Service package generation method and device based on patient data, equipment and storage medium

Abstract

The application relates to the technical field of big data, and discloses a service package generation method, device and equipment based on patient data and a storage medium. The method comprises the following steps: performing de-identification processing on collected original case data to obtain target treatment data of a patient; extracting a plurality of key events in the target treatment data, and performing fusion processing on the key events to obtain a medical information set of a disease corresponding to a disease type; inputting the target treatment data into a preset Bilstm model to obtain a medical feature vector of medical data, and performing pooling analysis on the medical feature vector to obtain a medical feedforward vector; and performing clustering analysis on the medical information set based on a preset cosine similarity algorithm according to the medical feedforward vector to obtain a medical service package corresponding to the disease. The application obtains different types of disease service packages with common characteristics by clustering the original case data, solves the technical problem of low medical service efficiency, and relieves medical pressure.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of big data, and particularly relates to a service package generation method and device based on patient data, equipment and a storage medium. BACKGROUND

[0002] The current medical situation is basically a special disease and a special treatment, that is, a user finds himself sick, goes to a hospital for treatment, a doctor makes a treatment plan of drug dispensing and reexamination for the patient after diagnosis, and the patient cooperates with the treatment according to the treatment plan. After the patient recovers, it is considered that the treatment is ended. This way has the following shortcomings:

[0003] 1. The traditional treatment mode does not follow up to let the patient participate in his own health management to understand the cause of the disease, to know the popular science of disease prevention related knowledge, and there is no prevention and treatment service of disease related complications.

[0004] 2. The traditional treatment mode is a passive treatment mode. There is no mechanism for early prevention. There is no improvement in the health link of the patient.

[0005] 3. The traditional treatment mode is basically a single disease and a single treatment of the doctor, without the comprehensive prevention and treatment of concurrent diseases and potential diseases.

[0006] At present, only a small number of people have their own family doctors, and most people do not have the ability of all-round management in health. How to let most people have early prevention when they are not sick, get all-round treatment when they are sick, and get all-round management in health is a common social problem. Therefore, how to provide convenient medical treatment for patients based on patient data and improve the treatment efficiency of patients has become a technical problem to be solved by the technical personnel in the field. SUMMARY

[0007] The main purpose of the present application is to solve the technical problem that the original case data cannot be analyzed in the prior art, and then the service package corresponding to different types of diseases is obtained, and the medical service efficiency is improved.

[0008] The first aspect of the present application provides a service package generation method based on patient data, comprising: collecting original case data of the same disease from a preset medical information platform, and extracting disease types and treatment data in the original case data; performing de-identification processing on the treatment data to obtain target treatment data; extracting a plurality of key events in the target treatment data, and performing fusion processing on the key events to obtain a medical information set of the disease corresponding to the disease type; inputting the medical information set into a preset Bilstm model for vector calculation to obtain a medical feature vector of the disease, and performing pooling analysis on the medical feature vector to obtain a medical feedforward vector; performing clustering analysis on the medical information set based on a preset cosine similarity algorithm according to the medical feedforward vector to obtain a medical service package corresponding to the disease.

[0009] Optionally, in the first implementation manner of the first aspect of the present application, after collecting the original case data of the same disease from the preset medical information platform and extracting the disease types and the treatment data in the original case data, the method further comprises: setting a diagnosis and treatment template of different disease types in advance; obtaining time information corresponding to the treatment data; obtaining a diagnosis and treatment template corresponding to the treatment data according to the disease type; and verifying the treatment data according to the diagnosis and treatment template to determine whether the time sequence of the treatment data is correct.

[0010] Optionally, in the second implementation manner of the first aspect of the present application, the de-identification processing on the treatment data to obtain the target treatment data comprises: constructing a treatment data query database according to the original case data; adding random noise to a sensitive attribute field in a data table of the treatment data query database according to a differential privacy algorithm; performing de-identification processing on the sensitive attribute field in the original case data according to the random noise to obtain an identifier field; and performing encryption processing on the identifier field to obtain the target treatment data.

[0011] Optionally, in the third implementation manner of the first aspect of the present application, the extracting a plurality of key events in the target treatment data comprises: obtaining a pre-defined key event set; performing screening processing on the target treatment data to filter invalid medical data in the target treatment data to obtain valid medical data; and extracting the plurality of key events in the target treatment data according to the key event set.

[0012] Optionally, in a fourth implementation form of the first aspect of the present application, the fusing processing of the key events to obtain the medical information set of the disease corresponding to the disease type comprises: determining attribute features corresponding to the key events and weight values of the attribute features; determining score data of the key events according to the attribute features and the weight values, and obtaining the medical information set of the disease corresponding to the disease type according to the score data.

[0013] Optionally, in a fifth implementation form of the first aspect of the present application, the pooling analysis of the medical feature vector to obtain the medical feedforward vector comprises: performing feature extraction on the medical feature vector to obtain target medical features of the medical feature vector; performing dimension reduction processing on the target medical features to obtain target data of a preset dimension; and performing pooling processing on the target data to obtain the medical feedforward vector of the target treatment data.

[0014] Optionally, in a sixth implementation form of the first aspect of the present application, the dimension reduction processing of the target medical features to obtain target data of a preset dimension comprises: obtaining frequencies of the target medical features; extracting first data corresponding to the target medical features in the target treatment data; performing data processing on different types of the first data to obtain standard data; and performing dimension reduction processing on the standard data to obtain target data of a preset dimension.

[0015] The second aspect of the present application provides a service package generation device based on patient data, comprising: an acquisition module configured to acquire original case data of the same disease from a preset medical information platform, and extract disease types and treatment data in the original case data; a de-identification module configured to perform de-identification processing on the treatment data to obtain target treatment data; a fusion module configured to extract a plurality of key events in the target treatment data, and perform fusing processing on the key events to obtain a medical information set of the disease corresponding to the disease type; a pooling module configured to input the medical information set into a preset Bilstm model for vector calculation to obtain a medical feature vector of the disease, and perform pooling analysis on the medical feature vector to obtain a medical feedforward vector; and a clustering module configured to perform clustering analysis on the medical information set based on a preset cosine similarity algorithm according to the medical feedforward vector to obtain a medical service package corresponding to the disease.

[0016] Optionally, in the first implementation of the second aspect of the present application, the service package generation device based on the patient data further comprises: a setting module for pre-setting diagnosis and treatment templates of different disease types; an acquisition module for acquiring time information corresponding to the treatment data; a diagnosis and treatment template corresponding to the treatment data is acquired according to the disease type; a verification module for verifying the treatment data according to the diagnosis and treatment template to determine whether the time sequence of the treatment data is correct.

[0017] Optionally, in the second implementation of the second aspect of the present application, the de-identification module is specifically configured to: construct a treatment data query database according to the original case data; add random noise to sensitive attribute fields in a data table of the treatment data query database according to a differential privacy algorithm; perform de-identification processing on the sensitive attribute fields in the original case data according to the random noise to obtain an identifier field; and perform encryption processing on the identifier field to obtain the target treatment data.

[0018] Optionally, in the third implementation of the second aspect of the present application, the fusion module is specifically configured to: acquire a pre-defined set of key events; perform filtering processing on the target treatment data to filter invalid medical data in the target treatment data to obtain valid medical data; and extract the valid medical data according to the set of key events to obtain a plurality of key events in the target treatment data.

[0019] Optionally, in the fourth implementation of the second aspect of the present application, the fusion module is specifically configured to: determine attribute features corresponding to the key events and weight values of the attribute features; determine score data of the key events according to the attribute features and the weight values, and obtain a medical information set of a disease corresponding to the disease type according to the score data.

