Medical document automatic generating and filing system and method

Abstract

The invention relates to the technical field of document generation processing, in particular to a medical document automatic generation and archiving system and method.The system comprises an information extraction and structuring module, a document generation and verification module, an abnormal content recognition module and a label generation and archiving module. According to the method, the electronic medical records are analyzed word by word and key information is structured, so that accurate extraction and rapid processing of the medical data are realized, the document generation efficiency is greatly enhanced, the verification process of document continuity and logicality is optimized, the document quality is improved, and through comparative analysis and accurate matching of medicine doses, the accuracy of the medical data is improved. Potential medical errors are effectively recognized, treatment safety is improved, document archiving is systematically performed through an automatic tag generation and data matching technology, document retrieval speed and accuracy are remarkably improved, safe storage of information is ensured, high efficiency of information processing and accuracy of archiving are ensured, and the method is suitable for popularization and application. And the medical document processing cost is obviously reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of document generation and processing, and in particular to a system and method for automatically generating and archiving medical documents. Background Art

[0002] The field of document generation and processing technology mainly involves methods and systems for automatically generating, editing, and managing documents from structured data sources, using advanced algorithms and data processing technologies to convert data into document formats, including the integration of text, charts, and other visual elements. This field combines natural language processing (NLP), machine learning, and artificial intelligence (AI) to simulate the writing and understanding process of human language to generate logical and well-formatted documents. With a wide range of application scenarios, automated document processing not only improves the efficiency of document creation, but also ensures the accuracy and consistency of information.

[0003] Among them, the medical document automatic generation and archiving system focuses on the automated processing of documents in the medical industry. Its main purpose is to automatically extract data from medical databases, electronic medical records or clinical trials, generate medical reports, treatment records, diagnostic documents and other medical-related documents, and store and archive the documents in accordance with the prescribed format and standards. It aims to improve the speed and accuracy of medical document processing, reduce the burden on medical staff in document processing, and ensure the safe storage and easy retrieval of documents, supporting the quality of medical services and management efficiency.

[0004] Existing document generation technologies have not yet fully addressed the speed and accuracy issues of information extraction and processing in the medical industry. In particular, in the conversion process from structured data sources to documents, there is a lack of effective content verification and error monitoring mechanisms, which can easily lead to erroneous information transmission. In addition, existing technologies fail to provide sufficient data linking and security control support, limiting the ability to quickly access and securely manage medical documents, which may hinder medical response speed in emergency situations and increase the security risk of patient information, thereby affecting the quality of medical services and management efficiency. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a medical document automatic generation and archiving system and method.

[0006] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: A medical document automatic generation and archiving system comprises: The information extraction and structuring module analyzes the text word by word based on electronic medical records, identifies and extracts the subject's diagnosis, treatment plan, and medication guide information, sorts the information in field order, and performs format standardization to generate key information structured output results; The document generation and verification module selects a medical document template, combines the key information structured output result, fills the key information into the medical document template, proofreads each filling point, performs a coherence and logic check on the document according to the proofreading result, corrects vocabulary and grammatical errors, and obtains a verified medical document; The abnormal content identification module scans the content of the verified medical document, compares it with the existing medical guidelines, identifies and marks the difference content, checks whether the drug dosage matches the condition according to the difference content identification and marking results, marks and records potential errors, and generates an abnormal content detection report; The label generation and archiving module combines the verified medical document and the abnormal content detection report, adds labels to the document through key information, uses data matching to link with clinical trial information, classifies and archives the document, and performs access control to obtain a record of the completion of medical document archiving.

[0007] As a further solution of the present invention, the steps of obtaining the structured output result of the key information are: Extracting complete text data from the electronic medical record, performing preliminary semantic analysis, identifying key information paragraphs, and generating a preliminary analysis result set; Based on the preliminary analysis result set, performing a named entity recognition operation, extracting the subject's diagnosis information, treatment plan and medication guide, and generating a detailed information set; The detailed information set is used to sort the information in the order of fields and standardize the information to generate a key information structured output result.

