WO2023029506A1 - Illness state analysis method and apparatus, electronic device, and storage medium - Google Patents

Abstract

Provided are an illness state analysis method and apparatus, an electronic device, and a storage medium. The method comprises: acquiring electronic medical record data (S101); performing entity feature extraction on the electronic medical record data to obtain target illness state information (S102); processing the target illness state information by using a pre-trained illness state recognition model to generate a treatment reminder corresponding to the target illness state information (S103); receiving treatment feedback data responded by a user side according to the treatment reminder (S104); generating a diagnosis conclusion label according to the treatment feedback data and the target illness state information (S105); and generating an illness state analysis report according to the diagnosis conclusion label (S106).

Description

Disease analysis method, device, electronic device and storage medium

This application claims the priority of the Chinese patent application with the application number 202111007202.5 and the title of the invention "disease analysis method, device, electronic equipment and storage medium" submitted to the China Patent Office on August 30, 2021, the entire contents of which are incorporated by reference in this application.

technical field

The present application relates to the field of artificial intelligence and digital medical technology, and in particular to a disease analysis method, device, electronic equipment and storage medium.

Background technique

At present, since the historical medical records of patients are often paper medical records or historical medical records cannot be shared, the inventor realized that this situation often brings inconvenience to patients' follow-up medical treatment and affects the accuracy and efficiency of disease analysis. Therefore, how to pass Analyzing the historical conditions of patients to obtain reference pathological data and improving the efficiency of condition analysis has become a technical problem that needs to be solved urgently.

technical problem

The following is an overview of the topics described in detail in this article. This summary is not intended to limit the scope of the claims.

The main purpose of the embodiments of the present application is to provide a disease analysis method, device, electronic equipment and storage medium, aiming to obtain reference pathological data by analyzing the patient's historical disease, and improve the efficiency of disease analysis.

technical solution

In the first aspect, the embodiment of the present application proposes a disease analysis method, the method comprising:

Access to electronic medical record data;

Extracting entity features from the electronic medical record data to obtain target disease information;

Processing the target condition information by using a pre-trained condition identification model to generate a treatment reminder corresponding to the target condition information;

receiving the treatment feedback data that the user terminal responds to according to the treatment reminder;

generating a diagnosis conclusion label according to the treatment feedback data and the target condition information;

A condition analysis report is generated according to the diagnostic conclusion label.

In the second aspect, the embodiment of the present application proposes a disease analysis device, which includes:

Electronic medical record data acquisition module, used to obtain electronic medical record data;

A feature extraction module, configured to extract entity features from the electronic medical record data to obtain target disease information;

A processing module, configured to use a pre-trained disease recognition model to process the target disease information, and generate a treatment reminder corresponding to the target disease information;

The treatment feedback data receiving module is used to receive the treatment feedback data that the client responds to according to the treatment reminder;

A diagnosis conclusion label generating module, configured to generate a diagnosis conclusion label according to the treatment feedback data and the target condition information;

A condition analysis report generating module, configured to generate a condition analysis report according to the diagnosis conclusion label.

In the third aspect, the embodiment of the present application provides an electronic device, the electronic device includes a memory, a processor, a program stored in the memory and operable on the processor, and a program for implementing the processor A data bus connecting and communicating with the memory, when the program is executed by the processor, a disease analysis method is implemented, wherein the disease analysis method includes: acquiring electronic medical record data; Perform entity feature extraction to obtain target condition information; use a pre-trained condition recognition model to process the target condition information to generate a treatment reminder corresponding to the target condition information; receive treatment feedback from the user terminal based on the treatment reminder response data; generate a diagnosis conclusion label according to the treatment feedback data and the target condition information; generate a condition analysis report according to the diagnosis conclusion label.

In the fourth aspect, the embodiment of the present application provides a computer-readable storage medium for computer-readable storage, the computer-readable storage medium stores one or more programs, and the one or more programs can be stored by one Or executed by multiple processors to implement a disease analysis method, wherein the disease analysis method includes: obtaining electronic medical record data; performing entity feature extraction on the electronic medical record data to obtain target disease information; using pre-trained disease The recognition model processes the target condition information to generate a treatment reminder corresponding to the target condition information; receives treatment feedback data from the user end according to the treatment reminder response; according to the treatment feedback data and the target condition information, Generate a diagnosis conclusion label; generate a condition analysis report according to the diagnosis conclusion label.

Beneficial effect

The disease analysis method, device, electronic equipment and storage medium proposed in this application obtain electronic medical record data and perform entity feature extraction on electronic medical record data to obtain target disease information. This method can realize feature extraction of electronic medical record data. Reducing the total amount of data makes it easier to extract the required disease information; then use the pre-trained disease identification model to process the target disease information, and generate treatment reminders corresponding to the target disease information, so that users can take medication according to the treatment reminder or Seek medical attention. Then receive the treatment feedback data from the user terminal according to the treatment reminder response, analyze the treatment feedback data and the target condition information, and obtain the diagnosis conclusion label, which can evaluate the treatment effect more conveniently; finally, generate the condition analysis report according to the diagnosis conclusion label, through the The user's historical condition is analyzed to obtain a condition analysis report, which provides reference pathological data for the subsequent diagnosis process, improves the efficiency of condition analysis, and can also reduce the cost of medical treatment and medication for patients.

Additional features and advantages of the application will be set forth in the description which follows, and, in part, will be obvious from the description, or may be learned by practice of the application. The objectives and other advantages of the application will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

Description of drawings

The accompanying drawings are used to provide a further understanding of the technical solution of the present application, and constitute a part of the specification, and are used together with the embodiments of the present application to explain the technical solution of the present application, and do not constitute a limitation to the technical solution of the present application.