[0020] Optionally, in the fifth implementation of the second aspect of the present application, the pooling module comprises: a feature extraction unit configured to perform feature extraction on the medical feature vector to obtain target medical features of the medical feature vector; a dimension reduction unit configured to perform dimension reduction processing on the target medical features to obtain target data of a preset dimension; and a pooling unit configured to perform pooling processing on the target data to obtain a medical feedforward vector of the target treatment data.

[0021] Optionally, in the sixth implementation of the second aspect of the present application, the dimension reduction unit is specifically configured to:

[0022] Obtaining the frequency of the target medical feature; extracting first data corresponding to the target medical feature in the target treatment data; performing data processing on different types of the first data to obtain standard data; performing dimension reduction processing on the standard data to obtain target data of a preset dimension.

[0023] The third aspect of the present application provides a service package generation device based on patient data, comprising: a memory and at least one processor, the memory having instructions stored therein, and the memory and the at least one processor being interconnected by a circuit;

[0024] The at least one processor invokes the instructions in the memory to enable the service package generation device based on patient data to perform the steps of the service package generation method based on patient data described above.

[0025] The fourth aspect of the present application provides a computer-readable storage medium having instructions stored therein, which, when executed on a computer, cause the computer to perform the steps of the service package generation method based on patient data described above.

[0026] In the technical solution provided by the present application, raw case data of the same disease is collected from a preset medical information platform, and the disease type and treatment data in the raw case data are extracted; the treatment data is de-identified to obtain target treatment data; a plurality of key events in the target treatment data are extracted and fused to obtain a medical information set of the disease corresponding to the disease type; the target treatment data is input into a preset Bilstm model to obtain a medical feature vector of the medical data, and the medical feature vector is subjected to pooling analysis to obtain a medical feedforward vector; based on a preset cosine similarity algorithm, the medical information set is subjected to clustering analysis according to the medical feedforward vector to obtain a medical service package corresponding to the disease. The present application obtains target treatment data with common characteristics by clustering analysis of raw case data, and generates service packages corresponding to different types of diseases according to the target treatment data, thereby solving the problems of lack of sustainable tracking service in current medical service, difficulty in timely and effective feedback, lack of effective integration of medical data information, and inability to guarantee the quality of mobile medical service. The efficiency and portability of medical service are effectively improved, the waiting time of medical personnel is reduced, and the medical pressure is relieved. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 The first embodiment schematic diagram of the service package generation method based on patient data provided by the present application;

[0028] Figure 2 The second embodiment schematic diagram of the service package generation method based on patient data provided by the present application;

[0029] Figure 3 The third embodiment of the service package generation method based on patient data provided by the present application is shown in the figure;

[0030] Figure 4 The fourth embodiment of the service package generation method based on patient data provided by the present application is shown in the figure;

[0031] Figure 5 The fifth embodiment of the service package generation method based on patient data provided by the present application is shown in the figure;

[0032] Figure 6 The first embodiment of the service package generation device based on patient data provided by the present application is shown in the figure;

[0033] Figure 7 The second embodiment of the service package generation device based on patient data provided by the present application is shown in the figure;

[0034] Figure 8 An embodiment of the service package generation device based on patient data provided by the present application is shown in the figure. DETAILED DESCRIPTION

[0035] The service package generation method, device, equipment and storage medium based on patient data provided by the embodiment of the present application first collect the original case data of the same disease from the preset medical information platform, and extract the disease type and treatment data in the original case data; The treatment data is de-identified to obtain target treatment data; Extract a plurality of key events in the target treatment data, and fuse the key events to obtain a medical information set corresponding to the disease type; The target treatment data is input into a preset Bilstm model to obtain a medical feature vector of the medical data, and the medical feature vector is analyzed by pooling to obtain a medical feedforward vector; According to the medical feedforward vector, the medical information set is analyzed based on a preset cosine similarity algorithm to obtain a medical service package corresponding to the disease. The present application obtains target treatment data with common characteristics by clustering analysis of original case data, and generates different types of disease corresponding service package according to target treatment data, solves the problem that current medical service lacks sustainable tracking service, and it is difficult to achieve timely and effective feedback, and lacks effective integration of medical data information, resulting in that the quality of mobile medical service cannot be guaranteed. Effectively improve the efficiency and portability of medical service, reduce the waiting time of medical personnel, and relieve the medical pressure.

[0036] The terms "first", "second", "third", "fourth" and the like in the description and in the claims of the present application, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the use of these terms herein is to be construed as interchangeable in order to comfort the embodiments described herein to be practiced in either order unless it is clear from the context that one technique precedes another technique. Furthermore, the terms "comprising", "having", "including", and the like, as can be used herein, are to be interpreted inclusively rather than exclusively. That is, unless otherwise indicated, these terms are to be construed to cover a process, method, system, product, or apparatus that comprises or includes a series of steps or units, but is not limited to only those steps or units that are explicitly listed.

[0037] For the purpose of facilitating understanding, the specific flow of the embodiments of the present application is described below, please refer to Figure 1 The first embodiment of the service package generation method based on patient data in the embodiments of the present application comprises:

[0038] 101. Collecting original case data of the same disease from a preset medical information platform, and extracting disease type and treatment data in the original case data;

[0039] In this embodiment, patients go to the Internet hospital built by Ping'an Health for medical treatment, which will form a large amount of data, including patient big data, diagnosis and treatment big data, disease big data, etc.

[0040] When designing a business system, the product will add data characteristic items required by the business on the interface:

[0041] For example, when a patient asks for a diagnosis, he or she needs to fill in the diagnosis card and medical record information first, then ask for a diagnosis, and the doctor will prescribe medicine or nursing according to the patient's condition. Each item will fill in the relevant information on the page. These information is the characteristic information that can be summarized for physical examination:

[0042] a. Diagnosis card information: patient's name, gender, age, birthday, household address, residence address, certificate type, certificate number, mobile phone number, social security card number, nationality, nationality, guardian information (name, phone, ID card, etc.), marital status, etc.

[0043] b. Patient condition information: affected part, related living habits, disease description, disease regularity, allergy history, medication history, disease duration, disease level, etc.

[0044] c. Diagnosis information: disease name, suspected disease name, disease severity, diagnosis suggestion, etc.

[0045] d. Doctor's prescription information: department, diagnosis result, prescription category, drug name, drug specification, drug brand, drug usage

[0046] e. Care information: disease name, disease information, care items (moxibustion, massage, physiotherapy, etc.), care frequency, etc.

[0047] 102. De-identify the treatment data to obtain target treatment data;

[0048] In this embodiment, de-identification refers to a data processing method that processes identifiers so that the processed information cannot identify a specific personal information subject. The biggest difference between the Personal Information Security Specification and the Personal Information De-identification Guidelines in China and the relevant laws in the United States and Canada regarding de-identification is whether to consider the possibility of indirect identification to prevent re-identification. China limits the "without additional information" when re-identifying, which denies the "indirect identification" situation, which is very similar to pseudonymization in GDPR; while CCPA and HIPAA laws require higher requirements for preventing re-identification, considering the possibility of re-identification combined with other additional information.

[0049] Specifically, de-identification refers to the process of removing the association between a set of identifiable data and data subjects. Through this process, data managers can delete or change the identification information in the data set, so that attackers can hardly or cannot identify the specific personal subject identity using the data set, so that the data set can be shared within the predetermined range for use. De-identification is one of the main tools of privacy preserving data publishing (PPDP) [1], by removing the association between the privacy attributes and the data subjects in the data set, and having sufficient ability to prevent re-identification, some attributes of the data set can be shared and published for processing and analysis by external business systems.