[0008] As a further solution of the present invention, the proofreading step of the filling point is: According to medical needs, select the document template of the target medical scenario, reveal the treatment details and patient information, and output the selected medical template; Combining the key information structured output results, filling the subject identification data, diagnosis and treatment content into the designated fields of the selected medical template to generate a filled draft document; Proofread each data item in the filled draft document, check the correctness of the information, compare the consistency between the original data and the template fields, correct errors and omissions, and generate a proofread document.

[0009] As a further solution of the present invention, the steps of obtaining the verified medical document are: Based on the proofread documents, confirm that the document timeline and event sequence are consistent with medical standards, review each paragraph to ensure there are no logical breaks or time errors, and generate coherence and logic review results; Analyze the coherence and logic review results, identify all vocabulary and grammatical errors in the document, check the correctness of spelling, grammar and terminology, adjust the misuse of professional terms, and generate grammar and vocabulary correction reports; According to the grammar and vocabulary correction report, adjust the punctuation and format in the document, confirm the completeness of the document and meet the submission standards, and generate a verified medical document.

[0010] As a further solution of the present invention, the step of identifying and marking the difference content is: Import the verified medical documents, perform text analysis, extract diagnosis, treatment and drug use information, unify the information format, and generate unified medical information; The unified medical information is compared with the existing medical guidelines using the formula: ; Calculate information difference , and obtain the preliminary difference comparison results, among which, represents the guideline value in the database, represents the extracted medical information value, Indicates the number of data points; The preliminary difference comparison results are analyzed, and the difference content is marked. All information points that are inconsistent with the medical guidelines are directly highlighted in the verified medical document to establish the marked difference content.

[0011] As a further solution of the present invention, the steps of obtaining the abnormal content detection report are: According to the marked difference content, the compatibility of the drug dosage and the condition in each medical record is checked one by one, the matching degree of each data is verified, the correlation between the suitability of the drug dosage and the condition is analyzed, and a matching verification result is generated; Marking the inconsistencies and errors in the matching verification results, automatically classifying them according to the error type and severity, and obtaining a marked error list; Integrate the matching verification results and the marked error list, record error details, recommend adjustment measures and optimization suggestions, and compile and generate an abnormal content detection report.

[0012] As a further solution of the present invention, the steps for obtaining the medical document archiving completion record are: Extracting key information from the verified medical document and the abnormal content detection report, recording pathological results and abnormal indicators, assigning medical terminology tags to each document, and generating a document tag set; The document tag set is used for deep matching and docking with the clinical trial database to identify the associated clinical trial information using the formula: ; Calculate the matching degree between labels and clinical trials , generate linked document data, where represents the pathology label matching score of the document, Represents the number of abnormal indicators, represents the relevance distance of the document to the clinical trial; According to the linked document data, classified archiving is performed, and access control rules are set to obtain a medical document archiving completion record.

