Integration method and system for integrating multi-source medical data set

Abstract

The invention relates to the technical field of multi-source medical data integration, in particular to an integration method and system for integrating multi-source medical data sets. Comprising the following steps: S1, collecting multi-source medical data into an LIMS system, and associating the multi-source medical data according to a patient identifier; s2, extracting historical patient diagnosis symptoms according to the patient identifier to establish a patient symptom model, inputting associated medical data after association according to the historical patient diagnosis symptoms, and outputting a symptom related report through the patient symptom model; the medical knowledge graph is extracted from the multi-source medical data, accuracy analysis is carried out on disease related reports, correction is carried out according to different types of disease related reports, the accuracy of diagnosis reports is greatly improved, then the corrected reports are integrated into a disease diagnosis model, training and optimization are carried out through a machine learning algorithm, and the accuracy of the diagnosis reports is improved. The model can learn a large amount of medical knowledge and diagnosis modes.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-source medical data integration, and in particular to an integration method and system for integrating multi-source medical data sets. Background Art

[0002] The processing and analysis of medical data plays a vital role in disease diagnosis, treatment plan formulation and medical research. At present, in the daily diagnosis and treatment process of hospitals, the HIS system is used to manage the basic information of patients, registration, charging and other processes, and the LIMS system (Laboratory Information Management System) focuses on the recording and analysis of laboratory test data. However, the data source is single, mainly concentrated in the diagnosis and treatment data within the hospital, and it is difficult to obtain comprehensive medical information. External data such as scientific research databases, public health data platforms and clinical research project data are difficult to effectively integrate, which seriously affects the comprehensive assessment of patients' conditions and in-depth research on diseases. Therefore, an integrated method and system for integrating multi-source medical data sets are proposed. Summary of the invention

[0003] The object of the present invention is to provide an integration method and system for integrating multi-source medical data sets to solve the problems raised in the above background technology.

[0004] In order to solve the above technical problems, one of the objectives of the present invention is to provide an integration method for integrating multi-source medical data sets, comprising the following steps: S1. Collect multi-source medical data into the LIMS system and associate the multi-source medical data according to the patient identification; S2. Extracting historical patient diagnosis symptoms according to the patient identifier to establish a patient symptom model, inputting the associated medical data after association according to the historical patient diagnosis symptoms, and outputting a symptom-related report through the patient symptom model; S3. Extract medical knowledge graphs from multi-source medical data, aggregate and associate the disease-related reports of each patient according to the disease type, and then perform accuracy analysis on the disease-related reports in the disease type according to the medical knowledge graph; S4. Conduct mutual correction analysis on low-accuracy disease-related reports based on the corresponding accuracy and disease type; S5. Integrate the disease-related reports after the corrective analysis into the LIMS system as a disease diagnosis model, and use the disease diagnosis model to perform disease analysis on the patient.

[0005] As a further improvement of the present technical solution, the S1 realizes the simultaneous external and internal acquisition of medical data by connecting various medical data generating devices, scientific research databases, public health data platforms, and clinical research project data, and then builds a medical data warehouse in the LIMS system by combining relational databases with non-relational databases.

[0006] As a further improvement of the present technical solution, S1 extracts patient data from a medical data warehouse, generates separate patient identifiers based on the patient data, and associates multi-source medical data based on the patient identifiers to obtain associated medical data.

[0007] As a further improvement of the technical solution, the steps of S2 are as follows: S2.1. Extract the corresponding historical patient diagnosis symptoms from the multi-source medical data according to the patient identifier, and then extract the historical symptom type according to the historical patient diagnosis symptoms; S2.2, analyze the relevant values ​​of the symptoms of the associated medical data according to the historical patient diagnosis, set the relevant threshold for the correlation of the symptoms according to the patient information, and then filter the associated medical data by combining the correlation threshold with the relevant values ​​of the symptoms; S2.3. Input the screened associated medical data and historical disease types into the patient disease model, and then output the corresponding disease-related report for each disease in the historical disease type through the patient medical record model.