[0050] 103. Extract a plurality of key events in the target treatment data, and fuse the key events to obtain a medical information set of the disease corresponding to the disease type;

[0051] In this embodiment, the key event can refer to a core event set according to different diseases of the target object in the treatment data, which can represent the diagnosis and treatment record or the diagnosis and treatment scheme corresponding to the disease, for example, assuming that the disease of the target object in the treatment data can be breast cancer disease, since the treatment strategies corresponding to breast cancer diseases of different clinical stages are completely different, the treatment choices of breast cancer diseases of the same clinical stage according to different molecular typing and pathological diagnosis are also different, for example, for HER2 (human epidermal growth factor receptor) positive breast cancer patients, the timing of targeted drug trastuzumab and the judgment of treatment efficacy are particularly important, therefore, it is necessary to set the key event of breast cancer disease, so as to facilitate scientific research, help to review and analyze the survival analysis of different treatment choices of breast cancer disease patients, and further optimize the treatment strategy of breast cancer disease patients; specifically, combining the characteristics of breast cancer disease, the key event in the treatment data can be the first diagnosis and treatment event, can also be a local treatment related event, of course, can also be a drug treatment related event, an efficacy evaluation event, a drug adverse reaction event, etc., and the present example embodiment is not limited thereto.

[0052] In this embodiment, the data fusion technology refers to using a computer to automatically analyze and synthesize a plurality of observation information obtained in time sequence under certain criteria to complete the required decision and evaluation tasks. The data fusion technology includes collection, transmission, synthesis, filtering, correlation and synthesis of useful information given by various information sources, so as to assist people in situation / environment judgment, planning, detection, verification and diagnosis.

[0053] 104. The medical information set is input into the preset Bilstm model for vector calculation to obtain a medical feature vector of the disease, and the medical feature vector is subjected to pooling analysis to obtain a medical feedforward vector;

[0054] In this embodiment, the target treatment data is converted from a text form to a vector form, which is used for finally calculating the similarity between vectors, and determining the medical word most matching the analysis statement from a plurality of medical words according to the similarity; the Bilstm model is a natural language processing neural network model, and the Bilstm model performs vector conversion on the target treatment data. For example, the analysis statement is "eye swelling, eye pain, photophobia, eyeball hard, weak vision"; the first medical word is "glaucoma, acute angle-closure glaucoma, chronic angle-closure glaucoma, primary open-angle glaucoma, filtering bleb separation", the second medical word is "myopia", and the third medical word is "keratitis".

[0055] In this embodiment, the Bilstm model splices the target treatment data to obtain an original statement vector HL, that is, the original statement vector HL{eye swelling + eye pain + photophobia + eyeball hard + weak vision} and the like.

[0056] The pooling analysis is a pooling based on a convolutional neural network. In this embodiment, max pooling and average pooling are used to process the target treatment data. The max pooling and average pooling are combined to perform a parallel double pooling operation, thereby retaining the deeper semantic information of the target treatment data.

[0057] 105. Based on the preset cosine similarity algorithm, the medical information set is clustered according to the medical feedforward vector, and the service corresponding to the disease is obtained.

[0058] In this embodiment, the cosine similarity, also known as the cosine similarity, is evaluated by calculating the cosine value of the included angle of two vectors. The cosine similarity draws the vector according to the coordinate value into the vector space, such as the most common two-dimensional space.

[0059] The cosine similarity measures the similarity between two vectors by measuring the cosine value of the included angle of the two vectors. The cosine value of the 0-degree angle is 1, while the cosine value of any other angle is not greater than 1; and the minimum value is -1. Thus, the cosine value of the angle between two vectors determines whether the two vectors point in the same direction. When two vectors have the same direction, the cosine similarity value is 1; when the included angle between two vectors is 90°, the cosine similarity value is 0; and when two vectors point in completely opposite directions, the cosine similarity value is -1. The result is independent of the length of the vector and only related to the pointing direction of the vector. Cosine similarity is usually used in positive space, so the value given is between -1 and 1.

[0060] Note that this upper and lower bound applies to any dimensional vector space, and cosine similarity is most commonly used in high-dimensional positive space. For example, in information retrieval, each term is assigned a different dimension, and a dimension is represented by a vector, the value of each dimension corresponding to the frequency of the term appearing in the document. Cosine similarity can thus give the similarity of two documents in terms of their topics. In addition, it is commonly used for file comparison in text mining. In addition, in the field of data mining, it is used to measure the cohesion within the cluster.

[0061] Among them, the cluster analysis refers to the analysis process of grouping a set of physical or abstract objects into multiple classes composed of similar objects. The goal is to collect data for classification on the basis of similarity. Clustering originates from many fields, including mathematics, computer science, statistics, biology and economics. In different application fields, many clustering techniques have been developed, which are used to describe data, measure the similarity between different data sources, and classify data sources into different clusters.

[0062] In this embodiment, the service package is also called a card commodity. Specifically, according to the obtained target treatment data, basic services are first created, and the services are combined into a service package. One or more services are included in a service package, such as disease management, summer moxibustion, chronic obstructive pulmonary disease, electrocardiogram monitoring, pediatric ophthalmic examination, quick consultation, and inpatient worry-free service.

[0063] The services in one service package are often a combination of one or a group of services with greater relevance. For example, the summer moxibustion service package includes five times of summer moxibustion service, which is specially designed for the elderly to prevent winter diseases in summer. For example, the inpatient secretary service package of the cardiology department is specially designed for patients with cardiology diseases, and includes three services of cardiology offline consultation green pass, cardiology offline examination green pass, and cardiology inpatient green pass.

[0064] Service package classification: Each service package includes one or more classification tags, and the classification tags are selected from various common phenomena, such as children, adolescents, young people, middle-aged people, and the elderly according to age. According to the department, it is divided into ophthalmology, cardiology, thoracic surgery, etc. According to the applicable population, it is divided into students, office workers, pregnant women, and three high-risk groups.

[0065] In the embodiment of the application, the original case data of the same disease is collected from the preset medical information platform, and the disease type and treatment data in the original case data are extracted; the treatment data is de-identified to obtain target treatment data; a plurality of key events in the target treatment data are extracted, and the key events are fused to obtain a medical information set corresponding to the disease type; the target treatment data is input into a preset Bilstm model to obtain a medical feature vector of the medical data, and the medical feature vector is analyzed by pooling to obtain a medical feedforward vector; according to the medical feedforward vector, the medical information set is clustered based on a preset cosine similarity algorithm to obtain a medical service package corresponding to the disease. The application obtains target treatment data with common characteristics by clustering analysis of original case data, and generates different types of disease corresponding service packages according to target treatment data, solves the problem that current medical services lack sustainable tracking services and are difficult to feedback in time, and effectively integrates medical data information, which solves the problem that the quality of mobile medical services cannot be guaranteed. The efficiency and portability of medical services are effectively improved, the waiting time of medical personnel is reduced, and the medical pressure is relieved.

[0066] Please refer to Figure 2 The second embodiment of the service package generation method based on patient data in the embodiment of the application includes:

[0067] 201, collect original case data of the same disease from a preset medical information platform, and extract disease type and treatment data from the original case data;

[0068] 202、previously set diagnosis and treatment templates of different disease types;

[0069] In this embodiment, before collecting original case data of the same disease from the preset medical information platform and extracting disease types and treatment data in the original case data, different diagnosis and treatment templates are configured for different types of patients. Different types of patients refer to patients with different reasons for visiting. When patients have different reasons for visiting (the reasons for visiting specifically include chest pain, stroke, and trauma), there are corresponding diagnosis and treatment processes for examination, diagnosis, and treatment, so diagnosis and treatment templates can be constructed according to the corresponding diagnosis and treatment processes to facilitate time sequence verification of subsequent information entry.

