Management system for intelligently capturing fracture records and capturing method thereof

Through the intelligent capture of fracture record management system, the special disease database and intelligent relationship system model are used, combined with TF-IDF and KMP algorithms, the problem of low efficiency in querying bone scabbar fracture information is solved, and efficient and accurate information extraction and management are achieved. , reducing costs and misjudgment risks.

CN118899093AInactive Publication Date: 2024-11-05GENERAL HOSPITAL OF NUCLEAR IND
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202411398825.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-09
Publication Date
2024-11-05
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

In the prevention and treatment of bone loose fractures and refractory, the existing technology is difficult to extract fracture information from massive medical data efficiently and accurately, which makes information inquiry time and effort, and it is easy to cause missed judgments or misjudgments due to human factors, affecting diagnosis and treatment.

Method used

Adopt a management system that intelligently captures fracture records, and builds a special disease database and an intelligent relational system model, and uses TF-IDF algorithm and KMP matching algorithm to screen and match fracture records from the medical business data center, combining manual review and data cleaning to achieve Dynamic updates and optimization of keyword databases.

Benefits of technology

It improves the efficiency of obtaining specialized disease information, reduces time and labor costs, avoids misjudgment and misjudgment, integrates it into the existing medical information system to have good compatibility and scalability, and supports complex and diverse fracture types.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118899093A_ABST
    Figure CN118899093A_ABST
Patent Text Reader

Abstract

The invention relates to a management system for intelligently capturing fracture records and a capturing method thereof. The capturing method of the management system for intelligently capturing the fracture records comprises the following steps that a special disease database is constructed in advance, and the special disease database comprises special disease types and a special disease keyword library; constructing a special disease intelligent relation system model by taking the special disease database as an information source, and capturing treatment records conforming to keywords in the special disease database from the business data set through the special disease intelligent relation system model; and marking or displaying the grabbed doctor-seeing records according to use requirements, marking the doctor-seeing records screened out from the business data set as special disease records, and performing marking or classified display in a medical management system. According to the management system for intelligently capturing the fracture records and the capturing method thereof disclosed by the invention, through recording and capturing, the efficiency of acquiring special disease information is improved, and the time and labor cost is reduced; and missed judgment and misjudgment are avoided, and good compatibility and expansibility are achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a management system for intelligently capturing fracture records and a capturing method thereof. Background Art

[0002] In the prevention and treatment of osteoporotic fracture refracture, timely and accurate understanding of the patient's fracture situation is undoubtedly a key link in preventing refracture, optimizing treatment strategies and improving the patient's quality of life. Fracture records are scattered in different orthopedic departments such as joint surgery and spinal surgery, and each department will generate detailed medical records, imaging data, treatment plans and follow-up information when dealing with fracture cases in their respective fields. These records are huge and diverse. The traditional method often relies on doctors to log in to multiple information systems to query one by one, and even requires cross-department communication to obtain complete information. This process is not only time-consuming and labor-intensive, but also very easy to cause information omissions or errors due to human factors, which in turn affects the accuracy of diagnosis and the formulation of treatment plans. With the rapid development of medical informatization, the growth rate of medical data is unprecedented, which not only provides the possibility for precision medicine, but also puts higher requirements on data management and analysis capabilities.

[0003] Therefore, how to efficiently and accurately extract effective fracture information from massive data has become an urgent problem to be solved. Summary of the invention

[0004] The present invention overcomes the shortcomings of the prior art and provides a management system for intelligently capturing fracture records and a capture method thereof. By capturing records, the efficiency of obtaining disease-specific information is improved, and time and labor costs are reduced; missed judgments and misjudgments are avoided, and the system can be integrated into existing medical information systems with good compatibility and scalability.

[0005] To achieve the above object, the technical solution adopted by the present invention is: a method for intelligently grabbing fracture records of a management system, comprising the following steps: Pre-constructing a disease-specific database, which includes disease-specific types and a disease-specific keyword library. The establishment of disease-specific types and disease-specific keyword libraries includes: collecting disease-specific type information, extracting keywords from the collected disease-specific type information, and constructing a disease-specific database including disease-specific types and a disease-specific keyword library by collecting disease-specific type information and the extracted keywords; Use the disease-specific database as an information source to build a disease-specific intelligent relationship system model; The steps for building the disease-specific intelligent relationship system model include: Extract medical records matching the keywords in the disease-specific database from the medical business data set; The captured medical records are marked or displayed according to usage requirements, the medical records filtered out from the medical business data set are marked as special disease records, and marked or classified and displayed in the medical management system.