[0008] As a further improvement of the technical solution, the steps of generating a symptom-related report in S2 are as follows: ; Where T is the set of disease types, is the set of disease types that appear in the patient's historical diagnosis data, t is the disease type in the set T, is the disease type field corresponding to disease type t, D is the patient's historical diagnosis data, d is the data record of multi-source medical data, is the disease field in the data record d, and f is the mapping function from the disease field to the disease type; ; Among them, r is the correlation coefficient, n is the total number of historical patient diagnosis diseases, is the disease status value of the i-th sample disease, is the physical diagnostic feature value of the i-th sample disease, and are the means of x and y respectively; ; in, is the correlation threshold set, α is the mean correlation coefficient set according to the historical patient diagnosis, The standard deviation of the correlation coefficient was set for the diagnosis of the disease based on historical patients; ; Among them, S is the filtered data set, is the correlation coefficient corresponding to each data point v in the associated medical data set, and K is the original associated medical data; ; in, is the predicted value of the continuous value related to the ith sample disease, is the jth physical diagnosis feature value of the i-th sample disease, is the intercept, is the corresponding coefficient of the jth physical diagnosis feature value, and P is the number of features.

[0009] As a further improvement of the technical solution, the steps of S3 are as follows: S3.1. Extract medical knowledge graph from multi-source medical data, and then summarize and associate each patient’s disease-related reports according to disease type; S3.2. Perform accuracy analysis on disease-related reports in disease types based on the medical knowledge graph to obtain the accuracy of each disease-related report in the disease type. Then set the accuracy mean based on the total accuracy of each case type and the case-related report data. Values ​​above the mean are considered high accuracy, and vice versa, values ​​below the mean are considered low accuracy.

[0010] As a further improvement of the technical solution, the steps of determining low-accuracy disease-related reports and high-accuracy disease-related reports in S3 are as follows: ; in, is the total disease-related reports of disease type t, E is the total number of patients, For the wth patient, For patients Reports on symptoms of The disease type field in the disease-related report g; ; in, is the accuracy score of the disease-related report g, KG is the medical knowledge graph, is the medical knowledge graph and the relevant medical knowledge of disease type t, f is the accuracy score calculation function, when the relevant medical knowledge about disease type t in the knowledge graph is accurately reflected in the disease-related report g, the accuracy score is improved, if there is a contradiction, the accuracy score is reduced; ; in, is the mean accuracy of disease type t, The overall accuracy reported for condition type t.

[0011] As a further improvement of the technical solution, the step of S4 is as follows: S4.1. For low-accuracy disease-related reports, they are divided into multi-type disease-related reports and single-type disease-related reports according to the disease type data involved; S4.2. extracting multiple high-accuracy disease-related reports of different disease types according to the disease types involved in the multi-type disease-related reports, and then combining the extracted multiple high-accuracy disease-related reports of different disease types with the low-accuracy multi-type disease-related reports for mutual correction analysis; S4.3. For single-type disease-related reports, extract high-accuracy disease-related reports of the same type based on the disease type involved and conduct mutual correction analysis.

[0012] A second object of the present invention is to provide an integration system for integrating multi-source medical data sets, including any one of the above-mentioned integration methods for integrating multi-source medical data sets, including a data integration module, an accurate analysis module and a disease model establishment module; The data integration module is used to collect multi-source medical data into the LIMS system, extract historical patient diagnosis symptoms according to patient identification to establish a patient symptom model, and output symptom-related reports through the patient symptom model; The accuracy analysis module is used to extract a medical knowledge graph from multi-source medical data, and then perform accuracy analysis on the disease-related reports in the disease type according to the medical knowledge graph; The disease model building module is used to perform mutual correction analysis on low-accuracy disease-related reports in combination with corresponding accuracy and disease types, integrate the disease-related reports after correction analysis into the LIMS system into a disease diagnosis model, and use the disease diagnosis model to perform disease analysis on patients.

[0013] Compared with the prior art, the present invention has the following beneficial effects: 1. An integrated method and system for integrating multi-source medical data sets. By extracting medical knowledge graphs from multi-source medical data, the accuracy of disease-related reports is analyzed, and corrections are made according to different types of disease-related reports, which greatly improves the accuracy of diagnosis reports. The corrected reports are then integrated into a disease diagnosis model and trained and optimized using machine learning algorithms. The model can learn a large amount of medical knowledge and diagnostic patterns. When analyzing patients' diseases, it can quickly and accurately give diagnostic suggestions and assist doctors in making more scientific decisions.

