Device, method, equipment and medium for identifying and managing critical illness physical examination users

Abstract

The present invention provides a device, method, equipment and medium for identifying and managing users of critical medical examinations. The method is based on the original data of the physical examination system, and after relevant preprocessing, a standardized data set is obtained; then, a positive-negative prediction is performed, and the positive data set is used as input to determine whether it meets the specific critical illness according to the screening criteria for critical illness, and finally whether the physical examination user is a critical illness user, as well as relevant instructions. The method provided by the present invention has an average recognition accuracy of 95.75% in 44 types of critical diseases that are the focus of attention during the physical examination process.

Description

Technical Field

[0001] The present invention belongs to the technical field of medical system management, and relates to a device, method, equipment and medium for identifying and managing critically ill physical examination users during a physical examination. Background Art

[0002] With the rapid development of AI technology and its disruptive applications in certain fields, AI's position in the industry is becoming more and more prominent, and its popularity and penetration in business are increasing. Natural language processing, as one of the most important components of AI technology, has been put into practical application in many aspects of life. The accuracy of AI determines the degree of market recognition and promotion of an AI technology, so high-accuracy AI technology is crucial.

[0003] At present, natural language processing rarely involves the field of screening for critical illnesses during physical examinations, and no accuracy baseline has been established in this field. One reason is that the medical screening logic for critical illnesses is very complex and sophisticated enough. How to apply the medical screening logic for critical illnesses to the physical examination process is a very difficult problem. Another reason is that a large amount of physical examination data is generated during the physical examination process. Using manual methods to screen out the less common critical examination data from the massive amount of physical examination data is a boring, repetitive, inefficient, and extremely time-consuming task.

[0004] As a physical examination unit, the physical examination center has the obligation to notify critically ill users as soon as possible during the examination process. Therefore, it urgently needs a system for critical illness identification and management. Summary of the invention

[0005] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a device, method, equipment and medium for identifying and managing users of critical and severe physical examinations. The method converts the raw data provided by the physical examination system into a standard data format, and then performs a positive and negative prediction on the processed data set, thereby identifying and managing the relevant information of various users in the main body of the physical examination data, including critical and severe users.

[0006] To achieve this object, the present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a device for identifying and managing critical illness physical examination users, the device comprising:

[0008] The original physical examination data collection module is used to obtain the physical examination data of the physical examination user from the physical examination system;

[0009] The standardized physical examination data module is used to standardize the test data and examination data in the original physical examination data to obtain data that conforms to the standard format of the critical care system;

[0010] The standardized physical examination data module includes a structured text data unit, which is used to perform structured processing on the examination data, and after the structured processing, the organs, tissues and symptoms in the examination data meet the atomicity requirements;

[0011] A module for predicting the positive and negative nature of physical examination data, used to predict the positive and negative nature of the test data and the structured inspection data, and retain the positive data;

[0012] The critical illness identification and screening module identifies positive data one by one according to the critical illness screening criteria, determines the existence of critical illness of the physical examination user corresponding to the original physical examination data, and retains the physical examination data that meets the critical illness criteria;

[0013] The critical care front-end rendering and display module renders and displays the relevant data of critical care users on the front-end for doctors or people in need to review.

[0014] The device for identifying and managing critically ill physical examination users described in the present invention can collect real physical examination data and standardize the physical examination data, so as to predict whether the physical examination person has the risk of suffering from critical illness; of course, this step is not the final diagnosis result, and the final diagnosis still needs to be completed by the doctor. The device predicts risks and retains the relevant data of critically ill users, which is convenient for doctors or people in need to manage.

[0015] In the standardized physical examination data module of the present invention, the operations of structurally processing the examination data include:

[0016] An original complete medical description sentence is split into multiple sentences, each sentence only contains one organ and its corresponding symptom, and the components of each sentence are analyzed, including organs, tissues and symptoms.

[0017] Generally speaking, original physical examination data includes: test data and examination data.

[0018] Test data are generally numerical data, such as blood routine, urine routine or blood lipid data, while examination data are generally text data, such as X-ray, CT, MRI, ultrasound, electrocardiogram or pathology data.

[0019] Preferably, the critical care system standard format includes: user id, name, value (content), unit, normal range, type, user id and hospital id, etc.

[0020] During the preprocessing process, the test data and the inspection data need to be split into the standard format that conforms to the critical care system. For items such as units or normal ranges, if there is specific content, take the specific content, if not, take None.