[0013] A method for automatically generating and archiving medical documents comprises the following steps: S1: Extract complete text data based on electronic medical records, perform preliminary semantic analysis, identify key information paragraphs, perform named entity recognition operations, sort according to field order, standardize the information, and generate key information structured output results; S2: Select a document template for the target medical scenario, combine the key information structured output results, fill the subject identification data, diagnosis and treatment content into the specified fields of the selected medical template, check the correctness of the information, and generate a proofread document; S3: Based on the proofread document, check each paragraph for logical breaks and time errors, identify all vocabulary and grammatical errors in the document, adjust the misuse of professional terms, correct the punctuation and format in the document, and generate a proofread medical document; S4: Import the verified medical document, perform text analysis, unify the information format, calculate the information difference, mark the difference content, and establish the marked difference content; S5: according to the marked difference content, verify the matching degree of each data, analyze the relationship between the appropriateness of the drug dosage and the condition, automatically classify according to the error type and severity, record the error details, recommend adjustment measures and optimization suggestions, and compile and generate an abnormal content detection report; S6: Extract key information from the verified medical document and the abnormal content detection report, assign medical terminology labels to each document, calculate the matching degree between the label and the clinical trial, classify and archive, and set access control rules to obtain a record of the completion of medical document archiving.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by analyzing electronic medical records word by word and structuring key information, accurate extraction and rapid processing of medical data are achieved, the efficiency of document generation is greatly enhanced, the verification process of document consistency and logic is optimized, and the quality of documents is improved. By comparative analysis and accurate matching of drug dosages, potential medical errors are effectively identified and the safety of treatment is improved. Through automated label generation and data matching technology, document archiving is systematically performed, which significantly improves the speed and accuracy of document retrieval, ensures the safe storage of information, ensures the efficiency of information processing and the accuracy of archiving, and significantly reduces the document processing burden of medical personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is a system flow chart of the present invention; Figure 2 A flowchart for obtaining the structured output results of key information of the present invention; Figure 3 It is a proofreading flow chart of the filling point of the present invention; Figure 4 This is a flow chart of obtaining medical documents after verification of the present invention; Figure 5 A flow chart of identification marks for the different contents of the present invention; Figure 6 The following is a flowchart of obtaining an abnormal content detection report of the present invention; Figure 7 A flow chart for obtaining completed records for medical document archiving of the present invention. DETAILED DESCRIPTION

[0016] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0017] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0018] See also Figure 1 , a medical document automatic generation and archiving system includes: The information extraction and structuring module analyzes the text word by word based on electronic medical records, identifies and extracts the subject's diagnosis, treatment plan, and medication guide information, sorts the information in field order, and performs format standardization to generate key information structured output results; The document generation and verification module selects a medical document template, combines the key information structured output results, fills the key information into the medical document template, proofreads each filling point, checks the document's coherence and logic based on the proofreading results, corrects vocabulary and grammatical errors, and obtains the verified medical document; The abnormal content identification module scans the content of the verified medical documents, compares them with the existing medical guidelines, identifies and marks the difference content, checks whether the drug dosage matches the condition based on the difference content identification and marking results, marks and records potential errors, and generates an abnormal content detection report; The label generation and archiving module combines the verified medical documents and abnormal content detection reports, adds labels to the documents through key information, uses data matching to link with clinical trial information, classifies and archives the documents, and performs access control to obtain the completion record of medical document archiving.

[0019] The structured output results of key information include name field, diagnosis field, treatment plan field and medication guide field; the verified medical documents include document integrity check results, document coherence analysis results, logic check results and grammatical accuracy detection records; the abnormal content detection report includes abnormal content tag recognition results and error records; the medical document archiving completion record includes label index and archiving classification results.

[0020] See also Figure 2 , the steps to obtain the structured output results of key information are: Extract complete text data from electronic medical records, perform preliminary semantic analysis, identify key information paragraphs, and generate a preliminary analysis result set; The text data extracted from the electronic medical record system contains the patient's name, age, gender, and detailed medical history. This process first ensures the legal use of the data through access control, and then uses text preprocessing technology to remove irrelevant characters and stop words to improve the quality of data processing. Through natural language processing engines such as word segmentation and syntactic analysis tools, text is structured and pasted text is decomposed into recognizable medical information units, such as symptom descriptions, diagnosis results, and treatment recommendations, thereby generating a preliminary set of data analysis results, laying a data foundation for subsequent in-depth analysis.

[0021] Based on the preliminary analysis result set, named entity recognition is performed to extract the subject's diagnosis information, treatment plan, and medication guide to generate a detailed information set; Based on the preliminary data analysis results, machine learning models, especially named entity recognition technology based on deep learning, are applied to further identify and extract key medical information, such as the specific disease name of the subject, recommended treatment plan, and medication guide. In this step, annotated medical records are used as training sets, and model training is used to identify entities in the text and the relationships between them. The extracted information is organized into a structure that is easy to understand and retrieve, and a detailed information set is generated to ensure the accuracy and availability of the information, thereby supporting subsequent medical decision-making and patient management.