[0006] In a preferred embodiment of the present invention, the extracted keywords corresponding to the fracture types are stored in a memory database, and used as a disease-specific database that can dynamically update disease-specific types and disease-specific keyword libraries; The update of the disease-specific database includes: extracting keywords that collect disease-specific information based on the TF-IDF algorithm and adding them to the disease-specific database.

[0007] In a preferred embodiment of the present invention, the establishment of the medical service data set includes: using a big data ETL tool to perform data preprocessing on the medical service data to form medical data, and converting the medical data into a medical service data set that meets the screening requirements through a data conversion method; Capturing medical records that match special disease keywords in a special disease database from a medical business data set includes: performing keyword matching on a preprocessed medical business data set through an established special disease type and a special disease keyword library to obtain a matching result; and screening out medical records that match special disease keywords in the special disease keyword library based on the matching result.

[0008] In a preferred embodiment of the present invention, the data conversion method includes: data cleaning, data formatting; Data formatting includes unifying the data format of medical data after data cleaning; Data cleaning includes removing duplicate medical data and invalid medical data.

[0009] In a preferred embodiment of the present invention, keyword matching includes the following steps: using a KMP matching algorithm to perform data matching between a medical service data set and a disease-specific database; The KMP matching algorithm includes building a KMP keyword table as a disease-specific keyword library based on the examination result character string in the medical business data set. The KMP keyword table is used to determine the next position of the text pointer when there is no match; the KMP keyword table records the length of the longest equal prefix and suffix before each position of the examination result character string; When a mismatch occurs, the KMP matching algorithm moves the inspection result string to the right to an appropriate position to continue matching without checking the inspection result string that has already matched; and combines full-text search and regular expression matching.

[0010] In a preferred embodiment of the present invention, after screening out the medical records matching the disease-specific keywords according to the matching results of the keyword matching, the following steps are also included: Establish a verification and audit mechanism and set up manual review steps. Manual review is used to confirm relevant records to avoid misjudgment or omissions; And / or, regularly update the disease-specific database and disease-specific intelligent relationship system model to adapt to different application scenarios and needs.

[0011] In a preferred embodiment of the present invention, the keywords include: specific disease names, symptom descriptions, and diagnostic terms; The sources of information collected for special disease types include: medical literature, medical clinical guidelines, and medical expert opinions.

[0012] In a preferred embodiment of the present invention, the TF-IDF algorithm includes: After the TF-IDF value is obtained using the TF-IDF algorithm, it is compared with the set threshold. If it exceeds the set value, it is considered to be a disease-specific keyword and included in the disease-specific keyword library; And / or, ETL algorithm is used to describe the process of extracting, transforming, and loading data from the source to the destination.

[0013] In a preferred embodiment of the present invention, the KMP algorithm is a string matching algorithm, which uses information after matching failure to minimize the number of matches between the pattern string and the main string to achieve the purpose of fast matching; it is implemented through the next() function, and the function itself contains local matching information of the pattern string.

[0014] In a preferred embodiment of the present invention, a management system for intelligently grabbing fracture records is used to implement a grabbing method of the management system for intelligently grabbing fracture records; The system structure of the management system includes: Special disease database, which includes data collection module, data sorting module and special disease data collection module with visualization function; The special disease intelligent relationship system module is interconnected with the special disease database. The special disease intelligent relationship system module includes a patient identification module which is set up with the special disease database as the information source, a file construction module, an auxiliary education module, a follow-up management module, a same-screen display module, a special disease management full-process module, a data quality control module, and a retrieval and export module.

[0015] In a preferred embodiment of the present invention, a management system for intelligently grabbing fracture records is used to implement a grabbing method of the management system for intelligently grabbing fracture records; The application framework of the management system includes: front-end module, gateway module, back-end module and technical platform module which are interconnected; The front-end module includes an interconnected operating platform layer, a technical tool layer, and an application implementation layer; the operating platform layer is used to provide an encapsulated operating environment, the technical tool layer is used to provide basic tools and frameworks, and the application implementation layer is used to implement business functions; the operating environment includes a PC and a Web browser; the basic tools and frameworks include logs and visual interactions; business functions include state models, functional components, and page routing; The back-end module includes an interconnected interface adaptation layer, a business logic layer, and a technical implementation layer; the interface adaptation layer is used to adapt the return results, protocols, and formats of different requests; the business logic layer is used to process core business logic, perform business orchestration, and task scheduling; the technical implementation layer is used to implement database connections, remote service calls, and configuration reading.

