Medical information integration processing method

Abstract

The invention relates to the technical field of information processing, and discloses a medical information integration processing method comprising the following steps: 1, receiving medical information; 2, analyzing the medical information; step 3, secondary re-checking is carried out; step 4, generating review form information; 5, integrating examination form information; and step 6, generating medical display information. Data from different systems are converted into a unified internal format by using a data adapter or a converter, the compatibility of the data is ensured, then the accuracy of information extraction is improved through a machine learning model and a natural language processing technology, and secondary rechecking is performed by using a verification rule and a data integrity check algorithm, so that the accuracy of information extraction is improved. The method ensures the accuracy and integrity of data, and achieves the beneficial effects of compatibility of information data of an electronic medical record system, an image system and a laboratory during data integration, and avoidance of data loss.

Description

Technical Field

[0001] The present invention relates to the technical field of information processing, and in particular to a method for integrated processing of medical information. Background Art

[0002] Rational drug use review refers to the process of automatically collecting and analyzing drug-related information through the system during clinical diagnosis and treatment, comprehensively judging the rationality of drug use and reminding of drug use problems. For outpatient prescriptions, there is less information and the form of review is relatively clear; while a large amount of information is generated every day during a hospitalization of an inpatient, including drug order information, test information, surgery information, vital signs information, etc. The rationality review of inpatient orders has the characteristics of many drug varieties, many factors affecting drug use, and long diagnosis and treatment cycle. Therefore, the review of inpatient orders is more complicated than the review of outpatient prescriptions. At present, there is a lack of effective review methods for inpatient drugs; and the prior art publication number is CN 106469255 A, a method for integrating and processing medical information, the method includes: receiving medical information; parsing the medical information, obtaining generation time information and medical content information, the medical content information includes event medical information and continuation medical information; according to the time information, generating review form information from the medical content information; integrating and processing the review form to generate integrated review form information; using the integrated review form information to generate medical display information, which effectively solves the problem of medical information integration.

[0003] Although the above-mentioned technical solution generates integrated information through integrated expression processing, since the form contains different systems or equipment, such as electronic medical record systems, imaging systems, and laboratory information systems, the data formats, standards, and quality of these systems are inconsistent, resulting in data incompatibility or data loss during the integration process. At the same time, when parsing medical information, it is necessary to accurately extract the generation time information and medical content information. If there are missing errors in the above-mentioned data, the extracted information will be further inaccurate. Summary of the invention

[0004] Technical issues solved: In view of the deficiencies of the prior art, the present invention provides a method for integrating and processing medical information, which has the advantages of making the electronic medical record system, imaging system, and laboratory information data compatible when integrating data to avoid data loss; and performing a secondary review after extracting the medical content information to ensure the accuracy of the information content, thereby solving the problems of the above-mentioned technology.

[0005] Technical solution: To achieve the above object, the present invention provides the following technical solution: a method for integrating and processing medical information, comprising the following steps: Step 1: Receive medical information: Use data adapters or converters to convert data from different sources such as electronic medical record systems, imaging systems, and laboratory information systems into a unified internal format; Step 2: Analyze medical information: Use machine learning models and natural language processing technology to improve the accuracy of information extraction and the ability to handle abnormal data; Step 3: Secondary review: Introduce a secondary review mechanism to recheck and verify the extracted medical content information, using verification rules and data integrity check algorithms to automatically detect and correct errors in the extracted information; Step 4: Generate review form information: Generate the review form in a templated and structured manner; Step 5: Integrate the review form information: Integrate the review form, including supplementary diagnosis, revised diagnosis, setting of test index validity and judgment of medical order type; Step 6. Generate medical display information: Use the integrated review form information to generate intuitive medical display information, including icons, graphs, and lists.

[0006] Preferably, in order to ensure the accuracy and efficiency of data conversion in step 1, n different data sources are provided, and the data format of each data source is Indicates that the target internal format is It means that the conversion function To convert the data from each data source into the internal format: The data source is represented as follows: Assume that the data format of each data source is A vector containing multiple fields , No. The data of each data source is represented as:

[0007] Using conversion functions , will Data from the data source Convert to internal format , the conversion function is expressed as: .