[0022] Use detailed information sets to sort by field order and standardize the information to generate structured output results of key information; For the detailed information sets extracted from medical texts, field sorting and data standardization are adopted to ensure that each data is strictly formatted according to medical standards, such as unifying the date and time into the international standard format, classifying and labeling the diagnosis and treatment information, using data cleaning technology to eliminate redundant and erroneous information, and checking the data consistency through automated scripts and making necessary corrections. The structured output results finally generated not only optimize the storage efficiency of the data, but also improve the accessibility and readability of the data, which provides healthcare providers with an accurate and consistent source of information for clinical decision support and health management.

[0023] See also Figure 3 , the proofreading steps of the filling points are: According to medical needs, select the document template of the target medical scenario, reveal the treatment details and patient information, and output the selected medical template; In the process of selecting the template, we first conduct a medical needs analysis and examine in detail the specific needs of different medical scenarios for data display. Through collaboration with medical staff, we collect necessary functional requirements, such as display of basic patient information, diagnosis history records, treatment effect tracking, etc. According to the needs, we screen out multiple candidate templates in the existing template library, evaluate the data processing capabilities and user interface of each template, and compare how each template supports dynamic filling of data and error prompts. Then we select the final template, which not only supports complex data structures, but also can adapt to possible subsequent template upgrades or customization needs, thereby ensuring the flexibility and accuracy of data display.

[0024] Combined with the key information structured output results, fill the subject identification data, diagnosis and treatment content into the specified fields of the selected medical template to generate a filled draft document; In the process of filling key information into the template, the structured data output results are first matched one by one with the fields of the template through a programmed script. The matching process involves field identification and data format conversion to ensure that all data such as medical diagnosis, treatment plan, etc. are displayed correctly according to the requirements of the template. For specific data items such as medical codes and treatment results, additional formatting operations are performed, such as converting diagnosis codes into readable text descriptions. At the same time, the integrity and consistency of the data after filling are monitored in real time, and data integrity testing and verification are performed to ensure that all required items have been filled in correctly without omissions or errors.

[0025] Proofread every data item in the filled draft document, check the correctness of the information, compare the consistency between the original data and the template fields, correct errors and omissions, and generate a proofread document; Proofreading the filled documents is a process involving multi-dimensional checks. It starts with verifying each piece of data one by one, double-checking each field for logical consistency and format correctness, and using advanced data comparison technology to accurately match the filled data with the original electronic health record. Pay special attention to the timestamp and update status of the data to ensure the timeliness and accuracy of the treatment information. For any differences found, immediately call the source data for correction and take necessary corrective measures, such as updating outdated information or correcting input errors. Finally, confirm that each record meets the standards of medical records and ensure the data integrity and accuracy of the entire document.

[0026] See also Figure 4 , the steps to obtain the medical documents after verification are: Based on the proofread documents, confirm that the document timeline and event sequence are consistent with medical standards, review each paragraph to ensure there are no logical breaks or time errors, and generate coherence and logic review results; When starting to check the continuity and logic of the document, first conduct a structural review of the entire document to determine whether each part of the document, such as the description of the condition, treatment measures, doctor's observations and patient reactions, is arranged in a reasonable chronological order according to the medical service to ensure the accuracy of the timeline. Then compare the patient's diagnosis record and treatment results section by section, and find and mark the logical inconsistencies or gaps in the description. For example, a treatment process description begins without mentioning the specific diagnosis results. In this case, it is necessary to trace back the document to find out whether there are omissions or recording errors. Through this meticulous inspection process, a logical and continuity review report of the document is generated.