[0070] 203、obtaining time information corresponding to the treatment data;

[0071] In this embodiment, an information entry template is configured to collect various treatment data of a patient entered by a user.

[0072] Before obtaining the treatment data, the information entry template is configured, and the user enters information according to the prompt of the pre-configured information entry template to collect patient information, reduce the error and omission of information entry, and reduce the frequency of subsequent modification. Specifically, the information entry template includes a sub-template for each diagnosis and treatment item of each type of patient.

[0073] Preferably, the time information of each treatment data is obtained, specifically: the entry time of each treatment data of the patient entered by the user is obtained as the corresponding time information. Generally, when a patient performs each diagnosis and treatment item, the user will simultaneously enter each treatment data, so the entry time of the treatment data can be used as the time information of the corresponding treatment data, automatic entry of the time information is realized, and the workload is reduced. At the same time, after automatic entry of the time information, the user can be prompted to check the automatically entered time information, and if there is an error, the automatically entered time information can be modified.

[0074] 204、obtaining a diagnosis and treatment template corresponding to the treatment data according to the disease type;

[0075] In this embodiment, the diagnosis and treatment template is generated in combination with the disease type of the patient and the sequence requirements thereof. Specifically, the diagnosis and treatment template corresponding to the STEMI type of patient is set, which includes ten diagnosis and treatment items of onset, in-hospital first electrocardiogram acquisition, in-hospital first electrocardiogram diagnosis, start of informed consent, start of catheter room, activation of catheter room, arrival of patient at catheter room, decision of interventional surgery, signing of informed consent, and guide wire passing, and three sequence requirements that the start of informed consent time should be later than the start of catheter room time, the start of catheter room time should be later than the decision of interventional surgery time, and the signing of informed consent time should be later than the start of informed consent time. The treatment data of the corresponding type of patient is verified according to the diagnosis and treatment template.

[0076] 205. verifying the treatment data according to the diagnosis and treatment template to determine whether the time sequence of the treatment data is correct;

[0077] In this embodiment, the diagnosis and treatment template further includes time length indicators of each diagnosis and treatment item. The actual time length of each diagnosis and treatment item is calculated according to the time information of each treatment data of the current patient. It is determined whether the actual time length meets the corresponding time length indicator. If yes, the treatment data is reported, otherwise, a warning prompt is given.

[0078] In addition to verifying the time sequence requirements of each diagnosis and treatment item, the time length requirements of each diagnosis and treatment item are also verified in this embodiment. When setting the diagnosis and treatment template of the STEMI type of patient, the time length indicators of each diagnosis and treatment item are set, such as greater than, less than or within a set time range. Then, the actual time length of each diagnosis and treatment item is calculated according to the time information of the input treatment data, and it is determined whether the actual time length meets the time length indicator.

[0079] 206. performing de-identification processing on the treatment data to obtain target treatment data;

[0080] 207. extracting a plurality of key events in the target treatment data, and performing fusion processing on the key events to obtain a medical information set of a disease corresponding to the disease type;

[0081] 208. inputting the medical information set into a preset Bilstm model to perform vector calculation to obtain a medical feature vector of the disease, and performing pooling analysis on the medical feature vector to obtain a medical feedforward vector;

[0082] 209. performing clustering analysis on the medical information set based on the preset cosine similarity algorithm according to the medical feedforward vector to obtain a medical service package corresponding to the disease.

[0083] Steps 201, 206-209 in this embodiment are similar to steps 101, 102-105 in the first embodiment, which will not be described here again.

[0084] In the embodiment of the present application, the original case data of the same disease is collected from the preset medical information platform, and the disease type and treatment data in the original case data are extracted; the treatment data is de-identified to obtain target treatment data; a plurality of key events in the target treatment data are extracted and fused to obtain a medical information set of the disease corresponding to the disease type; the target treatment data is input into a preset Bilstm model to obtain a medical feature vector of the medical data, and the medical feature vector is pooled and analyzed to obtain a medical feedforward vector; and the medical information set is clustered based on the preset cosine similarity algorithm according to the medical feedforward vector to obtain a medical service package corresponding to the disease. The present application obtains target treatment data with common characteristics through clustering analysis of original case data, and generates service packages corresponding to different types of diseases according to the target treatment data, thereby solving the problems of lack of sustainable tracking service in current medical service, difficulty in timely and effective feedback, lack of effective integration of medical data information, and inability to guarantee the quality of mobile medical service. The efficiency and portability of medical service are effectively improved, the waiting time of medical personnel is reduced, and the medical pressure is relieved.

[0085] Please refer to Figure 3 The third embodiment of the service package generation method based on patient data in the embodiment of the present application includes:

[0086] 301. Collecting original case data of the same disease from a preset medical information platform, and extracting disease types and treatment data in the original case data;

[0087] 302. Constructing a treatment data query database according to the original case data;

[0088] In this embodiment, the treatment data query database is constructed according to the disease influence characteristics and clinical treatment data, combined with the treatment data of the HIS, RIS and PACS databases of the patient, and the patient privacy data is de-identified by combining the differential privacy algorithm with the encryption algorithm, so as to update the treatment data query database, so that the clinical image query diagnosis and treatment system can be established according to the updated treatment data query database. The embodiment of the present application not only meets the privacy requirement in the differential privacy protection model, but also guarantees the reliability of the published data in the database, so as to help clinical researchers to query and collect past cases, analyze and evaluate big data, promote the automation of medical data statistics, eliminate information islands, and provide decision support to lay a good foundation.

[0089] 303. Adding random noise to the sensitive attribute field in the data table of the treatment data query database according to the differential privacy algorithm;

[0090] In this embodiment, the differential privacy algorithm adds appropriate noise to the statistical results to ensure that modifying an individual record in the data set does not significantly affect the statistical results, thereby meeting the requirements of privacy protection.

[0091] Assuming that D1 and D2 are adjacent data sets, S is all possible outputs of the random function A, and Pr is the probability that A(D1) obtains a certain value, then the algorithm is said to meet the standard of ε-differential privacy as long as the algorithm satisfies the following formula.

[0092] Pr[A(D1)∈S]≤e∈×Pr[A(D2)∈S]

[0093] where the probability Pr[·] represents the risk of privacy leakage, which is controlled by the randomness of the algorithm A(D); ε is a privacy protection budget parameter, which is used to adjust the balance between data privacy security and data reliability, and privacy protection is achieved by adding random noise, that is, the smaller the ε, the more noise is added, and the higher the privacy protection, and vice versa.

[0094] Optionally, different noise mechanisms are used to add random noise to the original treatment data of the sensitive fields of different data types. For example, Laplace mechanism is used to add random noise to the original treatment data of the sensitive fields of numerical type, and exponential mechanism is used to add random noise to the original treatment data of the sensitive fields of non-numerical type.

[0095] The Laplace mechanism processes numerical type data (continuous data), such as patient age, and adds random noise to the numerical result to achieve differential privacy. The exponential mechanism processes non-numerical type (discrete data) data, and returns a result with a certain probability value, and the output is a set of discrete data, which can be determined by a scoring function. The higher the score, the higher the output probability, and the lower the score, the lower the output probability.

[0096] Specifically, Laplace noise is added to the numerical type sensitive attributes such as patient age and examination date in the table, and exponential noise is added to the attributes such as gender, education level, region, examination device type, and disease in the data table to obtain noise results, which are replaced into the data table.