Fig. 1 is the flowchart of the condition analysis method provided by the embodiment of the present application;

Fig. 2 is the flowchart of step S102 in Fig. 1;

Fig. 3 is the flowchart of step S103 in Fig. 1;

Fig. 4 is a partial flowchart of a disease analysis method provided by another embodiment of the present application;

Fig. 5 is the flowchart of step S105 in Fig. 1;

Fig. 6 is the flowchart of step S106 in Fig. 1;

Fig. 7 is a schematic structural diagram of a disease analysis device provided by an embodiment of the present application;

FIG. 8 is a schematic diagram of a hardware structure of an electronic device provided by an embodiment of the present application.

Embodiment of this application

In order to make the purpose, technical solution and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application, not to limit the present application.

It should be noted that although the functional modules are divided in the schematic diagram of the device, and the logical sequence is shown in the flowchart, in some cases, it can be executed in a different order than the module division in the device or the flowchart in the flowchart. steps shown or described. The terms "first", "second" and the like in the specification and claims and the above drawings are used to distinguish similar objects, and not necessarily used to describe a specific sequence or sequence.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application, and are not intended to limit the present application.

First, analyze some nouns involved in this application:

Artificial Intelligence (AI): It is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence; artificial intelligence is a branch of computer science. Intelligence attempts to understand the essence of intelligence and produce a new intelligent machine that can respond in a manner similar to human intelligence. Research in this field includes robotics, language recognition, image recognition, natural language processing, and expert systems. Artificial intelligence can simulate the information process of human consciousness and thinking. Artificial intelligence is also a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

Natural language processing (NLP): NLP uses computers to process, understand and use human languages (such as Chinese, English, etc.). NLP belongs to a branch of artificial intelligence and is an interdisciplinary subject between computer science and linguistics. Known as computational linguistics. Natural language processing includes syntax analysis, semantic analysis, text understanding, etc. Natural language processing is often used in technical fields such as machine translation, handwritten and printed character recognition, speech recognition and text-to-speech conversion, information retrieval, information extraction and filtering, text classification and clustering, public opinion analysis and opinion mining. It involves language processing Related data mining, machine learning, knowledge acquisition, knowledge engineering, artificial intelligence research and linguistics research related to language computing, etc.

Information Extraction (Information Extraction, NER): A text processing technology that extracts specified types of factual information such as entities, relationships, and events from natural language texts, and forms structured data output. Information extraction is a technique to extract specific information from text data. Text data is composed of some specific units, such as sentences, paragraphs, and chapters. Text information is composed of some small specific units, such as words, words, phrases, sentences, paragraphs, or combinations of these specific units. . Extracting noun phrases, personal names, and place names in text data is all text information extraction. Of course, the information extracted by text information extraction technology can be various types of information.

Association analysis: Association analysis is a simple and practical analysis technique, which is to discover the association or correlation existing in a large number of data sets, so as to describe the laws and patterns of the simultaneous appearance of certain attributes in a thing. Association analysis is to discover interesting associations and correlations between item sets from a large amount of data.

(1) Support count: An item set appears in several transactions, and its support count is how many.

(2) Support: support count divided by the total number of transactions.

(3) Confidence: For the rule {Diaper}→{Beer}, the support count of {Diaper, Beer} divided by the support count of {Diaper} is the confidence of this rule.

(4) Strong association rules: Rules greater than or equal to the minimum support threshold and minimum confidence threshold are called strong association rules. The ultimate goal of association analysis is to find out strong association rules.

(5) Apriori algorithm: It is a basic algorithm for mining frequent itemsets required to generate Boolean association rules. Apriori property: Any subset of a frequent itemset should also be a frequent itemset. Prove that by definition, if an itemset I does not meet the minimum support threshold min_sup, then I is not frequent, that is, P(I)<min_sup. If an item A is added to the item set I, the resulting new item set (I∪A) is not frequent, and the number of occurrences in the entire transaction database cannot be more than the number of occurrences of the original item set I, so P (I∪A)<min_sup, that is (I∪A) is not frequent. In this way, it can be easily confirmed that the Apriori property holds according to the inverse axiom.

(6) FP-growth algorithm (Frequent Pattern Tree, referred to as FP-tree;): It is an algorithm for discovering frequent patterns based on frequent pattern trees. In the FP-growth algorithm, by scanning the transaction database twice, each transaction The included frequent items are compressed and stored in FP-tree in descending order of their support. In the process of discovering frequent patterns in the future, there is no need to scan the transaction database, but only to search in the FP-Tree, and the frequent pattern is directly generated by recursively calling the FP-growth method, so the whole discovery process is also Candidate patterns need not be generated.

Collaborative filtering algorithm: It is a relatively well-known and commonly used recommendation algorithm. It discovers the user's preferences based on the mining of user historical behavior data, and predicts the products that users may like to recommend, or finds similar users (based on users) or Items (based on items). The realization of the user-based collaborative filtering algorithm mainly needs to solve two problems. One is how to find people who have similar hobbies as you, that is, to calculate the similarity of data.

BERT (Bidirectional Encoder Representations from Transformers): It is a language representation model (language representation model). BERT uses the Transformer Encoder block for connection, which is a typical two-way encoding model.