[0021] In the present invention, the inspection data needs to be structured, specifically: an original complete medical description sentence needs to be split into multiple sentences, each sentence only contains one place with one symptom.

[0022] In the present invention, the method for predicting positive and negative sex by the physical examination data positive and negative sex prediction module includes:

[0023] For test data, the data within the normal range is negative, and the data outside the normal range is positive;

[0024] For structured inspection data, each sentence contains only one place and one symptom, so only the symptom needs to be predicted. That is, the yin and yang of the sentence has nothing to do with the place (organ, tissue), but only with the symptom.

[0025] For symptom prediction of yin and yang, the BiLSTM-attention deep learning framework can be used for negative and positive binary classification.

[0026] In the present invention, the screening criteria for critical illness are divided into two situations: single index and multiple indexes.

[0027] The single indicator is: if a certain indicator of a physical examination user meets the conditions, it is a specific critical illness;

[0028] The multiple indicators are: a physical examination user needs to meet multiple indicators at the same time to be considered a specific critical illness.

[0029] The critical screening criteria cover two types of data: tests and examinations.

[0030] The format of the critical illness screening criteria is stored in json format and read by the system.

[0031] The standard format of the critical illness test is as follows:

[0032]

[0033] The standard format of the critical examination category is as follows:

[0034]

[0035]

[0036] For the critical illness criteria of the test and examination categories, the id, criterion_name and risk_level are consistent.

[0037] The default value of id is Null. After the system is started, the SEQUENCE of the database takes a constant value within the life cycle. The value of criterion_name is based on the specific name of the critical disease, and the value of risk_level is based on the critical or severe disease.

[0038] In the test data, check_name is based on the specific name of the test; type is generally numeric; source is the test; the normal range is infimum and supremum;

[0039] In the inspection data, type is generally textual; source is inspection;

[0040] In the examination data, if organ needs to specify a specific organ word, the value is ["organ"]; if it does not need to be specified, the value is [None]; if only one or more of the multiple organs are required, the value is ["organ1|organ2|organ3|...organN"];

[0041] In the inspection data, if the word tissue needs to specify a specific tissue word, the value is ["tissue"]; if it does not need to be specified, the value is [None]; if only one or more of the multiple organs need to be satisfied, the value is ["tissue1|tissue2|tissue3|...tissueN"]; if all tissues need to be satisfied at the same time, the value is ["tissue1", "tissue2", "tissue3"..., "tissueN"];

[0042] The logical relationships of or and can be combined arbitrarily, for example: ["tissue11|tissue12","tissue21","tissue31|tissue32|tissue33"].

[0043] In the inspection data, the value logic of direction is consistent with that of tissue.

[0044] In the examination data, if disease needs to specify a specific disease, the value is ["disease"]; if it does not need to be specified, the value is [None]; if only one or several diseases of multiple organs need to be satisfied, the value is ["disease1|disease2|disease3|...|diseaseN"]; if all diseases need to be satisfied at the same time, the value is ["disease1", "disease2", "disease3",..., "diseaseN"];

[0045] The logical relationships of or and can be combined arbitrarily, for example: ["disease11|disease12|disease13", "disease21|disease22", "disease31"];

[0046] If a disease cannot exist, the value is [^(disease)];

[0047] If multiple diseases cannot exist, the value is [^(disease1|disease2)];

[0048] The logical relationships of or, and, and not can be combined arbitrarily, for example: ["disease11", "disease21|disease22", "^(disease31|disease32)"].

[0049] In the inspection data, the value logic of digree is consistent with that of disease.

[0050] In the inspection data, if source needs to indicate the source, the value is source; the value of source is 'discription' or 'impression'; under certain conditions, 'impression' may not exist.

[0051] If an existence constraint condition needs to be met, that is, if impresion does not exist, there is no requirement; if impresion exists, a constraint is required, then source takes the value if:source.

[0052] In the present invention, the information of the medical examination user with a critical illness includes:

[0053] User's personal information, such as name, gender, age and contact information;

[0054] The user's original physical examination information, such as blood routine, abdominal ultrasound and pathology, etc.;

[0055] The user's critical illness information, such as the name of the critical illness and the user's physical examination data that meets the critical illness criteria; the physical examination data includes name, value, unit, etc.;

[0056] Information describing critical illness, such as: description of conditions that meet the criteria for critical illness, description of the basis for screening for critical illness, etc.