[0097] 304、According to the random noise, the sensitive attribute fields in the original case data are de-identified to obtain an identifier field;

[0098] In this embodiment, the original database structure is roughly divided into the following categories: display identifier: a set of attributes that can uniquely identify a single individual, such as patient name and patient number fields in the table; sensitive attributes: a set of attributes containing private data: patient gender, age, education, region, examination device type, disease, etc.; non-sensitive attributes: a set of attributes other than the above categories.

[0099] 305, encrypt the identifier field to obtain the target treatment data of the patient;

[0100] In this embodiment, the identifier field of the data table is encrypted according to the data encryption standard DES encryption algorithm and Base64 encoding; DES symmetric encryption is a relatively traditional encryption method with very high security. Base64 is a method of representing binary data based on 64 printable characters. Combining these two methods meets the encryption needs of sensitive attributes such as patient name, examination number, and image number, which uniquely identify.

[0101] 306, extract a plurality of key events in the target treatment data, and fuse the key events to obtain a medical information set corresponding to the disease type of the disease;

[0102] 307, input the medical information set into a preset Bilstm model for vector calculation to obtain a medical feature vector of the disease, and perform pooling analysis on the medical feature vector to obtain a medical feedforward vector;

[0103] 308, based on the medical feedforward vector, perform clustering analysis on the medical information set based on a preset cosine similarity algorithm to obtain a medical service package corresponding to the disease.

[0104] In this embodiment, steps 301, 306-308 are similar to steps 101, 103-105 in the first embodiment, and will not be described here.

[0105] In the embodiment of the present application, the original case data of the same disease is collected from the preset medical information platform, and the disease type and treatment data in the original case data are extracted; the treatment data is de-identified to obtain target treatment data; a plurality of key events in the target treatment data are extracted, and the key events are fused to obtain a medical information set of the disease corresponding to the disease type; the target treatment data is input into a preset Bilstm model to obtain a medical feature vector of the medical data, and the medical feature vector is pooled and analyzed to obtain a medical feedforward vector; and the medical information set is clustered based on the preset cosine similarity algorithm according to the medical feedforward vector to obtain a medical service package corresponding to the disease. The present application obtains target treatment data with common characteristics through clustering analysis of original case data, and generates service packages corresponding to different types of diseases according to the target treatment data, thereby solving the problems of lack of sustainable tracking service in current medical service, difficulty in timely and effective feedback, lack of effective integration of medical data information, and inability to guarantee the quality of mobile medical service. The efficiency and portability of medical service are effectively improved, the waiting time of medical personnel is reduced, and the medical pressure is relieved.

[0106] Please refer to Figure 4 The fourth embodiment of the service package generation method based on patient data in the embodiment of the present application includes:

[0107] 401. Collecting original case data of the same disease from a preset medical information platform, and extracting disease types and treatment data in the original case data;

[0108] 402. De-identifying the treatment data to obtain target treatment data;

[0109] 403. Obtaining a predefined key event set;

[0110] In this embodiment, the predefined key event set is obtained; and a plurality of key events corresponding to the target object are extracted from the treatment data according to the key event set. The predefined key event set can be a set of key events corresponding to the disease set by the staff according to the special characteristics of various diseases at different clinical stages, that is, the predefined key event set can include key events of different clinical stages corresponding to the target disease. The plurality of key events corresponding to the target object are extracted from the treatment data according to the predefined key event set. The predefined key event set can effectively ensure the accuracy of the extracted plurality of key events, and also ensure the extraction efficiency of the extracted key events.

[0111] Preferably, the predefined set of key events can be periodically updated, or updated in real time when new key events occur, which can be updated manually or by a crawler tool or artificial intelligence, and the present example embodiment does not make special limitations.

[0112] 404, screening the target treatment data to filter out invalid medical data in the target treatment data, to obtain valid medical data;

[0113] In the present embodiment, the treatment data is screened to filter out invalid medical data in the treatment data; and a plurality of key events corresponding to the target object are extracted from the filtered treatment data according to the set of key events. Invalid medical data can refer to treatment data in the treatment data that is meaningless for the diagnosis and treatment of the target disease of the target object, for example, invalid medical data can be treatment data in the treatment data that treats the target object for a non-target disease, or can refer to incomplete treatment data in the treatment data that treats the target disease (for example, treatment data that ends treatment halfway due to special reasons), and of course, it can also be other treatment data in the treatment data that is meaningless for the diagnosis and treatment of the target disease of the target object, and the present example embodiment does not make special limitations. By filtering out invalid medical data in the treatment data, the accuracy of the treatment data can be effectively improved, the extraction accuracy of the key events is further ensured, unnecessary calculations are reduced, and computing resources are saved.

[0114] 405, extracting the valid medical data according to the set of key events to obtain a plurality of key events in the target treatment data;

[0115] In one example embodiment, the second key event can refer to an event determined by extracting key events from a plurality of treatment data as a whole, for example, the second key event can be the first diagnosis and treatment event in all treatment data corresponding to the series of the target object, or can be a drug adverse reaction event in all treatment data, and of course, the second key event can also be an event determined by jointly judging the treatment data across multiple medical records, for example, for the first chemotherapy event after recurrence, first determine the first medical visit for the first recurrence, then extract the chemotherapy drug in all medical drug orders after this medical visit, and finally find the first medical visit where the chemotherapy drug appears, which can determine the first chemotherapy event after recurrence. Of course, this is only an illustrative example and should not limit the present example embodiment in any special way.

[0116] In another example embodiment, the first key event extracted from the sorted single treatment data and the second key event extracted from the sorted multiple treatment data jointly constitute the key event corresponding to the treatment data of the target object in chronological order. By constructing the key event from the first key event extracted from the sorted single treatment data and the second key event extracted from the sorted multiple treatment data, the problem of missing key events or inaccurate key events due to insufficient screening or due to the presence of events that require joint determination of treatment data across multiple clinic records can be avoided, and the accuracy of the key event can be improved.

[0117] 406, determine the attribute feature corresponding to the key event and the weight value of the attribute feature;

[0118] In this embodiment, the attribute feature corresponding to the key event and the weight value of the attribute feature are determined. The weight value of the attribute feature is determined. The score data of the plurality of target key events is determined according to the attribute feature and the weight value, and the plurality of target key events are de-duplicated according to the score data to filter target key events with the same meaning. The weight value can refer to the pre-set weight of the attribute feature, for example, the weight of the time attribute can be 0.3, and the weight of the recurrence type can be 0.7. Of course, the weight value of the attribute feature can be customized according to actual conditions, and this example embodiment does not specially limit it. The score data can refer to the similarity data used to determine whether two key events are similar key events. The higher the score data of two key events, the more similar the two key events are, and only one of them is retained.

[0119] 407, determine the score data of the key event according to the attribute feature and the weight value, and obtain the medical information set of the disease corresponding to the disease type according to the score data;

[0120] In this embodiment, the fusion processing can refer to the process of constructing the target data from the plurality of key events in a specific order, for example, the fusion processing can include but is not limited to arranging the plurality of key events in chronological order, and this example embodiment does not specially limit it. The target data can refer to the structured data of the treatment data of the target object generated by the fusion processing of the plurality of key events.

[0121] Before the fusion processing of the plurality of key events, the attribute feature corresponding to the plurality of key events is first determined. The attribute feature can refer to different attributes corresponding to the key event, for example, the attribute feature can be the time attribute corresponding to the key event, or the recurrence type corresponding to the key event, and of course, the attribute feature can also be other attributes corresponding to the key event, and this example embodiment does not specially limit it.