[0008] Preferably, the step 1 further includes a unified data format, and the unified data format is calculated as: Internal format Is a vector, containing multiple fields The internal format is:

[0009] The conversion formula is:

[0010] in, Is the first Data source The field is converted to the internal format Function of a field.

[0011] Preferably, the text classification calculation in step 2 is as follows: there are n text samples, each sample xi belongs to a certain category yi, and a machine learning model is used to learn the mapping from text to category:

[0012] in, The model predicts The probability that a sample belongs to the positive class; The machine learning also includes an optimization target, which is calculated as:

[0013] in, are model parameters, is the regularization coefficient.

[0014] Preferably, the step 2 further includes a named entity recognition calculation, which is used to identify key entities in the text, such as diseases, drugs and examinations to improve accuracy, and the specific calculation uses a BiLSTM-CRF model: BiLSTM layer:

[0015] CRF layer:

[0016] in, is the transfer matrix, is the emission matrix, is the normalization factor.

[0017] Preferably, the verification rule in step 3 is calculated as:

[0018] in, and are the upper and lower limits of the value.

[0019] Preferably, in order to generate a structured and standardized review form in step 4, a template engine and data filling technology are used to define a form template, and the extracted and reviewed medical information is filled into the corresponding fields: The template is filled with: Set up form template , which contains several placeholders , the corresponding filling data is ; .

[0020] Preferably, before filling the data in step 4, the correctness and completeness of the data need to be verified, and the data verification calculation is:

[0021] The overall verification is:

[0022] in: .

[0023] Preferably, the supplementary diagnosis calculation in step 5 is: Set diagnostic set , the diagnostic set that needs to be supplemented is :

[0024] Revised diagnosis: Set up diagnosis Need to be corrected to : .

[0025] Preferably, the icon in step six is ​​generated as follows: Set multiple categories of information , each category There is a corresponding number ; .

[0026] Compared with the prior art, the present invention provides a method for integrating and processing medical information, which has the following beneficial effects: 1. The present invention converts data from different systems into a unified internal format by using a data adapter or converter, thereby ensuring data compatibility. Then, the accuracy of information extraction is improved through machine learning models and natural language processing technology, and verification rules and data integrity check algorithms are used for secondary review to ensure data accuracy and completeness. On this basis, a templated and structured approach is used to generate review forms, and data verification is used to ensure that the data in each field meets the requirements. When integrating the review form information, the diagnostic information is supplemented and corrected to ensure that all relevant information is correctly processed and recorded, thereby achieving the beneficial effect of making the electronic medical record system, imaging system, and laboratory information data compatible when integrating data and avoiding data loss.

[0027] 2. The present invention ensures the accuracy of the extracted medical content information by introducing a secondary review mechanism. Specifically, in step two, machine learning models and natural language processing technology are used to parse medical information to improve the accuracy of information extraction and the ability to handle abnormal data. On this basis, the secondary review link in step three uses verification rules and data integrity check algorithms to re-check and verify the extracted information, and automatically detect and correct errors therein. This double safeguard significantly improves the accuracy of the information, avoids the problem of inaccurate information caused by deviations or errors in the initial extraction process, and achieves the beneficial effect of secondary review after extracting medical content information to ensure the accuracy of the information content. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a schematic diagram of the steps of the method of the present invention; DETAILED DESCRIPTION

[0029] 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.

[0030] See also Figure 1 , a method for integrating and processing medical information, comprising the following steps: Step 1: Receive medical information: Use data adapters or converters to convert data from different sources such as electronic medical record systems, imaging systems, and laboratory information systems into a unified internal format; Step 2: Analyze medical information: Use machine learning models and natural language processing technology to improve the accuracy of information extraction and the ability to handle abnormal data; Step 3: Secondary review: Introduce a secondary review mechanism to recheck and verify the extracted medical content information, using verification rules and data integrity check algorithms to automatically detect and correct errors in the extracted information; Step 4: Generate review form information: Generate the review form in a templated and structured manner; Step 5: Integrate the review form information: Integrate the review form, including supplementary diagnosis, revised diagnosis, setting of test index validity and judgment of medical order type; Step 6. Generate medical display information: Use the integrated review form information to generate intuitive medical display information, including icons, graphs, and lists.