[0027] Analyze the results of the coherence and logic review, identify all vocabulary and grammatical errors in the document, check the correctness of spelling, grammar and terminology, adjust the misuse of professional terms, and generate grammar and vocabulary correction reports; After completing the initial check for coherence and logic, the document is further analyzed in depth for vocabulary and grammar, focusing on whether the terminology used is standard from a medical professional perspective, such as whether the occasions in which "heart attack" and "myocardial infarction" are used are accurate, and the consistency of these terms in the document. At the same time, the document's grammatical structure is checked to ensure that the sentences are constructed correctly and punctuation is used appropriately to avoid difficulty in understanding due to the complexity of long sentences. This step involves precise proofreading of each diagnosis and treatment description to ensure the accurate communication of medical information, correct any errors found, and record them in a grammar and vocabulary correction report.

[0028] Adjust punctuation and formatting in documents based on grammar and vocabulary revision reports, confirm document integrity and compliance with submission standards, and generate verified medical documents; According to the grammar and vocabulary revision report, the final document review is carried out. This stage mainly confirms that all parts of the document have been correctly revised in accordance with medical standards, without missing information or unresolved errors, especially in the format and style of the document. For example, ensure that all medical terms are spelled consistently, the format of dates and numbers is unified, and the titles and subtitles in the document meet the format requirements of medical documents. Through the adjustment of details, the professionalism and accuracy of the document are guaranteed, and the final verified medical document is generated to complete the document standardization process.

[0029] See also Figure 5 , the steps for identifying and marking the difference content are: Import verified medical documents, perform text analysis, extract diagnosis, treatment and drug use information, unify information formats, and generate unified medical information; In the process of importing verified medical documents, we first use text parsing technology to extract key medical information, including diagnosis, treatment and drug use. Through detailed analysis of the original medical records, we ensure that the extracted data is of high quality and criticality. Then, we structure the information based on text parsing tools to ensure the organization and operability of the data. During the structuring process, the attributes and values ​​of each type of information are recorded in detail to make the data more standardized and facilitate subsequent data processing and storage. This process not only improves the data processing efficiency, but also provides preparation for subsequent comparative analysis, generates structured medical information, and makes full preparation for the next step of comparative analysis.

[0030] Compare the content of the unified medical information with the existing medical guidelines using the formula: ; Calculate information difference , and obtain the preliminary difference comparison results, among which, represents the guideline value in the database, represents the extracted medical information value, Indicates the number of data points; According to the verified medical documents, the patient's actual blood glucose level was 9.5mmol / L and the daily insulin injection dose was 40 units.

[0031] For patients with the same diabetes symptoms, the standard blood sugar target recommended by medical guidelines is to keep it below 7.0mmol / L and the daily insulin dose is 30 units.

[0032] After data transformation, we get: Extracted information values: blood glucose level 9.5mmol / L, insulin dose 40 units.

[0033] Values ​​in the guidelines: blood glucose target 7.0mmol / L, insulin dose 30 units.

[0034] Right now: Blood sugar level , ; Insulin dosage , ; Compute the dot product: ; calculate The sum of squares: ; calculate The sum of squares: ; Calculate the modulus length: ; Final calculation: ; Calculation results A very small negative difference is shown. Theoretically, the difference should be between 0 and 1, where 0 indicates complete agreement and 1 indicates complete disagreement. Negative values ​​are caused by rounding errors and should actually be interpreted as being very close to 0, indicating that the patient's treatment plan is very close to the medical guidelines, that is, the patient's current treatment status is almost consistent with the recommended medical guidelines. This confirms that the patient's current treatment plan is very close to the standard recommended by the medical guidelines.

[0035] Analyze the preliminary difference comparison results, mark the difference content, directly highlight all information points that do not conform to the medical guidelines in the verified medical documents, and establish the marked difference content; The preliminary difference comparison results are analyzed in detail. The specific operations involved include using a marking algorithm to identify the difference points in the comparison results. Each difference point is recorded and analyzed in detail. During this analysis process, a highlighted marking method is used to visualize information that is inconsistent with medical guidelines. The corresponding marks are automatically generated in the document through the program. The marks help medical personnel and clinical trial personnel quickly identify the difference information that needs attention, and provide support for subsequent diagnosis and treatment decisions. The marked difference content is accurately located and described to ensure the accuracy and efficiency of medical document review.