Multilayer Perceptron (MLP): MLP is a forward-structured artificial neural network that maps a set of input vectors to a set of output vectors. An MLP can be viewed as a directed graph consisting of multiple layers of nodes, each fully connected to the next layer. Except for the input node, each node is a neuron (or processing unit) with a nonlinear activation function. MLP is an extension of the perceptron, which overcomes the weakness that the perceptron cannot recognize linearly inseparable data. The simplest MLP is a three-layer structure (input layer-hidden layer-output layer). The layers of the multi-layer perceptron are fully connected, that is, any neuron in each layer is connected to all neurons in the previous layer. This connection actually represents a weight summation.

Error backpropagation algorithm (Backpropagation, abbreviated as BP): It is a common method used in combination with optimization methods (such as gradient descent method) to train artificial neural networks. This method computes the gradient of the loss function for all weights in the network. This gradient is fed back to the optimization method to update the weights to minimize the loss function. BP algorithm is suitable for a learning algorithm of multi-layer neural network, which is based on the gradient descent method. The input-output relationship of the BP network is essentially a mapping relationship: the function completed by a BP neural network with n inputs and m outputs is a continuous mapping from n-dimensional Euclidean space to a finite field in m-dimensional Euclidean space. Mapping is highly non-linear. Its information processing ability comes from the multiple compounding of simple nonlinear functions, so it has a strong ability to reproduce functions. The learning process of BP algorithm is composed of forward propagation process and back propagation process. In the forward propagation process, the input information is processed layer by layer through the hidden layer through the input layer and transmitted to the output layer. If the desired output value cannot be obtained at the output layer, take the sum of the squares of the output and the expected error as the objective function, transfer to backpropagation, and calculate the partial derivative of the objective function with respect to the weight of each neuron layer by layer to form the objective The gradient of the function to the weight vector is used as the basis for modifying the weight, and the learning of the network is completed in the process of modifying the weight. When the error reaches the expected value, the network learning ends.

Logistic function (Logistic function or Logistic curve): Logistic function or Logistic curve is a common S-shaped function, and the generalized Logistic curve can imitate the S-shaped curve of population growth (P) in some cases. The initial stage is roughly exponential growth; then the increase slows as the initiation becomes saturated; finally, the increase stops when maturity is reached.

The embodiments of the present application may acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is a theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. .

Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation/interaction systems, and mechatronics. Artificial intelligence software technology mainly includes computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning/deep learning.

The embodiment of the present application can realize the analysis of the patient's historical condition based on the medical cloud technology. Among them, medical cloud refers to the use of "cloud computing" to create a medical and health service cloud based on new technologies such as cloud computing, mobile technology, multimedia, 4G communication, big data, and the Internet of Things, combined with medical technology. The platform realizes the sharing of medical resources and the expansion of medical coverage. Because of the combination of cloud computing technology, medical cloud improves the efficiency of medical institutions and facilitates residents to seek medical treatment. For example, appointment registration, electronic medical records, and medical insurance in hospitals are all products of the combination of cloud computing and the medical field. Medical cloud also has the advantages of data security, information sharing, dynamic expansion, and overall layout.

Based on this, the embodiment of the present application provides a condition analysis method, device, electronic equipment and storage medium, which can obtain a condition analysis report by analyzing the patient's historical condition, and provide reference pathological data for the subsequent diagnosis process, improving the The efficiency of disease analysis can also reduce the cost of medical treatment and medication for patients.

The disease analysis method, device, electronic device, and storage medium provided in the embodiments of the present application are specifically described through the following embodiments. First, the disease analysis method in the embodiments of the present application is described.

The disease analysis method provided in the embodiment of the present application relates to the fields of artificial intelligence and digital medical technology. The disease analysis method provided in the embodiment of the present application can be applied to a terminal, can also be applied to a server, and can also be software running on a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, etc.; the server end can be configured as an independent physical server, or can be configured as a server cluster or a distributed system composed of multiple physical servers, or It can be configured as a cloud that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server; the software can be the application of the disease analysis method, etc., but is not limited to the above forms.

Fig. 1 is an optional flowchart of the disease analysis method provided by the embodiment of the present application. The method in Fig. 1 may include but not limited to include steps S101 to S106.

Step S101, obtaining electronic medical record data;

Step S102, extracting entity features from electronic medical record data to obtain target disease information;

Step S103, using the pre-trained disease identification model to process the target disease information, and generate a treatment reminder corresponding to the target disease information;

Step S104, receiving the treatment feedback data that the client responds to according to the treatment reminder;

Step S105, generating a diagnosis conclusion label according to the treatment feedback data and the target condition information;

Step S106, generating a disease analysis report according to the diagnosis conclusion label.

After the above steps S101 to S106, firstly, an electronic medical record is generated according to the personal historical medical records uploaded by the patient, and the electronic medical record data is obtained, wherein the electronic medical record data includes patient information, medical records, medication records and so on. The entity feature extraction is performed on the electronic medical record data to obtain the target condition information, among which the target condition information includes the patient's age, gender, basic health indicators, medical diagnosis data, medication records, adverse drug reaction information, operation records, etc. This method can realize the feature extraction of electronic medical record data, reduce the total amount of data, and make it easier to extract the required disease information. Use the pre-trained condition recognition model to process the above target condition information, identify the patient's historical target condition, generate a treatment reminder corresponding to the target condition information, and feed back this treatment reminder to the patient, so that the patient can follow the treatment reminder Seek medical attention or take medication. After the treatment reminder is fed back to the patient, it is also necessary to obtain treatment feedback from the patient, for example, the treatment feedback data that the user terminal responds to according to the treatment reminder can be received. By analyzing the treatment feedback data and target condition information, the diagnostic conclusion label can be obtained, and the treatment effect can be evaluated more conveniently; it should be noted that when analyzing the treatment feedback data and target condition information, a big data analysis model can be used To match the treatment feedback data with the target condition information, generate the corresponding diagnostic conclusion label according to the matching situation, and finally generate the condition analysis report according to the diagnosis conclusion label, and obtain the condition analysis report by analyzing the patient's historical condition, which will be used for the subsequent diagnosis process Provide reference pathological data, improve the efficiency of disease analysis, and also reduce the cost of medical treatment and medication for patients.

