses;

[0202] An allocation unit is used to allocate corresponding medical resources to the patient based on the patient distribution structure diagram.

[0203] In some embodiments, the device for processing the intention to see a doctor may further include:

[0204] A deconstruction unit is used to deconstruct different types of cases based on the medical treatment intentions corresponding to the cases after the identification unit identifies the medical treatment intentions corresponding to the cases based on the judgment results of the subcategories, so as to assist in the prediction and diagnosis of the disease.

[0205] The embodiment of the present invention further provides a storage medium storing executable instructions, which, when executed, are used to implement the above-mentioned method for processing the intention to see a doctor provided in the embodiment of the present invention. The storage medium may specifically be a computer-readable storage medium, such as a ferromagnetic random access memory (FRAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a flash memory, a magnetic surface memory, an optical disc or a compact disc read-only memory (CD-ROM), and the like.

[0206] In summary, the technical solution of the embodiment of the present invention has the following beneficial effects:

[0207] 1) By utilizing the characteristics of the case text itself, the patient's treatment intention can be automatically and accurately identified, providing available data for subsequent case structuring, auxiliary diagnosis, and case association, thereby improving the efficiency of medical management.

[0208] 2) Classifying cases into diagnosis and treatment types and non-diagnosis and treatment types and then identifying the intention of medical treatment can further improve the efficiency of identifying the intention of medical treatment.

[0209] 3) In order to improve the efficiency of the classification model in classifying case features, the embodiment of the present invention performs the classification process in the order of follow-up visit, examination, prescription and incomplete information.

[0210] 4) Allocate reasonable medical resources to each patient based on the identified medical treatment intention, so as to achieve the rational allocation and use of medical resources.

[0211] In the above description, the terms "first", "second", etc. are only used to distinguish similar objects and do not represent a specific order or precedence for the objects. It is understood that "first", "second", etc. can be interchanged with a specific order or precedence where permitted, so that the embodiments of the present invention described herein can be implemented in an order other than that shown or described.

[0212] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for processing medical treatment intention, characterized in that: include: Obtain case records; Extracting features from the case to obtain corresponding case features, wherein the case features include at least one of the following: the number of symptoms, the number of part words, the length of the case, case keywords, and case fragment hit words; Determining the type of the case based on the characteristics of the case; When it is determined that the type of the case is non-diagnosis and treatment, the feature vectors corresponding to the case features are classified by the classification model according to the priority order of the subclasses, and the probability of the case belonging to each subclass is obtained accordingly, wherein the subclasses include at least one of the following: follow-up class, examination class, prescription class, and incomplete information class; When the obtained probability is greater than a second probability threshold, determining the judgment result of the subclass of the probability as yes; Merging the judgment results of each subclass to obtain a combined result, wherein when the judgment results of at least two of the subclasses are yes, the combined result is yes; Outputting the subcategories for which the merge result is yes in the order of priority, wherein the order of priority is: follow-up visit category, examination category, prescription category, and incomplete information category; Based on the output result, the medical treatment intention corresponding to the subclass with the highest priority among the subclasses with the judgment result of yes is identified as the medical treatment intention corresponding to the case; Based on the medical treatment intention corresponding to the case, a medical treatment distribution structure diagram for the medical treatment object is constructed; Based on the medical consultation distribution structure diagram, corresponding medical resources are allocated to the medical consultation objects.

2. The method according to claim 1, characterized in that Determining the type of the case based on the case characteristics comprises: Classify the feature vector corresponding to the case feature through the classification model to obtain the probability that the case is a diagnosis and treatment class output by the classification model; When the obtained probability is greater than a first probability threshold, determining the type of the case as diagnosis and treatment type; When the obtained probability is less than or equal to the first probability threshold, the type of the case is determined to be non-diagnosis and treatment type.

3. The method according to claim 1, characterized in that The extracting features of the case to obtain corresponding case features includes at least one of the following: Identifying the symptoms in the case, and counting the identified symptoms to obtain the corresponding number of symptoms; Identify the part words in the case, and count the identified part words to obtain the corresponding number of part words; Counting the number of words included in the case to obtain the corresponding case length; The keywords or fragment hits with classification characteristics in the case are screened to obtain corresponding case keywords or case fragment hits.

4. The method according to claim 1, characterized in that The method further comprises: When it is determined that the type of the case is diagnosis and treatment, the consultation intention corresponding to the case is identified as diagnosis and treatment.

5. The method according to any one of claims 1 to 4, characterized in that: The method further comprises: After identifying the medical treatment intention corresponding to the subclass with the highest priority as the medical treatment intention corresponding to the case, Based on the medical treatment intentions corresponding to the cases, different types of cases are deconstructed to assist in the prediction and diagnosis of diseases.

6. A device for processing medical consultation intention, characterized in that: include: An acquisition unit, used for acquiring cases; An extraction unit, used for extracting features from the case to obtain corresponding case features, wherein the case features include at least one of the following: the number of symptoms, the number of part words, the case length, case keywords, and case fragment hit words; A determination unit, configured to determine the type of the case based on the case characteristics; A classification unit, for, when it is determined that the type of the case is a non-diagnosis and treatment type, classifying the feature vectors corresponding to the case features through the classification model according to the priority order of the subclasses, and correspondingly obtaining the probability that the case belongs to each subclass, wherein the subclasses include at least one of the following: a follow-up visit type, an examination type, a prescription type, and an incomplete information type; when the obtained probability is greater than a second probability threshold, determining the judgment result of the subclass of the probability as yes; An identification unit is used to merge the judgment results of each subclass to obtain a merged result, wherein when the judgment results of at least two of the subclasses are yes, the merged result is yes; each of the subclasses for which the merged result is yes is output in the order of priority, wherein the order of priority is: follow-up visit, examination, prescription, and incomplete information; based on the output result, the medical treatment intention corresponding to the subclass with the highest priority among the subclasses for which the judgment result is yes is identified as the medical treatment intention corresponding to the case; A construction unit, used for constructing a medical treatment distribution structure diagram for a medical treatment object based on the medical treatment intention corresponding to the case; An allocation unit is used to allocate corresponding medical resources to the patient based on the patient distribution structure diagram.

7. A device for processing medical consultation intention, characterized in that: include: A memory for storing executable instructions; A processor, used to implement the method for processing medical consultation intention as described in any one of claims 1 to 5 when executing the executable instructions stored in the memory.

8. A computer-readable storage medium, characterized in that: Executable instructions are stored, and when the executable instructions are executed, they are used to implement the method for processing the medical treatment intention described in any one of claims 1 to 5.

9. A computer program comprising executable instructions, characterized in that When the executable instructions are executed, they are used to implement the method for processing the medical treatment intention described in any one of claims 1 to 5.