[0036] See also Figure 6 , the steps to obtain the abnormal content detection report are: According to the marked differences, the compatibility of the drug dosage and the condition in each medical record is checked one by one, the matching degree of each data is verified, the correlation between the appropriateness of the drug dosage and the condition is analyzed, and the matching verification results are generated; In the process of applying the matching verification algorithm to analyze the marked difference content, the marked difference content is first imported, including the description of the condition and the corresponding drug dosage information. The matching degree of each data is checked by the algorithm to ensure the compatibility of the drug dosage with the condition description. The algorithm processes each record one by one, marks the unmatched records, and generates detailed matching verification results. The process involves calling and comparing multiple parameters, such as the severity of the condition, the recommended drug dosage parameters, etc. Each parameter is precisely calculated to determine its adaptability to the condition, providing clear guidance for medical personnel, indicating which drug dosages need to be adjusted.

[0037] Mark the inconsistencies and errors in the matching verification results, automatically classify them according to the error type and severity, and obtain a marked error list; In the process of using the error marking system to mark in detail the inconsistencies or errors in the matching verification results, the matching verification results from the previous step are first received, and then automatically classified and marked according to the error type and severity. The automated process uses advanced classification algorithms, such as decision trees or logistic regression, to identify error patterns and classify them. Each error is assigned a severity mark so that medical staff can prioritize the most serious problems. The system also records the specific location of the error and the recommended corrective measures. The generated error list not only records all identified problems in detail, but also includes specific instructions and suggestions for each problem, providing important decision support for the medical team.

[0038] Integrate the matching verification results and the marked error list, record the error details, recommend adjustment measures and optimization suggestions, and compile and generate abnormal content detection reports; The process of compiling an abnormal content detection report involves integrating all marked errors and verification results into a detailed document, automatically summarizing all data and conclusions, including detailed descriptions of the errors, recommended adjustment measures, and optimization suggestions. During the report generation process, each error and suggestion is systematically calculated and analyzed in detail to ensure that the information provided to the medical team is accurate and practical. The report will finally generate a comprehensive document, which is not only convenient for the medical team to review, but also convenient for archiving and tracking various problems during the treatment process, ensuring that all important information is properly managed and utilized.

[0039] See also Figure 7 , the steps to obtain the completed records of medical document archiving are: Extract key information from verified medical documents and abnormal content detection reports, record pathological results and abnormal indicators, assign medical terminology tags to each document, and generate document tag sets; Extract key information from verified medical documents and abnormal content detection reports, such as pathological results and abnormal indicators. This process involves screening out clinically significant feature information from a large amount of document data. First, natural language processing is performed on the text data in the document to extract key words such as "cancer" and "hypertension" and other pathological terms. Subsequently, data mining technology is used to identify abnormal indicators, such as abnormally high white blood cell counts or abnormally low hemoglobin values. The indicators are obtained by analyzing and comparing with the known range of health indicators. Through this process, each document is annotated with specific medical terminology tags for subsequent processing. Medical terminology tags provide a basis for further analysis of documents and data matching.

[0040] Use document tag sets for deep matching, connect with clinical trial database, identify related clinical trial information, and use the formula: ; Calculate the matching degree between labels and clinical trials , generate linked document data, where represents the pathology label matching score of the document, Represents the number of abnormal indicators, represents the relevance distance of the document to the clinical trial; The pathology label matching score of the document can be obtained by counting the occurrence frequency of pathology terms in the document and the number of matching times of pathology conditions required by clinical trials. Assume that a document has 5 records matching the pathology conditions of a specific clinical trial. ; Represents the number of abnormal indicators. If a document shows that there are 3 abnormal indicators related to clinical trials, ; Represents the relevance distance between the document and the clinical trial, which is the time difference between the date the document is recorded and the start or end date of the clinical trial (from 1 to 5, indicating that the distance is getting farther). If the time distance between the document and the clinical trial is relatively close, it is assumed that . Substituting the values ​​into the formula: ; The results show that the document matches the clinical trial at a degree of 8.66. The higher this value is, the better the match between the document and the clinical trial is, thereby determining the relevance and importance of the document to the clinical trial.