[0031] Data compatibility: By using data adapters or converters, data from different systems (such as electronic medical records, imaging systems, and laboratory information systems) can be converted into a unified internal format, achieving data compatibility and integration.

[0032] Information extraction accuracy: The use of machine learning models and natural language processing technology improves the accuracy of information extraction from unstructured and semi-structured data, especially for the processing of text data.

[0033] Quality control: A secondary review mechanism is introduced to automatically detect and correct errors in extracted information through verification rules and data integrity check algorithms to ensure the accuracy and reliability of the data.

[0034] Structured and standardized: The generated review forms are structured and standardized to facilitate further data analysis and processing, while also complying with the standards and specifications of the medical industry.

[0035] Intelligent processing: When integrating the review form information, it can automatically perform supplementary diagnosis, correct diagnosis, set the effectiveness of test indicators and judge the type of medical order, thus improving work efficiency and accuracy.

[0036] Intuitive display: The generated medical display information is in the form of icons, graphs, and lists, allowing medical staff to understand and analyze medical data more intuitively and facilitate clinical decision-making Specifically, in order to ensure the accuracy and efficiency of data conversion in step 1, n different data sources are set, and the data format of each data source is Indicates that the target internal format is It means that the conversion function To convert the data from each data source into the internal format: The data source is represented as follows: Assume that the data format of each data source is A vector containing multiple fields , No. The data of each data source is represented as:

[0037] Using conversion functions , will Data from the data source Convert to internal format , the conversion function is expressed as: .

[0038] Specifically, step one also includes unifying the data format, and the unified data format is calculated as follows: Internal format Is a vector, containing multiple fields The internal format is:

[0039] The conversion formula is:

[0040] in, Is the first Data source The field is converted to the internal format Function of a field.

[0041] Specifically, the text classification calculation in step 2 is as follows: suppose there are n text samples, each sample xi belongs to a certain category yi, and a machine learning model is used to learn the mapping from text to category:

[0042] in, The model predicts The probability that a sample belongs to the positive class; Machine learning also includes optimization objectives, which are calculated as:

[0043] in, are model parameters, is the regularization coefficient.

[0044] Specifically, step 2 also includes named entity recognition calculations, which are used to identify key entities in the text, such as diseases, drugs, and examinations to improve accuracy. The specific calculation uses the BiLSTM-CRF model: BiLSTM layer:

[0045] CRF layer:

[0046] in, is the transfer matrix, is the emission matrix, is the normalization factor.

[0047] Specifically, the verification rule calculation in step 3 is:

[0048] in, and are the upper and lower limits of the value.

[0049] Specifically, in step 4, in order to generate a structured and standardized review form, a template engine and data filling technology are used to define a form template, and the extracted and reviewed medical information is filled into the corresponding fields: The template is filled with: Set up form template , which contains several placeholders , the corresponding filling data is ; .

[0050] Specifically, in step 4, before filling in the data, the correctness and completeness of the data need to be verified. The data verification calculation is:

[0051] The overall verification is:

[0052] in: .

[0053] Specifically, the supplementary diagnosis calculation in step 5 is: Set diagnostic set , the diagnostic set that needs to be supplemented is :

[0054] Revised diagnosis: Set up diagnosis Need to be corrected to : .

[0055] Specifically, the icon generated in step 6 is: Set multiple categories of information , each category There is a corresponding number ; .