[0014] 2. An integration method and system for integrating multi-source medical data sets. With patient identification as the core, multi-source medical data are associated to break data silos and form a complete patient diagnosis and treatment data chain. A medical data warehouse is built by combining relational and non-relational databases. It can store structured data and process unstructured data, realize efficient integration and management of data, and facilitate medical staff to fully understand the patient's condition. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is the overall flow chart of the present invention; Figure 2 A flowchart of extracting historical disease types based on historical patient diagnosis diseases in the present invention; Figure 3 A flowchart of the present invention for extracting a medical knowledge graph from multi-source medical data; Figure 4 A flowchart of the present invention is divided into multiple-type disease-related reports and single-type disease-related reports; Figure 5 It is a structural principle diagram of the integrated system of the present invention.

[0016] The meaning of each number in the figure is: 10. Data integration module; 20. Accurate analysis module; 30. Disease model building module. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] like Figure 1 - Figure 5 As shown, one of the purposes of the present invention is to provide an integration method for integrating multi-source medical data sets, comprising the following steps: S1. Collect multi-source medical data into the LIMS system and associate the multi-source medical data according to the patient identification; By connecting various medical data generation equipment, scientific research databases, public health data platforms, and clinical research project data, medical data can be obtained both externally and internally. Then, a medical data warehouse can be built in the LIMS system by combining relational and non-relational databases. Use DICOM protocol to connect CT, MRI and other imaging equipment, connect inspection equipment through RS-232, RS-485 and other serial communication protocols or TCP / IP network protocol, and collect medical data generated by the equipment in real time; For scientific research databases, public health data platforms, and clinical research project data, if SQL queries are supported, you can use the JDBC or ODBC interface to obtain specific data by writing SQL query statements; Use relational databases such as MySQL and Oracle to store structured data, such as basic patient information and test results; Use MongoDB to store unstructured data, such as medical images, text reports, etc.

[0019] By extracting patient data from a medical data warehouse, generating separate patient identifiers for the patient data according to the different owners, and at the same time associating multi-source medical data according to the patient identifiers, the associated medical data is obtained.

[0020] S2. Extracting historical patient diagnosis symptoms according to the patient identifier to establish a patient symptom model, inputting the associated medical data after association according to the historical patient diagnosis symptoms, and outputting a symptom-related report through the patient symptom model; The steps of S2 are as follows: S2.1. Extract the corresponding historical patient diagnosis symptoms from the multi-source medical data according to the patient identification, and then extract the historical symptom type according to the historical patient diagnosis symptoms. Suppose the corresponding relationship between the symptom ID and the symptom type is represented by the mapping function f, that is, , the specific formula is as follows: ; Where T is the set of disease types, is the set of disease types that appear in the patient's historical diagnosis data, t is the disease type in the set T, is the disease type field corresponding to disease type t, D is the patient's historical diagnosis data, d is the data record of multi-source medical data, is the disease field in the data record d, and f is the mapping function from the disease field to the disease type; S2.2. Perform symptom-related numerical analysis on the associated medical data based on the historical patient diagnosis symptoms, set a relevant threshold for symptom relevance based on the patient information, and then filter the associated medical data based on the relevance threshold combined with the symptom-related numerical value. The specific formula is as follows: ; Among them, r is the correlation coefficient, n is the total number of historical patient diagnosis diseases, is the disease status value of the i-th sample disease, which can usually be represented by 0 and 1 to indicate whether a disease exists, 1 for the disease exists, and 0 for the disease does not exist. is the physical diagnostic feature value of the i-th sample disease. For example, when studying the relationship between blood sugar values ​​and a certain disease, It can be the blood sugar test value of the i-th patient's diagnosis of the disease, and are the means of x and y respectively; ; in, is the correlation threshold set, α is the mean correlation coefficient set according to the historical patient diagnosis, The standard deviation of the correlation coefficient was set for the diagnosis of the disease based on historical patients; ; Among them, S is the filtered data set, is the correlation coefficient corresponding to each data point v in the associated medical data set, and K is the original associated medical data; S2.3. Input the screened associated medical data and historical disease types into the patient disease model, and then output the corresponding disease-related report for each disease in the historical disease type through the patient medical record model. The specific formula is as follows: ; in, is the predicted value of the continuous value related to the ith sample disease, is the jth physical diagnosis feature value of the i-th sample disease, is the intercept, is the corresponding coefficient of the jth physical diagnostic feature value, and P is the number of features; The screened associated medical data and historical disease categories are input into the patient disease model for prediction, and a report is generated based on the model prediction results.