[0041] According to the linked document data, classified archiving is performed, and access control rules are set to obtain the completion record of medical document archiving; Classification and archiving are performed based on the linked document data. This process involves appropriate classification and storage of documents with high matching degrees. First, all documents are sorted according to their matching degree values. Then, the documents are classified according to the predetermined archiving rules. For example, documents with a matching degree higher than 8 are stored in the "high priority" folder. At the same time, specific access control rules are set, such as only allowing researchers with specific qualifications to access documents in the "high priority" folder. Through this precise management process, the safe storage and efficient management of medical documents are ultimately ensured, providing support for clinical research and ensuring the accessibility and compliance of documents.

[0042] A method for automatically generating and archiving medical documents comprises the following steps: S1: Extract complete text data based on electronic medical records, perform preliminary semantic analysis, identify key information paragraphs, perform named entity recognition operations, sort according to field order, standardize the information, and generate key information structured output results; S2: Select the document template of the target medical scenario, combine the key information structured output results, fill the subject identification data, diagnosis and treatment content into the specified fields of the selected medical template, check the correctness of the information, and generate a proofread document; S3: Based on the proofread document, check each paragraph for logical breaks and time errors, identify all vocabulary and grammatical errors in the document, adjust the misuse of professional terms, correct the punctuation and format in the document, and generate a proofread medical document; S4: Import the verified medical documents, perform text analysis, unify the information format, calculate the information difference, mark the difference content, and establish the marked difference content; S5: Based on the marked difference content, verify the matching degree of each data, analyze the relationship between the appropriateness of the drug dosage and the condition, automatically classify according to the error type and severity, record the error details, recommend adjustment measures and optimization suggestions, and compile and generate an abnormal content detection report; S6: Extract key information from the verified medical documents and abnormal content detection reports, assign medical terminology labels to each document, calculate the matching degree between the labels and clinical trials, classify and archive them, and set access control rules to obtain the completion record of medical document archiving.

[0043] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A medical document automatic generation and archiving system, characterized by: The system comprises: The information extraction and structuring module analyzes the text word by word based on electronic medical records, identifies and extracts the subject's diagnosis, treatment plan, and medication guide information, sorts the information in field order, and performs format standardization to generate key information structured output results; The document generation and verification module selects a medical document template, combines the key information structured output result, fills the key information into the medical document template, proofreads each filling point, performs a coherence and logic check on the document according to the proofreading result, corrects vocabulary and grammatical errors, and obtains a verified medical document; The abnormal content identification module scans the content of the verified medical document, compares it with the existing medical guidelines, identifies and marks the difference content, checks whether the drug dosage matches the condition according to the difference content identification and marking results, marks and records potential errors, and generates an abnormal content detection report; The label generation and archiving module combines the verified medical document and the abnormal content detection report, adds labels to the document through key information, uses data matching to link with clinical trial information, classifies and archives the document, and performs access control to obtain a record of the completion of medical document archiving.

2. The medical document automatic generation and archiving system according to claim 1, characterized in that: The steps for obtaining the structured output results of the key information are: Extracting complete text data from the electronic medical record, performing preliminary semantic analysis, identifying key information paragraphs, and generating a preliminary analysis result set; Based on the preliminary analysis result set, performing a named entity recognition operation, extracting the subject's diagnosis information, treatment plan and medication guide, and generating a detailed information set; The detailed information set is used to sort the information in the order of fields and standardize the information to generate a key information structured output result.

3. The medical document automatic generation and archiving system according to claim 2, characterized in that: The proofreading steps of the filling point are: According to medical needs, select the document template of the target medical scenario, reveal the treatment details and subject information, and output the selected medical template; Combining the key information structured output results, filling the subject's diagnosis and treatment content into the designated fields of the selected medical template to generate a filled draft document; Proofread each data item in the filled draft document, check the correctness of the information, compare the consistency between the original data and the template fields, correct errors and omissions, and generate a proofread document.