In some medical application scenarios, in a possible implementation manner, the above data is medical data, such as personal health records, prescriptions, examination reports and other data.

Referring to FIG. 2, in some embodiments, step S102 may include but not limited to include steps S201 to S205:

Step S201, extracting natural language text in electronic medical record data;

Step S202, using a preset lexical analysis model to identify entity features in the natural language text;

Step S203, segmenting the natural language text to obtain a natural language vocabulary sequence;

Step S204, constructing a list of feature sequences according to entity features and natural language vocabulary sequences;

Step S205, determine the target condition information according to the feature sequence table.

Specifically, the unstructured data in the electronic medical record data is first converted into unified structured data, and the required natural language text is extracted from the structured data. Use the preset lexical analysis model to identify entity features in natural language text. For example, a medical thesaurus is pre-built, and the medical thesaurus may include medical names, medical terms, non-medical names, folk names, international medical terms, etc. related to various medical pathologies. Through this medical lexicon, the preset lexical analysis model can enumerate medical specific names. Input the natural language text into the preset lexical analysis model, and identify the entity features in the natural language text through the medical specific names contained in the preset lexical analysis model and the preset part-of-speech categories, and the entity features may include The above-mentioned multi-dimensional entity vocabulary related to medical pathology, such as medical names, medical terms, non-medical names, folk names, international medical terms, modifiers, and time information. In order to extract entity features more accurately, a sequence classifier can also be built based on the bi-LSTM algorithm. In the model based on the bi-LSTM algorithm, the word wi and character embedding are input, and the long and short memory from left to right and the length from right to left are used. time memory, so that a single output layer is generated where the outputs are concatenated. The sequence classifier can pass the input entity features directly to the softmax classifier through this output layer, and create a probability distribution on the preset label through the softmax classifier, so as to mark and classify the entity parameters according to the probability distribution, and finally The entity features after the classification process are subjected to feature extraction to obtain the required entity features. In addition, in order to achieve data storage, the BERT encoder can also be used to convert the entity feature string from text form to encoded form through the preset encoding function to realize the storage of entity features. In addition, in addition to extracting entity features from natural language texts, it is also necessary to segment natural language texts to obtain natural language vocabulary sequences. For example, according to the part-of-speech condition, a corresponding part-of-speech is assigned to a word segment in the natural language text, such as noun, verb, adjective, and so on. According to the root and part of speech of the basic natural language, the continuous natural language text is segmented into natural language lexical sequences with semantic rationality and integrity. Furthermore, according to the corresponding relationship between entity features and natural language lexical sequences in parts of speech and roots, a feature sequence table is constructed, and the data in the electronic medical record data can be searched and compared according to the feature sequence table, and the target disease information can be determined more conveniently. .

In a possible implementation, the above-mentioned natural language text is a medical text, and the medical text can be an electronic healthcare record (Electronic Healthcare Record), an electronic personal health record, including medical records, electrocardiograms, medical images, etc. An electronic record of value.

Referring to FIG. 3, in some embodiments, step S103 may include but not limited to include step S301 or step S303:

Step S301, inputting the target condition information into the pre-trained condition identification model;

Step S302, performing quadrant partitioning on the target disease information through the disease identification model to obtain the partitioned disease information;

Step S303 , performing a fitting process on the partitioned disease information of each quadrant to generate a medication reminder or a medical reminder.

In some embodiments, the BP algorithm can be used for deep learning to establish a disease recognition model, and the target disease information is input into the pre-trained disease recognition model. The information is divided into tasks, and the condition information of the partitions is obtained. For example, according to the four-quadrant rule, the disease data in the target disease information is prioritized to obtain partitioned disease information of different importance levels. Furthermore, the condition information of each quadrant is fitted through the fitting function, and the fitting result is output to the appearance layer for the judgment of accuracy and excellent agent for iterative optimization. When the iterative optimization meets the preset When the iterative condition is met, the iterative optimization is stopped, the fitting result is output, and the corresponding medication reminder or medical reminder is generated according to the fitting result. It should be noted that the preset iteration condition may be that the number of iterations reaches the preset number, etc., or other conditions, and is not limited thereto. In addition, the disease recognition model is a multi-layer neural network, which contains two layers of processing units and two hidden layers, and each feedback can only be sent to the previous output layer or hidden layer. The multi-layer neural network is a neural network that uses a backward propagation algorithm to learn classification or prediction. Specifically, the construction process of the disease recognition model is as follows:

The first step: take the first training ancestor X={1,0,1} and its class label is 1; initialize all weights and biases of the neural network; the weights of the network are generally initially small random numbers (for example, -1.0 to 1.0 );

The second step: loop each training tuple ancestor under the termination loop condition, set X={1,0,1}, and its class label Y is 1;

The third step: cyclic input unit, wherein, the input of the input unit = output, and I ₁ =O ₁ =1; I ₂ =O ₂ =0; I ₃ =O ₂ =1; the fourth step: calculate the hidden layer or The input and output of the output layer, wherein, the input formula is as shown in formula (1), and the output formula is as shown in formula (2), according to the input and output formulas, the input I _j and output θ _j of the preset node can be obtained respectively;

I _j = Σ _i w _ij O _i + θ _j Formula (1)

Among them, w _ij is the weight of the connection from unit i to unit j in the previous layer; O _i is the output of unit i in the previous layer, and θ _j is the bias of unit j, which is used as a threshold to change the unit’s active;

O _j ＝1+(1+E[I _j ]) Formula (2)

Step 5: Calculate the errors of hidden nodes and output nodes respectively, where the output layer error formula is shown in formula (3); the hidden layer formula is shown in formula (4);

Among them, O _j is the actual output of unit j, and T _j is the known target value of j given the training tuple. It should be noted that O _j (1-O _j ) is the derivative of the logistic function;

where _wjk is the weight of the connection from unit k to unit j in the next higher layer, and

is the error of unit k.