[0021] S3. Extract medical knowledge graphs from multi-source medical data, aggregate and associate the disease-related reports of each patient according to the disease type, and then perform accuracy analysis on the disease-related reports in the disease type according to the medical knowledge graph; The steps for S3 are as follows: S3.1. Extract medical knowledge graph from multi-source medical data, and then summarize and associate each patient’s disease-related reports according to disease type; Construct a medical knowledge graph from multi-source medical data (such as electronic medical records, medical literature, clinical research reports, etc.) through technologies such as entity recognition, relationship extraction, and semantic annotation; S3.2. According to the medical knowledge graph, the accuracy of the disease-related reports in the disease type is analyzed to obtain the accuracy of each disease-related report in the disease type. Then, the accuracy mean is set according to the total accuracy of each case type and the case-related report data. The accuracy above the mean is regarded as high accuracy, and vice versa, the accuracy below the mean is regarded as low accuracy. The specific formula is as follows: ; in, is the total disease-related reports of disease type t, E is the total number of patients, For the wth patient, For patients Reports on symptoms of The disease type field in the disease-related report g; ; in, is the accuracy score of the disease-related report g, KG is the medical knowledge graph, is the medical knowledge graph and the relevant medical knowledge of disease type t, f is the accuracy score calculation function, when the relevant medical knowledge about disease type t in the knowledge graph is accurately reflected in the disease-related report g, the accuracy score is improved, if there is a contradiction, the accuracy score is reduced; ; in, is the mean accuracy of disease type t, The overall accuracy of reports related to condition type t; Reports with accuracy greater than or equal to the mean were considered high accuracy reports; Accuracies less than the mean are reported as low accuracy.

[0022] S4. Conduct mutual correction analysis on low-accuracy disease-related reports based on the corresponding accuracy and disease type; The steps of S4 are as follows: S4.1. For low-accuracy disease-related reports, they are divided into multi-type disease-related reports and single-type disease-related reports according to the disease type data involved. The specific steps are as follows: Classify multi-type reports: Check each report in the low-accuracy disease-related report set one by one. For each report, obtain the disease type set involved in this report. If the number of disease types in this disease type set is greater than 1, it means that this report involves multiple disease types. Such reports are classified into the multi-type disease-related report set; Classify single-type reports: Check each report in the low-accuracy symptom-related report set one by one. If the number of symptom types in the symptom type set involved in a report is equal to 1, it means that this report only involves one symptom type, and such reports are classified into the single-type symptom-related report set.

[0023] S4.2. Extract multiple high-accuracy disease-related reports of different disease types according to the disease types involved in the multi-type disease-related reports, and then combine the extracted multiple high-accuracy disease-related reports of different disease types with the low-accuracy multi-type disease-related reports for mutual correction analysis. The specific steps are as follows: Extracting high-accuracy reports: for each multi-type disease-related report in the multi-type disease-related report set, obtain a set of all disease types involved in the report, then check these disease types one by one, and for each disease type, select high-accuracy reports with the same disease type as the currently viewed disease type from all high-accuracy disease-related report sets, and these reports constitute a high-accuracy report subset of the corresponding disease type. Finally, all these high-accuracy report subsets are aggregated to obtain a set of high-accuracy reports of multiple different disease types corresponding to the disease-related reports of this type; Mutual correction analysis: The extracted high-accuracy report sets of multiple different disease types are compared and analyzed with the original low-accuracy multi-type disease-related reports. The key information in the reports, such as symptom descriptions, diagnostic procedures, treatment recommendations, etc., are carefully compared. By analyzing the differences between them, possible errors or inaccuracies in the low-accuracy reports are identified, and corrections are made with reference to the relevant content in the high-accuracy reports. If the diagnostic basis for a certain disease in the high-accuracy report is more detailed and authoritative, and the low-accuracy report is brief or biased in this regard, then the low-accuracy report will be supplemented and corrected based on the high-accuracy report.