[0122] 408. Input the medical information set into the preset Bilstm model for vector calculation to obtain the medical feature vector of the disease, and perform pooling analysis on the medical feature vector to obtain the medical feedforward vector;

[0123] 409. Based on the medical feedforward vector, cluster analysis is performed on the medical information set using a preset cosine similarity algorithm to obtain medical service packages corresponding to diseases.

[0124] In this embodiment, steps 401-402 and 408-409 are similar to steps 101-102 and 104-105 in the first embodiment, and will not be described again here.

[0125] In this embodiment of the invention, original case data of similar diseases are collected from a preset medical information platform, and disease type and treatment data are extracted from the original case data. The treatment data is de-identified to obtain target treatment data. Multiple key events are extracted from the target treatment data and fused to obtain a set of medical information corresponding to the disease type. The target treatment data is input into a preset Bilstm model to obtain medical feature vectors, and pooling analysis is performed on these medical feature vectors to obtain medical feedforward vectors. Based on the medical feedforward vectors, a preset cosine similarity algorithm is used to perform cluster analysis on the medical information set to obtain a medical service package corresponding to the disease. This invention obtains target treatment data with common characteristics by clustering analysis of original case data, and generates service packages corresponding to different types of diseases based on the target treatment data. This solves the current problems of lacking sustainable tracking services in medical services, difficulty in achieving timely and effective feedback, and lack of effective integration of medical data information, leading to unreliable quality of mobile medical services. It effectively improves the efficiency and portability of medical services, reduces waiting time for patients, and alleviates medical pressure.

[0126] Please see Figure 5 The fifth embodiment of the service package generation method based on patient data in this invention includes:

[0127] 501. Collect original case data of similar diseases from the preset medical information platform, and extract the disease type and treatment data from the original case data;

[0128] 502. De-identify the treatment data to obtain the target treatment data;

[0129] 503. Extract multiple key events from the target treatment data, and fuse these key events to obtain a set of medical information corresponding to the disease type;

[0130] 504. Extract features from the medical feature vector to obtain the target medical features of the medical feature vector;

[0131] In this embodiment, the medical feature vector is input to two different parallel pooling layers for dimension reduction after being input to convolution layers with different window sizes and the feature extraction of the filter units thereunder, i.e., an avg pooling layer and a max pooling layer, which fully combines the characteristics dynamically extracted by the max pooling layer and the contribution ability of the avg pooling layer to the average semantics of short texts, effectively reducing the loss of semantic information in dimension reduction; finally, necessary semantic splicing is performed in a concatenation layer to form the original feedforward vector and the medical feedforward vector; wherein the max pooling layer dynamic extraction method and the avg pooling layer average semantic method for short texts consider the influence of the height of the convolution kernel sliding window on the generated feature map. That is, the height of the convolution kernel is an important basis for the number M of down-sampling of the feature map, the higher the convolution kernel, the fewer the down-sampling number, and vice versa, the lower the convolution kernel height, the more the down-sampling number.

[0132] 505、acquiring the frequency of the target medical feature;

[0133] In this embodiment, the frequency of the target medical feature can be the frequency of the target medical feature appearing in different historical analysis models. For different target medical features, the frequency of appearing in multiple historical analysis models can be different. For example, for the surgical fee, it can appear in various different historical analysis models, and the frequency is high, while for the glucose content, it can only appear in the diabetes surgery analysis model, and the frequency is low.

[0134] Specifically, the higher the frequency of the target medical feature appearing in different historical analysis models, such as the target medical feature appearing in all historical analysis models, the more important the target medical feature can be determined to be. Thus, the determination of the target medical feature can be made according to the frequency of the feature appearing in different historical analysis models.

[0135] The frequency meeting the requirement can be that the frequency of the target medical feature appearing in the historical analysis model meets a certain threshold condition, or the frequency of the target medical feature appearing in the historical analysis model is sorted, and the target medical feature whose sorting meets a certain requirement is the target medical feature with the frequency meeting the requirement.

[0136] 506、extracting the first data corresponding to the target medical feature in the target treatment data;

[0137] In this embodiment, the multi-dimensional data can refer to all the data stored in the database, and can include newly added data and historical data before the data is changed each time. For example, for the medical insurance data, the initial data refers to the medical data generated after the user visits the doctor and stored under the user name, and can include historical medical data and current medical data, and can specifically include but is not limited to the data of the inquiry location, the inquiry time, the International Classification of Diseases (ICD), the registration department, the registration doctor information, the registration fee, the payment method, the examination item, the examination fee, the disease description, the medical advice, the drug list, the drug price, the drug dosage, the payment window, the drug pickup window, whether to return for a medical examination, the return time, the number of inquiries, and the like.

[0138] Specifically, the server can extract the initial data from the multi-dimensional data based on the selected target medical feature. The extraction of the initial data can be divided into multiple categories. For example, for medical insurance data, the initial data can include but is not limited to current medical expense data, current medical ICD data, and historical medical data. The current medical expense data can include but is not limited to surgical fees, drug fees, and examination fees. The current medical ICD data can include but is not limited to the cost of the current diagnosed ICD, the average cost of the ICD, and the like. The historical medical data can include but is not limited to the number of local outpatient visits, the number of local hospitalizations, the number of out-of-town outpatient visits, the number of out-of-town hospitalizations, the proportion of local outpatient visits, the proportion of out-of-town outpatient visits, and the like.

[0139] 507. Perform data processing on the different types of first data to obtain standard data;

[0140] In this embodiment, the extracted initial data can have a large difference in data magnitude due to different data types. For example, the drug fee is 500, and the total fee is 1,000,000, which have a huge difference in data magnitude.

[0141] The server can perform data processing on the initial data of different data magnitudes by using the same data magnitude data processing method to obtain standard data of the same data magnitude. For example, in the previous example, the same data magnitude data processing is performed on the drug fee and the total fee to obtain the drug fee and the total fee with a data magnitude between 0 and 100, i.e., the standard drug fee is 0.05 and the standard total fee is 100.

[0142] Specifically, the same data magnitude data processing method can be selected according to the difference in data type or the difference in data magnitude, for example, the square root, square, cube, exponential, logarithm, and the like can be selected, and the present application does not limit the same.

[0143] 508. Perform dimensionality reduction processing on the standard data to obtain target data of a preset dimension.

[0144] In this embodiment, the preset dimension can be a dimension pre-set by the terminal to the server according to the needs of subsequent data processing by the user, and the data quantity of the target data of the preset dimension can be less than that of the standard data. The method of nonlinear dimensionality reduction processing can include but is not limited to Isometric Feature Mapping (Isomap), Locally Linear Embedding (LLE), Modified Locally Linear Embedding (MLLE), Hessian Eigenmapping, Spectral Embedding, Local Tangent Space Alignment (LTSA), Multi-dimensional Scaling (MDS), t-distributed Stochastic Neighbor Embedding (t-SNE), etc.

[0145] In actual application, a linear dimensionality reduction processing method can also be used, which can include but is not limited to Principal Component Analysis (PCA), kernel PCA, Incremental PCA, etc. Specifically, the server can use the clustering characteristics of data in the Riemann space in multiple dimensions to map the standard data in multiple dimensions to low dimensions, for example, to 2 dimensions, to obtain target data according to the above method.

[0146] In the above data dimensionality reduction processing method, the target medical feature is obtained by the historical analysis model, then the initial data corresponding to the target medical feature in the multi-dimensional data is extracted, and the standard data is obtained after processing the data of the same order of magnitude. The standard data is subjected to nonlinear dimensionality reduction processing to obtain target data of a preset dimension. The generated target data is generated based on multi-dimensional data and is associated with multi-dimensional data, so that the characteristics of multi-dimensional data can be maintained, and then subsequent data processing and analysis can be performed through the target data.