Step 6: Extract the eliminated agents and recycle the corresponding root parts of speech after segmenting the corresponding natural language text.

Through the above process, a condition identification model that meets the requirements can be constructed, and then the reference condition information is input into the condition identification model, and the reference condition information is divided into quadrants through the condition identification model, and the reference condition information is assigned tasks according to the four-quadrant rule. Differentiate, and classify the disease data in the reference disease information according to priority, and obtain the partitioned disease information of different importance levels. Furthermore, the fitting function is used to fit the partition reference disease information of each quadrant, and the fitting result is output to the appearance layer for the judgment of accuracy and excellent agent for iterative optimization. When the iterative optimization meets the expected When the iterative condition is set, the iterative optimization is stopped, so as to complete the training of the pathological recognition model. Finally, the patient's historical target condition is identified through the trained condition recognition model. Specifically, the target condition information is input into the trained condition recognition model to obtain the fitting result, and the target condition information is generated according to the fitting result. The corresponding treatment reminder is given, and the treatment reminder is fed back to the patient, so that the patient can seek medical treatment or take medicine according to the treatment reminder.

Referring to FIG. 4, in some embodiments, after step S104, the method may include, but is not limited to, step S401 to step S402:

Step S401, analyzing the drug data in the target condition information to obtain drug reaction information; wherein, the drug reaction information includes adverse drug reaction information;

Step S402, identifying abnormal medication data in the treatment feedback data according to the adverse drug reaction information.

Specifically, in order to better analyze the patient's historical condition, it is also necessary to analyze the drug data in the target condition information, and extract the drug data contained in the patient's historical medical records and historical medication records, which can be based on key The automatic text generation model of words is used to extract drug data, and the automatic text generation model can perform different data processing according to the type of input data. Input keywords or text sentences or fields in historical medical records and historical medication records to the pre-trained text automatic generation model, if the input keywords, text sentences or fields can match the preset reference text, it indicates The current input meets the requirements. If the current input is a keyword, select the same sentence set as the input keyword in the basic corpus, and generate the corresponding drug data field according to this sentence set. If the current input is a text sentence or field, you need to select a candidate sentence in the basic corpus, and determine whether the selected candidate sentence meets the requirements; wherein, the candidate sentence is that the similarity between the basic corpus and the input text sentence or field is greater than the preset Threshold sentence, if the selected candidate sentence meets the requirements, the drug data field will be directly generated according to the candidate sentence; if the selected candidate sentence does not meet the requirements, the sentence supplement will be performed on the candidate sentence, for example, filling in synonyms, or according to the corresponding input information The candidate sentence is copied and supplemented, and the drug data field is generated according to the supplemented candidate sentence. Furthermore, based on relevant data in domestic and foreign literature, medical databases, medical data platforms, etc., these drug data fields are analyzed to obtain drug response information, which includes adverse drug reaction information, drug efficacy information, and medication guidance information. wait. Furthermore, by comparing and analyzing the adverse drug reaction information and the treatment feedback data, if the treatment feedback data involves data that matches the adverse drug reaction information, it can be determined that the data is abnormal drug use data, which can be more conveniently identified Abnormal medication data, and mark the abnormal medication data in the treatment feedback data, so as to play a prompt role in subsequent diagnosis and treatment, avoid the occurrence of abnormal medication again, and improve the reliability of medication.

In addition, in some specific embodiments, after step S104, the method also includes, but is not limited to:

Visualize the treatment feedback data and target disease information to generate a visualized data map.

Specifically, the treatment feedback data and target condition information can be written into local files through the logback component (open source log component) in the data management layer, and the interface log including treatment feedback data and target condition information can be recorded through the interceptor. The log collection system (Flume) in the management layer visualizes the treatment feedback data and the target disease information respectively to obtain the corresponding target visualization data, and then inputs these target visualization data into the Hive database and the HBase database in the data management layer for integration , and finally import the target visualization data from the PostgreSql database into the Hive database through the Sqoop transmission component in the data management layer, generate corresponding charts, and sort the series of charts according to the preset part-of-speech ranking sequence to generate visual data picture. For example, the preset part-of-speech level is that the level of nouns is higher than that of verbs, and the level of verbs is higher than that of adjectives. Through the visualized data graph, the historical condition of the patient can be presented more conveniently, so that the patient and the doctor can consult it.

Referring to FIG. 5, step S105 in some embodiments may include, but is not limited to, step S501 to step S502:

Step S501, encoding treatment feedback data and target condition information respectively to obtain treatment feedback data in encoded form and target condition information in encoded form;

Step S502, using the preset big data analysis model to perform data analysis on the coded treatment feedback data and the coded target condition information to generate a diagnostic conclusion label.