[0024] S4.3. For single-type disease-related reports, extract high-accuracy disease-related reports of the same type according to the disease type involved and conduct mutual correction analysis. The specific steps are as follows: Extracting high-accuracy reports of the same type: for each single-type disease-related report in the single-type disease-related report set, obtain the disease type involved in this report, and select high-accuracy reports with the same disease type as that involved in this single-type disease-related report from all high-accuracy disease-related report sets to form a high-accuracy report set of the same type corresponding to the single-type disease-related report; Mutual correction analysis: Compare the same type of high-accuracy report set with the single-type disease-related reports, and also conduct detailed comparisons on key information in the reports, such as symptom manifestations, examination indicators, diagnostic conclusions, etc. If it is found that the single-type low-accuracy report differs from the high-accuracy report in some key information, the high-accuracy report will be used as a reference to revise and improve the single-type low-accuracy report. If the treatment plan in the high-accuracy report has been clinically verified, and there are doubts about the treatment plan in the single-type low-accuracy report, then adjust the treatment plan in the single-type low-accuracy report according to the high-accuracy report.

[0025] S5. Integrate the disease-related reports after the corrective analysis into the LIMS system as a disease diagnosis model, and use the disease diagnosis model to perform disease analysis on the patient.

[0026] The second object of the present invention is to provide an integration system for integrating multi-source medical data sets, including any one of the above-mentioned integration methods for integrating multi-source medical data sets, including a data integration module 10, an accurate analysis module 20 and a disease model establishment module 30; The data integration module 10 is used to collect multi-source medical data into the LIMS system, extract historical patient diagnosis symptoms according to patient identification to establish a patient symptom model, and output symptom-related reports through the patient symptom model; The accuracy analysis module 20 is used to extract a medical knowledge graph from multi-source medical data, and then perform accuracy analysis on the disease-related reports in the disease type according to the medical knowledge graph; The disease model building module 30 is used to perform mutual correction analysis on low-accuracy disease-related reports in combination with corresponding accuracy and disease types, integrate the disease-related reports after correction analysis into the LIMS system into a disease diagnosis model, and use the disease diagnosis model to perform disease analysis on patients.

[0027] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An integration method for integrating multi-source medical data sets, characterized in that: The steps include: S1. Collect multi-source medical data into the LIMS system and associate the multi-source medical data according to the patient identification; S2. Extracting historical patient diagnosis symptoms according to the patient identifier to establish a patient symptom model, inputting the associated medical data after association according to the historical patient diagnosis symptoms, and outputting a symptom-related report through the patient symptom model; S3. Extract medical knowledge graphs from multi-source medical data, aggregate and associate the disease-related reports of each patient according to the disease type, and then perform accuracy analysis on the disease-related reports in the disease type according to the medical knowledge graph; S4. Conduct mutual correction analysis on low-accuracy disease-related reports based on the corresponding accuracy and disease type; S5. Integrate the disease-related reports after the corrective analysis into the LIMS system as a disease diagnosis model, and use the disease diagnosis model to perform disease analysis on the patient.

2. The method for integrating multi-source medical data sets according to claim 1, characterized in that: In S1, medical data is simultaneously acquired externally and internally by connecting various medical data generating devices, scientific research databases, public health data platforms, and clinical research project data, and then a medical data warehouse is constructed in the LIMS system by combining relational databases with non-relational databases.

3. The method for integrating multi-source medical data sets according to claim 1, characterized in that: In S1, patient data is extracted from a medical data warehouse, and then individual patient identifiers are generated for the patient data according to the different owners. At the same time, multi-source medical data are associated according to the patient identifiers to obtain associated medical data.

4. The method for integrating multi-source medical data sets according to claim 1, characterized in that: The steps of S2 are as follows: S2.

1. Extract the corresponding historical patient diagnosis symptoms from the multi-source medical data according to the patient identifier, and then extract the historical symptom type according to the historical patient diagnosis symptoms; S2.2, analyze the relevant values ​​of the symptoms of the associated medical data according to the historical patient diagnosis, set the relevant threshold for the correlation of the symptoms according to the patient information, and then filter the associated medical data by combining the correlation threshold with the relevant values ​​of the symptoms; S2.