[0147] 509, performing pooling processing on the target data to obtain a medical feedforward vector of the target treatment data;

[0148] In this embodiment, the pooling analysis is a pooling based on a convolutional neural network. In this embodiment, max pooling and avg pooling medical feature vectors are used for processing. A parallel double pooling layer combining max pooling and avg pooling is used for pooling operation, so as to retain deeper semantic information of the medical feature vector.

[0149] 510. According to the medical feedforward vector, the medical information set is analyzed based on a preset cosine similarity algorithm, and a medical service package corresponding to the disease is obtained.

[0150] Steps 501-503 and 510 in this embodiment are similar to steps 101-103 and 105 in the first embodiment, and will not be described here.

[0151] In the embodiment of the application, the original case data of the same disease is collected from the preset medical information platform, and the disease type and treatment data in the original case data are extracted. The treatment data is de-identified to obtain target treatment data. A plurality of key events in the target treatment data are extracted and fused to obtain a medical information set of the disease corresponding to the disease type. The target treatment data is input into a preset Bilstm model to obtain a medical feature vector of the medical data, and the medical feature vector is analyzed by pooling to obtain a medical feedforward vector. According to the medical feedforward vector, the medical information set is analyzed based on a preset cosine similarity algorithm, and a medical service package corresponding to the disease is obtained. The application obtains target treatment data with common characteristics by clustering analysis of original case data, and generates service packages corresponding to different types of diseases according to the target treatment data, solves the problem that current medical services lack sustainable tracking services and it is difficult to achieve timely and effective feedback, and effectively integrates medical data information, so as to solve the problem that the quality of mobile medical services cannot be guaranteed. The efficiency and portability of medical services are effectively improved, the waiting time of medical personnel is reduced, and the medical pressure is relieved.

[0152] The service package generation method based on patient data in the embodiment of the application is described above, and the service package generation device based on patient data in the embodiment of the application is described below. Please refer to Figure 6 The first embodiment of the service package generation device based on patient data in the embodiment of the application includes:

[0153] The acquisition module 601 is configured to collect original case data of the same disease from a preset medical information platform, and extract disease types and treatment data in the original case data.

[0154] The de-identification module 602 is configured to de-identify the treatment data to obtain target treatment data.

[0155] The fusion module 603 is configured to extract a plurality of key events in the target treatment data, and fuse the key events to obtain a medical information set of a disease corresponding to the disease type.

[0156] The pooling module 604 is configured to input the medical information set into a preset Bilstm model to perform vector calculation to obtain a medical feature vector of the disease, and perform pooling analysis on the medical feature vector to obtain a medical feedforward vector.

[0157] The clustering module 605 is configured to perform clustering analysis on the medical information set based on a preset cosine similarity algorithm according to the medical feedforward vector to obtain a medical service package corresponding to the disease.

[0158] In the embodiment of the application, the original case data of the same disease is collected from a preset medical information platform, and the disease type and treatment data in the original case data are extracted; the treatment data is de-identified to obtain target treatment data; a plurality of key events in the target treatment data are extracted, and the key events are fused to obtain a medical information set of a disease corresponding to the disease type; the target treatment data is input into a preset Bilstm model to obtain a medical feature vector of the medical data, and the medical feature vector is subjected to pooling analysis to obtain a medical feedforward vector; the medical information set is subjected to clustering analysis based on a preset cosine similarity algorithm according to the medical feedforward vector to obtain a medical service package corresponding to the disease. The application obtains target treatment data with common characteristics by clustering analysis on original case data, and generates different types of disease corresponding service packages according to the target treatment data, which solves the problem that the current medical service lacks sustainable tracking service, it is difficult to achieve timely and effective feedback, and medical data information lacks effective integration, resulting in the problem that the quality of mobile medical service cannot be guaranteed. The efficiency and portability of medical service are effectively improved, the waiting time of medical personnel is reduced, and the medical pressure is relieved.

[0159] Please refer to Figure 7 In the embodiment of the application, the second embodiment of the service package generation device based on patient data specifically includes:

[0160] The collection module 601 is configured to collect original case data of the same disease from a preset medical information platform, and extract the disease type and treatment data in the original case data;

[0161] The de-identification module 602 is configured to de-identify the treatment data to obtain target treatment data;

[0162] a fusion module 603, configured to extract a plurality of key events in the target treatment data, and perform fusion processing on the key events to obtain a medical information set of a disease corresponding to the disease type;

[0163] a pooling module 604, configured to input the medical information set into a preset Bilstm model to perform vector calculation to obtain a medical feature vector of the disease, and perform pooling analysis on the medical feature vector to obtain a medical feedforward vector;

[0164] a clustering module 605, configured to perform clustering analysis on the medical information set based on a preset cosine similarity algorithm according to the medical feedforward vector, to obtain a medical service package corresponding to the disease.

[0165] In this embodiment, the service package generation apparatus based on patient data further includes:

[0166] a setting module 606, configured to pre-set diagnosis and treatment templates of different disease types;

[0167] an acquisition module 607, configured to acquire time information corresponding to the treatment data, and acquire a diagnosis and treatment template corresponding to the treatment data according to the disease type;

[0168] a verification module 608, configured to verify the treatment data according to the diagnosis and treatment template, and determine whether the time sequence of the treatment data is correct.

[0169] In this embodiment, the de-identification module 602 is specifically configured to:

[0170] construct a treatment data query database according to the original case data;

[0171] add random noise to a sensitive attribute field in a data table of the treatment data query database according to a differential privacy algorithm;

[0172] perform de-identification processing on the sensitive attribute field in the original case data according to the random noise to obtain an identifier field;

[0173] perform encryption processing on the identifier field to obtain target treatment data.

[0174] In this embodiment, the fusion module 603 is specifically configured to:

[0175] acquire a predefined key event set;

[0176] perform screening processing on the target treatment data to filter invalid medical data in the target treatment data to obtain valid medical data;

[0177] According to the key event set, the effective medical data is extracted to obtain a plurality of key events in the target treatment data.

[0178] In this embodiment, the fusion module 603 is specifically further used for:

[0179] determining attribute features corresponding to the key events and weight values of the attribute features;

[0180] determining score data of the key events according to the attribute features and the weight values, and obtaining a medical information set of a disease corresponding to the disease type according to the score data.

[0181] In this embodiment, the pooling module 604 includes:

[0182] a feature extraction unit 6041, configured to perform feature extraction on the medical feature vector to obtain target medical features of the medical feature vector;

[0183] a dimension reduction unit 6042, configured to perform dimension reduction processing on the target medical features to obtain target data of a preset dimension;

[0184] a pooling unit 6043, configured to perform pooling processing on the target data to obtain a medical feedforward vector of the target treatment data.

[0185] In this embodiment, the dimension reduction unit 6042 is specifically used for:

[0186] obtaining frequencies of the target medical features;

[0187] extracting first data corresponding to the target medical features in the target treatment data;

[0188] performing data processing on different types of the first data to obtain standard data;

[0189] performing dimension reduction processing on the standard data to obtain target data of a preset dimension.