[0056] By using data adapters or converters to convert data from different systems into a unified internal format, data compatibility is ensured. Then, machine learning models and natural language processing technology are used to improve the accuracy of information extraction, and verification rules and data integrity check algorithms are used for secondary review to ensure the accuracy and completeness of the data. On this basis, a templated and structured approach is used to generate review forms, and data verification is used to ensure that the data in each field meets the requirements. When integrating the review form information, the diagnostic information is supplemented and corrected to ensure that all relevant information is correctly processed and recorded, so that the electronic medical record system, imaging system, and laboratory information data are compatible when integrating data. The accuracy of the extracted medical content information is ensured by introducing a secondary review mechanism. Specifically, in step 2, machine learning models and natural language processing technology are used to parse medical information to improve the accuracy of information extraction and the ability to handle abnormal data. On this basis, in the secondary review link in step 3, verification rules and data integrity check algorithms are used to re-check and verify the extracted information, and errors are automatically detected and corrected. This double guarantee measure significantly improves the accuracy of information and avoids inaccurate information caused by deviations or errors in the initial extraction process.

[0057] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for integrating and processing medical information, characterized in that: The following steps are involved: Step 1: Receive medical information: Use data adapters or converters to convert data from different sources such as electronic medical record systems, imaging systems, and laboratory information systems into a unified internal format; Step 2: Analyze medical information: Use machine learning models and natural language processing technology to improve the accuracy of information extraction and the ability to handle abnormal data; Step 3: Secondary review: Introduce a secondary review mechanism to recheck and verify the extracted medical content information, using verification rules and data integrity check algorithms to automatically detect and correct errors in the extracted information; Step 4: Generate review form information: Generate the review form in a templated and structured manner; Step 5: Integrate the review form information: Integrate the review form, including supplementary diagnosis, revised diagnosis, setting of test index validity and judgment of medical order type; Step 6. Generate medical display information: Use the integrated review form information to generate intuitive medical display information, including icons, graphs, and lists.

2. A method for integrating and processing medical information according to claim 1, characterized in that: In order to ensure the accuracy and efficiency of data conversion in step 1, n different data sources are provided, and the data format of each data source is Indicates that the target internal format is It means that the conversion function To convert the data from each data source into the internal format: The data source is represented as follows: Assume that the data format of each data source is A vector containing multiple fields , No. The data of each data source is represented as: Using conversion functions , will Data from the data source Convert to internal format , the conversion function is expressed as: .

3. A method for integrating and processing medical information according to claim 1, characterized in that: The step 1 also includes a unified data format, and the unified data format is calculated as: Internal format Is a vector, containing multiple fields The internal format is: The conversion formula is: in, Is the first Data source The field is converted to the internal format Function of a field.

4. The method for integrating and processing medical information according to claim 1, characterized in that: The text classification calculation in step 2 is as follows: suppose there are n text samples, each sample xi belongs to a certain category yi, and a machine learning model is used to learn the mapping from text to category: in, The model predicts The probability that a sample belongs to the positive class; The machine learning also includes an optimization target, which is calculated as: in, are model parameters, is the regularization coefficient.

5. The method for integrating and processing medical information according to claim 1, characterized in that: The step 2 also includes named entity recognition calculation, which is used to identify key entities in the text, such as diseases, drugs, and examinations to improve accuracy. The specific calculation uses the BiLSTM-CRF model: BiLSTM layer: CRF layer: in, is the transfer matrix, is the emission matrix, is the normalization factor.

6. A method for integrating and processing medical information according to claim 1, characterized in that: The verification rule in step 3 is calculated as: in, and are the upper and lower limits of the value.

7. A method for integrating and processing medical information according to claim 1, characterized in that: In step 4, in order to generate a structured and standardized review form, a form template is defined using a template engine and data filling technology, and the extracted and reviewed medical information is filled into the corresponding fields: The template is filled with: Set up form template , which contains several placeholders , the corresponding filling data is ; 。 8. The method for integrating and processing medical information according to claim 1, characterized in that: In step 4, before filling the data, the correctness and completeness of the data need to be verified. The data verification calculation is: The overall verification is: in: .

9. The method for integrating and processing medical information according to claim 1, characterized in that: The supplementary diagnosis calculation in step 5 is: Set diagnostic set , the diagnostic set that needs to be supplemented is : Revised diagnosis: Set up diagnosis Need to be corrected to : 。 10. The method for integrating and processing medical information according to claim 1, characterized in that: The icon generated in step 6 is: Set multiple categories of information , each category There is a corresponding number ; 。

Images