[0057] In the present invention, the device may also include a critically ill user management module for managing the information of critically ill users; for example: when new critically ill user data has not been reviewed by a doctor, it will be displayed as unreviewed information on the front end; when the doctor clicks to review the critically ill user information, the user will enter the history library and archive.

[0058] Preferably, the collection module of the original physical examination data is connected to the physical examination system of the hospital. However, the new physical examination data generated by the physical examination system of the hospital will be very large, especially for tertiary hospitals. Therefore, the device provided by the present invention should have the characteristics of high concurrency, low latency, strong fault tolerance and strong robustness.

[0059] In the present invention, the critical care system will have a larger and larger data volume over time. In order to ensure that the system can maintain high performance and lasting operation, it is necessary to avoid the bigTable situation. Therefore, the database should divide the main table into two parts: the current active part and the historical archive part. The current active part only stores the data of the last 15 days, and the historical archive part stores the data before 15 days.

[0060] Only the currently active table is frequently accessed, and the historical archived table is rarely accessed. In order to maintain the currently active table and the historical archived table, an ETL process is required to refresh in real time in the background to maintain the integrity and consistency of the currently active table and the historical archived table.

[0061] In order to make the critical care system achieve high fault tolerance, high concurrency, low latency, and strong robustness. The data processed by each module should be stored in the corresponding table of the database. Each module should act as an independent process, read the data in the corresponding table from the database, and write it to the downstream table. Considering the confidentiality and security of the hospital's internal physical examination data, the critical care system is generally deployed in the hospital's internal network and is not allowed to be accessed by external networks.

[0062] It should be noted that it should be understood that the division of the various modules of the above device is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. These modules can all be implemented in the form of software called by processing elements; they can also be all implemented in the form of hardware: some modules can also be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. For example, the collection module of the original physical examination data can be a separately established processing element, or it can be integrated in a certain chip. In addition, it can also be stored in the memory in the form of program code, and called and executed by a certain processing element. The function of the above acquisition module. The implementation of other modules is similar. In addition, all or part of these modules can be integrated together, or they can be implemented independently. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or instructions in the form of software.

[0063] In a second aspect, the present invention provides a method for critical illness identification and management for non-disease diagnosis and treatment purposes using the device as described in the first aspect, the method comprising the following steps:

[0064] S1. Obtaining original physical examination data based on the physical examination system;

[0065] S2. Preprocessing the original physical examination data to obtain data that conforms to the standard format of the critical care system;

[0066] Among them, the examination data is structured, and the organs, tissues and symptoms in each piece of data obtained after the structured processing meet the atomicity requirements;

[0067] S3, predicting the positive and negative data obtained after the preprocessing, and retaining the positive data;

[0068] S4. Screen the positive data according to the screening criteria for critical illness, and determine whether the critical illness exists in the physical examination user corresponding to the original physical examination data;

[0069] For users who are undergoing physical examinations and are in critical condition, generate reasons for judgment and retain the basis;

[0070] S5. Display the information of the user who is undergoing physical examination and has a critical illness to the doctor.

[0071] It should be noted that the method for identifying and managing critically ill physical examination users described in the present invention is not for the purpose of diagnosis and treatment of diseases. Its main purpose is to remind physical examination users that they may be at risk of disease and to remind patients to conduct more detailed screening. At the same time, the method described in the present invention can also screen physical examination data (including healthy users and critically ill users) to complete the management and preservation of physical examination user information.

[0072] In a third aspect, the present invention provides a computer device, the device comprising:

[0073] one or more processors;

[0074] A storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method for identifying and managing critical care examination users as described in the second aspect.

[0075] In a fourth aspect, a computer storage medium stores a computer program, which, when executed by a processor, implements the method for identifying and managing critical care examination users as described in the second aspect.

[0076] All or part of the steps of the above method can be completed by hardware related to the computer program. The above computer program can be stored in a computer-readable storage medium.

[0077] When the program is executed, it executes the steps of the above-mentioned method embodiments; the computer storage medium may include but is not limited to floppy disks, optical disks, CD-ROMs (read-only disk memories), magneto-optical disks, ROMs (read-only memory), RAMs (random access memory), EPROMs (erasable programmable read-only memory), EEPROMs (electrically erasable programmable read-only memory), magnetic cards or optical cards, flash memory, or other types of media / machine-readable media suitable for storing machine-executable instructions.

[0078] The computer storage medium may be a product that is not connected to a computer device, or a component that is connected to a computer device for use.