3. Input the screened associated medical data and historical disease types into the patient disease model, and then output the corresponding disease-related report for each disease in the historical disease type through the patient medical record model.

5. The method for integrating multi-source medical data sets according to claim 1, characterized in that: The steps of generating a symptom-related report in S2 are as follows: ； Where T is the set of disease types, is the set of disease types that appear in the patient's historical diagnosis data, t is the disease type in the set T, is the disease type field corresponding to disease type t, D is the patient's historical diagnosis data, d is the data record of multi-source medical data, is the disease field in the data record d, and f is the mapping function from the disease field to the disease type; ； Among them, r is the correlation coefficient, n is the total number of historical patient diagnosis diseases, is the disease status value of the i-th sample disease, is the physical diagnostic feature value of the i-th sample disease, and are the means of x and y respectively; ; in, is the correlation threshold set, α is the mean correlation coefficient set according to the historical patient diagnosis, The standard deviation of the correlation coefficient was set for the diagnosis of the disease based on historical patients; ； Among them, S is the filtered data set, is the correlation coefficient corresponding to each data point v in the associated medical data set, and K is the original associated medical data; ； in, is the predicted value of the continuous value related to the ith sample disease, is the jth physical diagnosis feature value of the i-th sample disease, is the intercept, is the corresponding coefficient of the jth physical diagnosis feature value, and P is the number of features.

6. The method for integrating multi-source medical data sets according to claim 1, characterized in that: The steps of S3 are as follows: S3.

1. Extract medical knowledge graph from multi-source medical data, and then summarize and associate each patient’s disease-related reports according to disease type; S3.

2. Perform accuracy analysis on disease-related reports in disease types based on the medical knowledge graph to obtain the accuracy of each disease-related report in the disease type. Then set the accuracy mean based on the total accuracy of each case type and the case-related report data. Values ​​above the mean are considered high accuracy, and vice versa, values ​​below the mean are considered low accuracy.

7. The method for integrating multi-source medical data sets according to claim 1, characterized in that: The steps of determining low-accuracy disease-related reports and high-accuracy disease-related reports in S3 are as follows: ； in, is the total disease-related reports of disease type t, E is the total number of patients, For the wth patient, For patients Reports on symptoms of The disease type field in the disease-related report g; ； in, is the accuracy score of the disease-related report g, KG is the medical knowledge graph, is the medical knowledge graph and the relevant medical knowledge of disease type t, f is the accuracy score calculation function, when the relevant medical knowledge about disease type t in the knowledge graph is accurately reflected in the disease-related report g, the accuracy score is improved, if there is a contradiction, the accuracy score is reduced; ； in, is the mean accuracy of disease type t, The overall accuracy reported for condition type t.

8. The method for integrating multi-source medical data sets according to claim 1, characterized in that: The steps of S4 are as follows: S4.

1. For low-accuracy disease-related reports, they are divided into multi-type disease-related reports and single-type disease-related reports according to the disease type data involved; S4.

2. extracting multiple high-accuracy disease-related reports of different disease types according to the disease types involved in the multi-type disease-related reports, and then combining the extracted multiple high-accuracy disease-related reports of different disease types with the low-accuracy multi-type disease-related reports for mutual correction analysis; S4.

3. For single-type disease-related reports, extract high-accuracy disease-related reports of the same type based on the disease type involved and conduct mutual correction analysis.

9. An integration system for integrating multi-source medical data sets, used to implement an integration method for integrating multi-source medical data sets as claimed in any one of claims 1 to 8, characterized in that: It includes a data integration module (10), an accurate analysis module (20) and a disease model building module (30); The data integration module (10) is used to collect multi-source medical data into the LIMS system, extract historical patient diagnosis symptoms according to patient identification, establish a patient symptom model, and output symptom-related reports through the patient symptom model; The accuracy analysis module (20) is used to extract a medical knowledge graph from multi-source medical data, and then perform accuracy analysis on the disease-related reports in the disease type according to the medical knowledge graph; The disease model building module (30) is used to perform mutual correction analysis on low-accuracy disease-related reports in combination with corresponding accuracy and disease types, integrate the disease-related reports after correction analysis into the LIMS system to form a disease diagnosis model, and use the disease diagnosis model to perform disease analysis on patients.

Images