[0190] In the embodiment of the present application, the original case data of the same disease is collected from the preset medical information platform, and the disease type and treatment data in the original case data are extracted; the treatment data is de-identified to obtain target treatment data; a plurality of key events in the target treatment data are extracted, and the key events are fused to obtain a medical information set of the disease corresponding to the disease type; the target treatment data is input into a preset Bilstm model to obtain a medical feature vector of the medical data, and the medical feature vector is pooled and analyzed to obtain a medical feedforward vector; and the medical information set is clustered based on a preset cosine similarity algorithm according to the medical feedforward vector to obtain a medical service package corresponding to the disease. The present application obtains target treatment data with common characteristics through clustering analysis of original case data, and generates different types of disease corresponding service packages according to the target treatment data, which solves the problem that the current medical service lacks sustainable tracking service, it is difficult to achieve timely and effective feedback, and medical data information lacks effective integration, resulting in that the quality of mobile medical service cannot be guaranteed. The efficiency and portability of medical service are effectively improved, the waiting time of medical personnel is reduced, and the medical pressure is relieved.

[0191] The above Figure 6 And Figure 7 The service package generation device based on patient data in the embodiment of the present application is described in detail from the perspective of modular functional entities, and the service package generation device based on patient data in the embodiment of the present application is described in detail from the perspective of hardware processing.

[0192] Figure 8 The structure diagram of the service package generation device based on patient data provided by the embodiment of the present application, the service package generation device 800 based on patient data can have great differences due to different configurations or performances, and can include one or more processors (central processing units, CPU) 810 (for example, one or more processors) and a memory 820, one or more storage application programs 833 or data 832 storage media 830 (for example, one or more mass storage devices). Among them, the memory 820 and the storage medium 830 can be temporary storage or persistent storage. The program stored in the storage medium 830 can include one or more modules (not shown in the figure), each module can include a series of instruction operations in the service package generation device 800 based on patient data. Further, the processor 810 can be configured to communicate with the storage medium 830, execute a series of instruction operations in the storage medium 830 on the service package generation device 800 based on patient data, to realize the steps of the service package generation method based on patient data provided by each method embodiment.

[0193] The patient data based service package generating device 800 can also include one or more power supplies 840, one or more wired or wireless network interfaces 850, one or more input / output interfaces 860, and / or one or more operating systems 831, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art can understand that Figure 8 The illustrated patient data based service package generating device structure does not constitute a limitation to the patient data based service package generating device provided in the present application, and can include more or fewer components than illustrated, or combine certain components, or arrange the components differently.

[0194] The present application also provides a computer readable storage medium, which can be a non-volatile computer readable storage medium, or a volatile computer readable storage medium. The computer readable storage medium stores instructions, which, when executed on a computer, cause the computer to perform the steps of the patient data based service package generating method described above.

[0195] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0196] The integrated units, if implemented in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application, essentially or the part that contributes to the prior art, or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program code storage media.

[0197] The above-described embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced by equivalent replacements; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A service package generation method based on patient data, characterized by, The service package generation method based on patient data comprises: Collecting original case data of the same disease from a preset medical information platform, and extracting disease types and treatment data in the original case data; De-identifying the treatment data to obtain target treatment data; Extracting a plurality of key events in the target treatment data, and fusing the key events to obtain a medical information set of the disease corresponding to the disease type; Inputting the medical information set into a preset Bilstm model for vector calculation to obtain a medical feature vector of the disease, and performing pooling analysis on the medical feature vector to obtain a medical feedforward vector; According to the medical feedforward vector, performing clustering analysis on the medical information set based on a preset cosine similarity algorithm to obtain a medical service package corresponding to the disease; The extracting a plurality of key events in the target treatment data comprises: obtaining a pre-defined key event set; filtering invalid medical data in the target treatment data by screening the target treatment data to obtain valid medical data; and extracting the valid medical data according to the key event set to obtain a plurality of key events in the target treatment data; The fusing the key events to obtain a medical information set of the disease corresponding to the disease type comprises: determining attribute characteristics corresponding to the key events and weight values of the attribute characteristics; determining score data of the key events according to the attribute characteristics and the weight values, and obtaining a medical information set of the disease corresponding to the disease type according to the score data; The pooling analysis on the medical feature vector to obtain a medical feedforward vector comprises: feature extraction on the medical feature vector to obtain target medical features of the medical feature vector; dimension reduction processing on the target medical features to obtain target data of a preset dimension; and pooling processing on the target data to obtain a medical feedforward vector of the target treatment data; The dimension reduction processing on the target medical features to obtain target data of a preset dimension comprises: obtaining a frequency of the target medical features; extracting first data corresponding to the target medical features in the target treatment data; data processing on different types of the first data to obtain standard data; and dimension reduction processing on the standard data to obtain target data of a preset dimension.

2. The service package generation method based on patient data according to claim 1, wherein, After the collecting original case data of the same disease from a preset medical information platform, and extracting disease types and treatment data in the original case data, further comprising: Pre-setting diagnosis and treatment templates of different disease types; Obtaining time information corresponding to the treatment data; According to the disease type, obtaining a diagnosis and treatment template corresponding to the treatment data; According to the diagnosis and treatment template, verifying the treatment data to determine whether the time sequence of the treatment data is correct.

3. The service package generation method based on patient data according to claim 1, wherein, The de-identifying the treatment data to obtain target treatment data comprises: Constructing a treatment data query database according to the original case data; adding random noise to a sensitive attribute field in a data table of the treatment data query database according to a differential privacy algorithm; de-identifying the sensitive attribute field in the original case data according to the random noise to obtain an identifier field; encrypting the identifier field to obtain target treatment data.

4. A service package generation apparatus based on patient data, characterized by, The service package generation device based on patient data comprises: a collection module configured to collect original case data of the same disease from a preset medical information platform, and extract a disease type and treatment data in the original case data; a de-identification module configured to de-identify the treatment data to obtain target treatment data; a fusion module configured to extract a plurality of key events in the target treatment data, and fuse the key events to obtain a medical information set of the disease corresponding to the disease type; a pooling module configured to input the medical information set into a preset Bilstm model for vector calculation to obtain a medical feature vector of the disease, and perform pooling analysis on the medical feature vector to obtain a medical feedforward vector; a clustering module configured to perform clustering analysis on the medical information set based on a preset cosine similarity algorithm according to the medical feedforward vector to obtain a medical service package corresponding to the disease; The fusion module is specifically configured to: obtain a predefined key event set; filter invalid medical data in the target treatment data by performing screening processing on the target treatment data to obtain valid medical data; and extract a plurality of key events in the target treatment data according to the key event set; The fusion module is specifically further configured to: determine an attribute feature corresponding to the key event and a weight value of the attribute feature; determine score data of the key event according to the attribute feature and the weight value, and obtain a medical information set of the disease corresponding to the disease type according to the score data; The pooling module comprises: a feature extraction unit configured to perform feature extraction on the medical feature vector to obtain a target medical feature of the medical feature vector; a dimension reduction unit configured to perform dimension reduction processing on the target medical feature to obtain target data of a preset dimension; and a pooling unit configured to perform pooling processing on the target data to obtain a medical feedforward vector of the target treatment data; The dimension reduction unit is specifically configured to: obtain a frequency of the target medical feature; extract first data corresponding to the target medical feature in the target treatment data; perform data processing on different types of the first data to obtain standard data; and perform dimension reduction processing on the standard data to obtain target data of a preset dimension.

5. A service package generating apparatus based on patient data, characterized by, The service package generation device based on patient data comprises a memory and at least one processor, the memory has instructions stored therein, and the memory and the at least one processor are interconnected by a circuit; The at least one processor invokes the instructions in the memory, so that the service package generation device based on patient data performs each step of the service package generation method based on patient data in any one of claims 1-3.

6. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program, which is executed by a processor, realizes each step of the service package generation method based on patient data as claimed in any one of claims 1-3.

Images