[0079] Compared with the prior art, the present invention has the following beneficial effects:

[0080] In the present invention, the device or method for identifying and managing critical health examination users can structure the textual health examination data, that is, the components such as organs, tissues, symptoms, etc. of the original health examination data are stripped one by one and redistributed, so that each clause contains only one position and one symptom;

[0081] At the same time, the present invention logically deconstructs the critical illness screening standards in a relatively authoritative manner, and uses a json language to simply and highly accurately model the critical illness screening standards, thereby obtaining an authoritative and consensus-based detection method and device that can detect, analyze and screen 44 types of critical illness conditions that are of focus during physical examinations, and the average accuracy rate for screening critically ill patients reaches 95.75%. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 It is an internal logic diagram of a solution for identifying and managing critically ill physical examination users during a physical examination, provided by an embodiment of the present invention.

[0083] Figure 2 It is a system architecture design diagram for identifying and managing critically ill physical examination users during a physical examination, provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0084] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments.

[0085] The components of the embodiments of the present application generally described and shown in the drawings herein may be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. For the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application.

[0086] Example 1

[0087] This embodiment provides a solution for identifying and managing critical and severe medical examination users during a medical examination. Figure 1 , the method comprising:

[0088] S1. Obtain original physical examination data; original physical examination data includes test data and examination data. Test data is generally numerical data, such as blood routine, urine routine, blood lipids and other data. Examination data is generally text data, such as X-ray, CT, MRI, ultrasound, electrocardiogram, pathology and other data.

[0089] S2. Preprocess the original physical examination data and send it to the critical illness identification and management system; the preprocessing includes formatting the inspection data and structuring the inspection data.

[0090] Standard format of critical care system: Test data and inspection data should be divided into basic formats. Each basic format includes: id, name, value (content), unit, normal range, type, user id, hospital id; for unit and normal range, if there is content, take the specific content, if not, take None.

[0091] Among them, the standard format of critical illness for laboratory tests is as follows:

[0092]

[0093]

[0094] In a specific example, the original sentence is "

Examination findings

Impression

[0095] 1. The cross-section of both lobes of the thyroid gland is normal in shape and size;

[0096] 2. The surface of the thyroid gland is smooth;

[0097] 3. The thyroid capsule is intact;

[0098] 4. The inside of the thyroid gland is a uniform fine dot-shaped echo;

[0099] 5. No obvious abnormality was found in the thyroid gland;

[0100] Also, analyze the components of each sentence, for example:

[0101] "The cross-sections of both lobes of the thyroid gland are normal in shape and size";

[0102] Sentence components: ["organ", "tissue", "symptom"];

[0103] Split interval: [[0,3],[3,6],[6,13]].

[0104] In another specific embodiment, the standard format of some critical illnesses of a certain physical examination user is as follows:

[0105]

[0106]

[0107] Among them, the id, criterion_name, and risk_level in the critical illness standards for tests and examinations are the same.

[0108] The default value of id is Null. After the system is started, the SEQUENCE of the database takes a constant value within the life cycle. The value of criterion_name is based on the specific name of the critical disease, and the value of risk_level is based on the critical or severe disease.

[0109] In the test data, check_name is based on the specific name of the test; type is generally numeric; source is the test; the normal range is infimum and supremum;

[0110] In the inspection data, type is generally textual; source is inspection;

[0111] In the data, the design logic of AND, OR, NOT, and EXISTENCE of organ, direction, tissue, disease, degree, and source is set as needed.

[0112] S3, predicting the positive and negative of the preprocessed inspection numerical data and structured inspection text data, and retaining the positive data;

[0113] For test data, the data within the normal range is negative, and the data outside the normal range is positive;

[0114] For structured inspection data, each sentence contains only one place and one symptom, so only the symptom needs to be predicted. That is, the yin and yang of the sentence has nothing to do with the place (organ, tissue), but only with the symptom.

[0115] For symptom prediction of yin and yang, the BiLSTM-attention deep learning framework can be used for negative and positive binary classification.

[0116] S4. According to the critical illness standard, the positive physical examination data of the user is judged to obtain the physical examination data that meets the critical illness standard;

[0117] Among them, the screening criteria for critical illness are divided into single indicator and multiple indicators.