4. The medical document automatic generation and archiving system according to claim 3, characterized in that: The steps for obtaining the verified medical document are: Based on the proofread documents, confirm that the document timeline and event sequence are consistent with medical standards, review each paragraph to ensure there are no logical breaks or time errors, and generate coherence and logic review results; Analyze the coherence and logic review results, identify all vocabulary and grammatical errors in the document, check the correctness of spelling, grammar and terminology, adjust the misuse of professional terms, and generate grammar and vocabulary correction reports; According to the grammar and vocabulary correction report, adjust the punctuation and format in the document, confirm the completeness of the document and meet the submission standards, and generate a verified medical document.

5. The medical document automatic generation and archiving system according to claim 4, characterized in that: The steps of identifying and marking the difference content are as follows: Import the verified medical documents, perform text analysis, extract diagnosis, treatment and drug use information, unify the information format, and generate unified medical information; The unified medical information is compared with the existing medical guidelines using the formula: ； Calculate information difference , and obtain the preliminary difference comparison results, among which, represents the guideline value in the database, represents the extracted medical information value, Indicates the number of data points; The preliminary difference comparison results are analyzed, and the difference content is marked. All information points that are inconsistent with the medical guidelines are directly highlighted in the verified medical document to establish the marked difference content.

6. The medical document automatic generation and archiving system according to claim 5, characterized in that: The steps for obtaining the abnormal content detection report are: According to the marked difference content, the compatibility of the drug dosage and the condition in each medical record is checked one by one, the matching degree of each data is verified, the correlation between the suitability of the drug dosage and the condition is analyzed, and a matching verification result is generated; Marking the inconsistencies and errors in the matching verification results, automatically classifying them according to the error type and severity, and obtaining a marked error list; Integrate the matching verification results and the marked error list, record error details, recommend adjustment measures and optimization suggestions, and compile and generate an abnormal content detection report.

7. The medical document automatic generation and archiving system according to claim 6, characterized in that: The steps for obtaining the medical document filing completion record are: Extracting key information from the verified medical document and the abnormal content detection report, recording pathological results and abnormal indicators, assigning medical terminology tags to each document, and generating a document tag set; The document tag set is used for deep matching and docking with the clinical trial database to identify the associated clinical trial information using the formula: ； Calculate the matching degree between labels and clinical trials , generate linked document data, where represents the pathology label matching score of the document, Represents the number of abnormal indicators, represents the relevance distance of the document to the clinical trial; According to the linked document data, classified archiving is performed, and access control rules are set to obtain a medical document archiving completion record.

8. A method for automatically generating and archiving medical documents, characterized in that: The system according to any one of claims 1 to 7 is implemented, comprising the following steps: Extract complete text data based on electronic medical records, conduct preliminary semantic analysis, identify key information paragraphs, perform named entity recognition operations, sort according to field order, and standardize the information to generate key information structured output results; Select a document template for the target medical scenario, combine the key information structured output results, fill the subject identification data, diagnosis and treatment content into the specified fields of the selected medical template, check the correctness of the information, and generate a proofread document; Based on the proofread document, check each paragraph for logical breaks and time errors, identify all vocabulary and grammatical errors in the document, adjust the misuse of professional terms, correct the punctuation and format in the document, and generate a verified medical document; Import the verified medical documents, perform text analysis, unify the information format, calculate the information difference, mark the difference content, and establish the marked difference content; According to the marked difference content, verify the matching degree of each data, analyze the relationship between the appropriateness of the drug dosage and the condition, automatically classify according to the error type and severity, record the error details, recommend adjustment measures and optimization suggestions, and compile and generate an abnormal content detection report; Extract key information from the verified medical documents and the abnormal content detection report, assign medical terminology tags to each document, calculate the matching degree between the tags and clinical trials, classify and archive them, set access control rules, and obtain a record of the completion of medical document archiving.

Images