In some embodiments, the treatment feedback data and the target condition information can be respectively encoded by a preset encoder, and the preset encoder can be a BERT-based encoder, that is, by obtaining the treatment feedback data and the target condition information , and tokenize treatment feedback data and target condition information, build a BERT token generator, pre-train the BERT token generator, and form a BERT encoder that meets the requirements, so that the BERT encoder can pass the preset The encoding function converts the treatment feedback data and the target condition information from the text form into the code form, and obtains the treatment feedback data in the code form and the target condition information in the code form. Furthermore, the collaborative filtering algorithm in the preset big data analysis model is used to calculate the similarity between the treatment feedback data in the coded form and the target disease information in the coded form. According to the degree of similarity, the treatment feedback data in the coded form and the target disease information in the coded form are correlated and matched to generate corresponding diagnostic conclusion labels. It should be noted that the collaborative filtering algorithm may be a Jaccard similarity coefficient method, an included angle cosine method, or a similarity measurement method such as Euclidean distance or Manhattan distance, without limitation. In some other embodiments, the preset big data analysis model may also use an association analysis algorithm to associate and match the encoded treatment feedback data and the encoded target disease information. Commonly used correlation analysis algorithms include Apriori algorithm, FP-growth algorithm and so on. This method improves the efficiency of data analysis, and also improves the matching accuracy of treatment feedback data and target disease information.

Referring to FIG. 6, in some embodiments, step S106 may include but not limited to include steps S601 to S603:

Step S601, purifying the diagnostic conclusion label according to the target condition information to obtain the purified diagnostic conclusion label;

Step S602, verifying and analyzing the purified diagnosis conclusion label to obtain a standard diagnosis conclusion label;

Step S603, generating a disease analysis report according to the standard diagnosis conclusion label.

Specifically, in order to improve the accuracy of the diagnostic conclusion label, the select statement function can also be preset, fill in the required option data in the preset select statement function, and use the select statement function filled with option data to extract the target disease information. The data set is compared with the diagnosis conclusion label, and the abnormal diagnosis conclusion label is eliminated to obtain the purified diagnosis conclusion label. Among them, the judgment of the abnormal diagnosis conclusion label can be determined according to the similarity between the data set and the diagnosis conclusion label. If the similarity between the two is less than the preset similarity threshold, the diagnosis conclusion label is determined to be an abnormal diagnosis conclusion label. Furthermore, the purified diagnostic conclusion label is verified and analyzed, the target condition information and treatment feedback data corresponding to the purified diagnostic conclusion label are reviewed, and the target condition information and treatment feedback data are corrected and adjusted to obtain a standard diagnosis conclusion label. Finally, the condition analysis report is generated according to the standard diagnosis conclusion label, and the condition analysis report is obtained by analyzing the patient's historical condition, which provides reference pathological data for the subsequent diagnosis process, improves the efficiency of condition analysis, and can also reduce the cost of medical treatment for patients and drug costs.

In the embodiment of the present application, by obtaining electronic medical record data, extracting entity features from electronic medical record data, and obtaining target disease information, this method can realize feature extraction of electronic medical record data, reduce the total amount of data, and make it more convenient to extract the required data. The patient’s condition information; and then use the pre-trained condition recognition model to process the target condition information, and generate a treatment reminder corresponding to the target condition information, so that the patient can take medication or seek medical treatment according to the treatment reminder. Then receive the treatment feedback data from the user terminal according to the treatment reminder response, analyze the treatment feedback data and the target condition information, and obtain the diagnosis conclusion label, which can evaluate the treatment effect more conveniently; finally, generate the condition analysis report according to the diagnosis conclusion label, through the The patient's historical condition is analyzed to obtain a condition analysis report, which provides reference pathological data for the subsequent diagnosis process, improves the efficiency of condition analysis, and can also reduce the cost of medical treatment and medication for patients.

Please refer to Figure 7, the embodiment of the present application also provides a disease analysis device, which can implement the above disease analysis method, the device includes:

Electronic medical record data acquisition module 701, used to acquire electronic medical record data;

A feature extraction module 702, configured to extract entity features from electronic medical record data to obtain target disease information;

The processing module 703 is configured to use the pre-trained disease recognition model to process the target disease information, and generate a treatment reminder corresponding to the target disease information;

The treatment feedback data receiving module 704 is used to receive the treatment feedback data that the client responds to according to the treatment reminder;

The diagnosis conclusion label generation module 705 is used for generating diagnosis conclusion labels according to the treatment feedback data and the target condition information;

The disease analysis report generating module 706 is configured to generate a disease analysis report according to the diagnosis conclusion label.

The embodiment of the present application also provides an electronic device, the electronic device includes: a memory, a processor, a program stored in the memory and operable on the processor, and a data bus for realizing connection and communication between the processor and the memory , when the program is executed by the processor, the above disease analysis method is realized. The electronic device may be any intelligent terminal including a tablet computer, a vehicle-mounted computer, and the like.

Please refer to FIG. 8. FIG. 8 illustrates a hardware structure of an electronic device in another embodiment. The electronic device includes:

The processor 801 may be implemented by a general-purpose CPU (Central Processing Unit, central processing unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute related programs to realize The technical solutions provided by the embodiments of the present application;

The memory 802 may be implemented in the form of a read-only memory (ReadOnlyMemory, ROM), a static storage device, a dynamic storage device, or a random access memory (RandomAccessMemory, RAM). The memory 802 can store operating systems and other application programs. When implementing the technical solutions provided by the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 802, and are invoked by the processor 801 to execute a condition Analytical method;

The input/output interface 803 is used to realize information input and output;

The communication interface 804 is used to realize the communication interaction between the device and other devices, and the communication can be realized through a wired method (such as USB, network cable, etc.), or can be realized through a wireless method (such as a mobile network, WIFI, Bluetooth, etc.); and

A bus 805, which transmits information between various components of the device (such as a processor 801, a memory 802, an input/output interface 803, and a communication interface 804);

The processor 801 , the memory 802 , the input/output interface 803 and the communication interface 804 are connected to each other within the device through the bus 805 .