[0118] It should be noted that the following points should be noted when generating critical illness screening criteria:

[0119] In some specific embodiments, the Expert Consensus on the Management of Important Abnormal Results of Health Examination, the interpretation of professional and authoritative doctors, and the real physical examination data can be organically combined as a reliable screening standard applicable to the critical care system, and the accuracy of the method described in the present invention depends largely on the screening standard;

[0120] The "Expert Consensus on the Management of Important Abnormal Results in Health Examinations" serves as an outline guiding direction in the present invention, and is not an instruction document for guiding development; at the same time, the accuracy of critical illness identification in the present invention depends in part on the screening criteria for critical illness. If, with further development of technology, this field can provide a more comprehensive and standard physical examination result evaluation standard, the present invention can be further improved and combined with the new evaluation standard to identify and manage critical illness, so as to achieve further optimization results.

[0121] At the same time, when drafting the screening criteria for critical illness, full consideration should be given to the role of organ, direction, tissue, disease, degree, and source in the criteria, as well as the potential impact of the design logic of and, or, not, and existence; after comprehensively considering the above impacts, a screening criterion with high coverage, high accuracy, high generalization rate, and extremely simpleness is used as the screening criterion for the method described in the present invention.

[0122] Here, the "Expert Consensus on the Management of Important Abnormal Results of Health Examinations" in the present invention is an outline guideline, not an instruction document for guiding development.

[0123] S5. Display the information of the critical illness physical examination user: including the user's personal information, the user's original physical examination information, the user's critical illness information, and related critical illness description information, etc.

[0124] In a specific embodiment, the user's critical illness information includes: the user's personal information, such as name, gender, age, and contact information; the user's original physical examination information, such as blood routine, abdominal color ultrasound, pathology, etc.; the user's critical illness information, such as the name of the critical illness, the user's physical examination data that meets the critical illness, including name, value, unit, etc.; critical illness description information, such as: description of the conditions that meet the critical illness, description of the basis for screening for critical illness, etc.

[0125] In certain specific embodiments, the method provided by the present invention enables the scheme to achieve an accuracy rate of 95.75% on real data in 44 critical cases, thereby meeting the requirements of the hospital physical examination department very well; even if subsequent screening standards put forward new requirements, the corresponding needs can be quickly met with very little change.

[0126] Example 2

[0127] This embodiment provides a device for identifying and managing critically ill physical examination users during a physical examination, which can implement the software environment relied on by the above-mentioned method for identifying and managing critically ill physical examination users during a physical examination, and provide stable, mature and standardized web services therefor.

[0128] like Figure 2 As shown, the device for identifying and managing critically ill physical examination users during a physical examination includes:

[0129] The module for collecting original physical examination data: scans the physical examination data from the physical examination system in real time, once every 5 minutes, and writes the newly generated physical examination data within the last 5 minutes into the rabbitMq queue;

[0130] In the present invention, since the data from hospital examinations are highly concurrent and rapidly growing data, new data is scanned from the system of the physical examination hospital and uploaded to the critical care system. This requires the use of message queue software similar to rabbitMq to meet the requirements of high concurrency, high fault tolerance, low latency, and strong stability, and further achieve the effect of low coupling.

[0131] Standardized physical examination data module: write rabbitMq queue data to the database process, read data from the rabbitMq queue, parse the data, and write it to the corresponding table of the Postgresql database;

[0132] Furthermore, the inspection data is formatted and the inspection data is structured.

[0133] Physical examination data positive and negative prediction module: reads standardized data from the database, performs positive and negative prediction, and finally writes positive data into the database.

[0134] Critical illness identification and screening module: positive data is identified one by one according to the user; in order to make the critical illness system achieve high fault tolerance, high concurrency, low latency and other performance, the data processed by each module should be stored in the corresponding table of the database; each module should act as an independent process, read the data in the corresponding table from the database, write it to the downstream table, and write the data that meets the critical illness into the corresponding table of the database.

[0135] The device described in the present invention also includes: the ETL process maintains the current active tables and historical archive tables in the database in real time; when designing the database, most of the core business tables must be designed into two parts: the current active and historical archive, and the ETL process is used to maintain their consistency in real time.

[0136] The device of the present invention also includes a front-end display service, which reads the corresponding information from the database, performs web rendering, and finally displays it to the doctor for operation by users, such as doctors, physical examination users or managers. Considering the confidentiality and security of physical examination data within the hospital, the critical care system is generally deployed in the hospital's intranet and is not allowed to be accessed by external networks.

[0137] When new critically ill user data has not been reviewed by a doctor, it will be displayed as unreviewed information on the front end; when the doctor clicks to review the critically ill user information, the user will enter the history library and archive.

[0138] The device of the present invention uses the form of microservices, and each process is responsible for a core module. This can achieve low coupling and high cohesion of the system, while improving the concurrency of the system and reducing the impact of bugs on the system.