Wherein, the condition analysis method provided by the embodiment of the present application includes:

Access to electronic medical record data;

Extract entity features from electronic medical record data to obtain target disease information;

Use the pre-trained disease recognition model to process the target disease information and generate treatment reminders corresponding to the target disease information;

Receive the treatment feedback data that the client responds to according to the treatment reminder;

Generate diagnostic conclusion labels based on treatment feedback data and target condition information;

Generate a condition analysis report based on the diagnostic conclusion label.

An embodiment of the present application also provides a computer-readable storage medium for computer-readable storage. The computer-readable storage medium may be non-volatile or volatile. The computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement a disease analysis method, wherein the disease analysis method includes: obtaining electronic medical record data; Extract entity features from medical record data to obtain target condition information; use pre-trained condition recognition model to process target condition information to generate treatment reminders corresponding to target condition information; receive treatment feedback data from user terminals based on treatment reminder responses; Feedback data and target condition information to generate a diagnosis conclusion label; generate a condition analysis report based on the diagnosis conclusion label.

As a non-transitory computer-readable storage medium, memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory optionally includes memory located remotely from the processor, and these remote memories may be connected to the processor via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The embodiments described in the embodiments of the present application are to illustrate the technical solutions of the embodiments of the present application more clearly, and do not constitute a limitation to the technical solutions provided by the embodiments of the present application. Those skilled in the art know that with the evolution of technology and new For the emergence of application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

Those skilled in the art can understand that the technical solutions shown in Figures 1-6 do not constitute a limitation to the embodiments of the present application, and may include more or fewer steps than those shown in the illustrations, or combine certain steps, or be different A step of.

The device embodiments described above are only illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, the functional modules/units in the system, and the device can be implemented as software, firmware, hardware, and an appropriate combination thereof.

In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, each unit may exist separately physically, or two or more units may be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

If the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or part of the contribution to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , including multiple instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method in each embodiment of the present application. The aforementioned storage media include: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), magnetic disk or optical disc, etc., which can store programs. medium.

The preferred embodiments of the embodiments of the present application have been described above with reference to the accompanying drawings, which does not limit the scope of rights of the embodiments of the present application. Any modifications, equivalent replacements and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall fall within the scope of rights of the embodiments of the present application.

Claims (20)

1. A disease analysis method, wherein the method comprises:

   Access to electronic medical record data;

   Extracting entity features from the electronic medical record data to obtain target disease information;

   Processing the target condition information by using a pre-trained condition identification model to generate a treatment reminder corresponding to the target condition information;

   receiving the treatment feedback data that the user terminal responds to according to the treatment reminder;

   generating a diagnosis conclusion label according to the treatment feedback data and the target condition information;

   A condition analysis report is generated according to the diagnostic conclusion label.
2. The disease analysis method according to claim 1, wherein the step of extracting entity features from the electronic medical record data to obtain target disease information includes:

   Extracting natural language text in the electronic medical record data;

   Using a preset lexical analysis model to identify entity features in the natural language text;

   Segmenting the natural language text to obtain a natural language vocabulary sequence;

   Constructing a feature sequence table according to the entity feature and the natural language vocabulary sequence;

   Determine the target condition information according to the feature sequence table.
3. The disease analysis method according to claim 1, wherein the step of using a pre-trained disease identification model to process the target disease information and generate a treatment reminder corresponding to the target disease information includes:

   Inputting the target condition information into the pre-trained condition identification model;

   performing quadrant partitioning on the target disease information through the disease identification model to obtain partitioned disease information;

   The condition information of each quadrant is fitted and processed to generate medication reminders or medical reminders.
4. The disease analysis method according to claim 1, wherein, after the step of receiving the treatment feedback data that the client responds to according to the treatment reminder, the method further comprises:

   Analyzing the drug data in the target condition information to obtain drug reaction information; wherein the drug reaction information includes adverse drug reaction information;

   Identifying abnormal medication data in the treatment feedback data according to the adverse drug reaction information.
5. The disease analysis method according to any one of claims 1 to 4, wherein, after the step of receiving the treatment feedback data that the client responds to according to the treatment reminder, the method further includes:

   The treatment feedback data and the target condition information are visualized to generate a visualized data map.
6. The condition analysis method according to any one of claims 1 to 4, wherein the step of generating a diagnostic conclusion label according to the treatment feedback data and the target condition information includes:

   Coding the treatment feedback data and the target condition information respectively to obtain the treatment feedback data in coded form and the target condition information in coded form;

   Using a preset big data analysis model to perform data analysis on the coded treatment feedback data and the coded target condition information to generate a diagnosis conclusion label.
7. The disease analysis method according to any one of claims 1 to 4, wherein the step of generating a disease analysis report according to the diagnostic conclusion label includes:

   Purifying the diagnostic conclusion label according to the target condition information to obtain the purified diagnostic conclusion label;

   Perform verification analysis on the purified diagnostic conclusion label to obtain a standard diagnostic conclusion label;

   According to the standard diagnosis conclusion label, a condition analysis report is generated.
8. A condition analysis device, wherein the device includes:

   Electronic medical record data acquisition module, used to obtain electronic medical record data;

   A feature extraction module, configured to extract entity features from the electronic medical record data to obtain target disease information;

   A processing module, configured to use a pre-trained disease recognition model to process the target disease information, and generate a treatment reminder corresponding to the target disease information;

   The treatment feedback data receiving module is used to receive the treatment feedback data that the client responds to according to the treatment reminder;

   A diagnosis conclusion label generating module, configured to generate a diagnosis conclusion label according to the treatment feedback data and the target condition information;

   A condition analysis report generating module, configured to generate a condition analysis report according to the diagnosis conclusion label.
9. An electronic device, wherein the electronic device includes a memory, a processor, a program stored on the memory and operable on the processor, and a program for realizing the connection between the processor and the memory A data bus for communication, when the program is executed by the processor, it is implemented as a disease analysis method, wherein the disease analysis method includes:

   Access to electronic medical record data;

   Extracting entity features from the electronic medical record data to obtain target disease information;

   Processing the target condition information by using a pre-trained condition identification model to generate a treatment reminder corresponding to the target condition information;

   receiving the treatment feedback data that the user terminal responds to according to the treatment reminder;

   generating a diagnosis conclusion label according to the treatment feedback data and the target condition information;

   A condition analysis report is generated according to the diagnostic conclusion label.
10. The electronic device according to claim 9, wherein the step of extracting entity features from the electronic medical record data to obtain target disease information includes:

    Extracting natural language text in the electronic medical record data;

    Using a preset lexical analysis model to identify entity features in the natural language text;

    Segmenting the natural language text to obtain a natural language vocabulary sequence;

    Constructing a feature sequence table according to the entity feature and the natural language vocabulary sequence;

    Determine the target condition information according to the feature sequence table.
11. The electronic device according to claim 9, wherein the step of using a pre-trained disease identification model to process the target disease information to generate a treatment reminder corresponding to the target disease information includes:

    Inputting the target condition information into the pre-trained condition identification model;

    performing quadrant partitioning on the target disease information through the disease identification model to obtain partitioned disease information;

    The condition information of each quadrant is fitted and processed to generate medication reminders or medical reminders.
12. The electronic device according to claim 9, wherein, after the step of receiving the treatment feedback data that the user terminal responds to according to the treatment reminder, the method further comprises:

    Analyzing the drug data in the target condition information to obtain drug reaction information; wherein the drug reaction information includes adverse drug reaction information;

    Identifying abnormal medication data in the treatment feedback data according to the adverse drug reaction information.
13. The electronic device according to any one of claims 9 to 12, wherein, after the step of receiving the treatment feedback data that the user terminal responds to according to the treatment reminder, the method further comprises:

    The treatment feedback data and the target condition information are visualized to generate a visualized data map.
14. The electronic device according to any one of claims 9 to 12, wherein the step of generating a diagnostic conclusion label according to the treatment feedback data and the target condition information includes:

    Coding the treatment feedback data and the target condition information respectively to obtain the treatment feedback data in coded form and the target condition information in coded form;

    Using a preset big data analysis model to perform data analysis on the coded treatment feedback data and the coded target condition information to generate a diagnosis conclusion label.
15. A computer-readable storage medium for computer-readable storage, wherein the computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to Realize a kind of condition analysis method, wherein, described condition analysis method comprises:

    Access to electronic medical record data;

    Extracting entity features from the electronic medical record data to obtain target disease information;

    Processing the target condition information using a pre-trained condition recognition model to generate a treatment reminder corresponding to the target condition information;

    receiving the treatment feedback data that the user terminal responds to according to the treatment reminder;

    generating a diagnosis conclusion label according to the treatment feedback data and the target condition information;

    A condition analysis report is generated according to the diagnostic conclusion label.
16. The computer-readable storage medium according to claim 15, wherein the step of extracting entity features from the electronic medical record data to obtain target disease information includes:

    Extracting natural language text in the electronic medical record data;

    Using a preset lexical analysis model to identify entity features in the natural language text;

    Segmenting the natural language text to obtain a natural language vocabulary sequence;

    Constructing a feature sequence table according to the entity feature and the natural language vocabulary sequence;

    Determine the target condition information according to the feature sequence table.
17. The computer-readable storage medium according to claim 15, wherein the step of using a pre-trained disease recognition model to process the target disease information to generate a treatment reminder corresponding to the target disease information includes:

    Inputting the target condition information into the pre-trained condition identification model;

    performing quadrant partitioning on the target disease information through the disease identification model to obtain partitioned disease information;

    The condition information of each quadrant is fitted and processed to generate medication reminders or medical reminders.
18. The computer-readable storage medium according to claim 15, wherein, after the step of receiving the treatment feedback data from the user terminal according to the treatment reminder response, the method further comprises:

    Analyzing the drug data in the target condition information to obtain drug reaction information; wherein the drug reaction information includes adverse drug reaction information;

    Identifying abnormal medication data in the treatment feedback data according to the adverse drug reaction information.
19. The computer-readable storage medium according to any one of claims 15 to 18, wherein, after the step of receiving the treatment feedback data that the user responds to according to the treatment reminder, the method further comprises:

    The treatment feedback data and the target condition information are visualized to generate a visualized data map.
20. The computer-readable storage medium according to any one of claims 15 to 18, wherein the step of generating a diagnostic conclusion label according to the treatment feedback data and the target condition information includes:

    Coding the treatment feedback data and the target condition information respectively to obtain the treatment feedback data in coded form and the target condition information in coded form;

    Using a preset big data analysis model to perform data analysis on the coded treatment feedback data and the coded target condition information to generate a diagnosis conclusion label.

Images