[0139] In the present invention, the screening standards for critical illness are summarized by interpreting representative real physical examination data and authoritative critical illness screening standards. The method and device constructed by combining the technology in the field of natural language processing have a very high accuracy rate in judging critically ill patients during the physical examination process, filling the gap in this field.

[0140] The applicant declares that the above is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Those skilled in the art should understand that any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention shall fall within the protection scope and disclosure scope of the present invention.

Claims

1. A device for identifying and managing critical care examination users, characterized in that: The device comprises: The original physical examination data collection module is used to obtain the physical examination data of the physical examination user from the physical examination system; the original physical examination data includes: test data and examination data, the test data includes blood routine, urine routine or blood lipids, and the examination data includes X-ray, CT, nuclear magnetic resonance, ultrasound, electrocardiogram or pathology; The standardized physical examination data module is used to standardize the test data and examination data in the original physical examination data to obtain data that conforms to the standard format of the critical care system; The standardized physical examination data module includes a structured text data unit, which is used to perform structured processing on the examination data, and after the structured processing, the organs, tissues and symptoms in the examination data meet the atomicity requirements; A module for predicting the positive and negative nature of physical examination data, used to predict the positive and negative nature of the test data and the structured inspection data, and retain the positive data; The critical illness identification and screening module identifies positive data one by one according to the critical illness screening criteria, determines the existence of critical illness of the physical examination user corresponding to the original physical examination data, and retains the physical examination data that meets the critical illness criteria; The critical illness front-end rendering and display module renders and displays the relevant data of critical illness users on the front-end; In the standardized physical examination data module, the operations of structurally processing the examination data include: Split an original complete medical description sentence into multiple sentences, each sentence only contains one organ and one corresponding symptom, and analyze the components of each sentence, including organs, tissues and symptoms; The method for the physical examination data positive / negative prediction module to perform positive / negative prediction includes: For test data, the data within the normal range is negative, and the data outside the normal range is positive; for structured inspection data, the symptom in each sentence is predicted to be positive or negative; for the prediction of the symptom to be positive or negative, the BiLSTM-attention deep learning framework is used to perform negative and positive binary classification; The screening criteria for critical illness are divided into two types: single indicator and multiple indicators; The single indicator is: if a physical examination user meets one indicator, he / she is judged to have a specific critical illness; The multiple indicators are: a user undergoing physical examination meets multiple indicators at the same time and is diagnosed with a specific critical illness; The critical screening criteria include two types of data: testing and examination; The format of the critical illness screening criteria is stored in json format and read by the system.

2. The device for identifying and managing critical illness physical examination users according to claim 1, characterized in that: The relevant data of the critically ill users include: The user's personal information, the user's original physical examination information, the user's critical illness information, and description information of the critical illness information.

3. The device for identifying and managing critical illness physical examination users according to claim 1, characterized in that: The device also includes a critically ill user management module for managing information of critically ill users.

4. The device for identifying and managing critical illness physical examination users according to claim 1, characterized in that: The table in the database of the device is divided into two parts: currently active and historical archive; Use the ETL process to maintain the currently active table and the historical archived table to achieve real-time background refresh; The data processed by each module of the system is stored in the corresponding table of the database, and each module acts as an independent process, reading the data in the corresponding table from the database and writing it to the downstream table.

5. A method for critical illness identification and management for non-disease diagnosis and treatment purposes using the device as claimed in any one of claims 1 to 4, characterized in that: The method comprises the following steps: S1. Obtaining original physical examination data based on the physical examination system; S2. Preprocessing the original physical examination data to obtain data that conforms to the standard format of the critical care system; Among them, the examination data is structured, and the organs, tissues and symptoms in each piece of data obtained after the structured processing meet the atomicity requirements; S3, predicting the positive and negative data obtained after the preprocessing, and retaining the positive data; S4. Screen the positive data according to the screening criteria for critical illness, and determine whether the critical illness exists in the physical examination user corresponding to the original physical examination data; For users who are undergoing physical examinations and are in critical condition, generate reasons for judgment and retain the basis; S5. Display the information of the user undergoing physical examination who has a critical illness.

6. A computer device, characterized in that: The device comprises: one or more processors; A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, enables the one or more processors to implement the method for identifying and managing critical care examination users as described in claim 5.

7. A computer storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for identifying and managing critical illness physical examination users as claimed in claim 5 is implemented.

Images