[0016] The present invention solves the defects existing in the technical background, and the beneficial technical effects of the present invention are: The management system and capture method of the fracture records of the present invention can improve the efficiency of obtaining disease-specific information and reduce time and labor costs by capturing records; it can avoid missed judgments and misjudgments, and can be integrated into the existing medical information system, with good compatibility and scalability. It has intelligent learning and optimization functions, and can continuously adjust and improve the keyword library and algorithm according to actual results. The comprehensive and detailed keyword library supports complex and diverse fracture types. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0018] Figure 1 It is a work flow chart of a preferred embodiment of the present invention; Figure 2 It is a data table of "OF keywords" in the preferred embodiment of the present invention; Figure 3 It is a data table of "osteoporosis / fragility fracture standard data set field" in the preferred embodiment of the present invention; Figure 4 is a schematic diagram of the system structure in a preferred embodiment of the present invention; Figure 5 is a schematic diagram of a system framework in a preferred embodiment of the present invention; Figure 6 It is a data table of "basic information of included patients" in the preferred embodiment of the present invention; Figure 7 is a schematic diagram of the process of patient data identification in a preferred embodiment of the present invention; Figure 8 It is a data table of "occurrence of re-fracture in included patients" in a preferred embodiment of the present invention; Fig. 9It is a data table of "inclusion of relevant data on management of patients with osteoporosis and fractures" in a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0019] The present invention will now be further described in detail with reference to the accompanying drawings and embodiments. These drawings are simplified schematic diagrams that only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention. Embodiment 1

[0020] like Figure 1-Figure 4 As shown, a method for intelligently grabbing fracture records of a management system includes the following steps: A disease-specific database is pre-constructed, and the disease-specific database includes disease-specific types and disease-specific keyword libraries. The establishment of disease-specific types and disease-specific keyword libraries includes: collecting disease-specific type information, extracting keywords from the collected disease-specific type information, and constructing a disease-specific database including disease-specific types and disease-specific keyword libraries by collecting disease-specific type information and extracted keywords. Specifically, the extracted keywords corresponding to the fracture types are stored in the memory database, and used as a disease-specific database that can dynamically update disease-specific types and disease-specific keyword libraries; The updating of the disease-specific database includes: extracting and collecting keywords of disease-specific type information according to the TF-IDF algorithm and adding them to the disease-specific database. The TF-IDF algorithm extracts keywords and adds them to update the disease-specific database to adapt to changes in medical knowledge. The TF-IDF algorithm includes: after obtaining the TF-IDF value using the TF-IDF algorithm, it is compared with the set threshold, and keywords exceeding the set value are considered to be fracture type keywords and included in the keyword library. Furthermore, the disease-specific database and the disease-specific intelligent relationship system model are regularly updated through the TF-IDF algorithm to adapt to different application scenarios and needs. Keywords include: disease name, symptom description, and diagnostic terminology. The information collection sources for collecting disease-specific type information include: medical literature in related fields, clinical guidelines, and expert opinions.

[0021] A disease-specific intelligent relationship system model is constructed using the disease-specific database as an information source; the steps of constructing the disease-specific intelligent relationship system model include: using the disease-specific intelligent relationship system model to capture medical records that match the keywords in the disease-specific database from the medical business data set.

[0022] Furthermore, the establishment of a medical business data set includes: using a big data ETL tool to pre-process the medical business data to form medical data, and the medical data is converted into a medical business data set that meets the screening requirements through a data conversion method. Specifically, the data conversion method includes: data cleaning and data formatting. The data formatting includes unifying the data format of the medical data after data cleaning. Data cleaning includes removing duplicate data and invalid data of medical data. Among them, the big data ETL tool (Kettle) is used to pre-process the medical business data set, and the required in-hospital outpatient, hospitalization, examination and other medical data are converted into a standard data set that meets the screening requirements. The value selection method includes data cleaning (removing duplicate and invalid data), formatting (unifying data format), etc., to ensure the accuracy and availability of the data. The ETL algorithm is used to describe the process of extracting, converting, and loading data from the source end to the destination end.

[0023] Capturing medical records that match the disease-specific keywords in the disease-specific database from the medical business data set includes: performing keyword matching on the preprocessed medical business data set through the established disease-specific type and disease-specific keyword library to obtain matching results; and filtering out medical records that match the disease-specific keywords in the disease-specific keyword library according to the matching results. Keyword matching includes the following steps: using the KMP matching algorithm to match data. The KMP matching algorithm includes: constructing a KMP keyword table as a disease-specific keyword library according to the inspection result string in the data set, and the KMP keyword table is used to determine the next position of the text pointer when there is no match; the KMP keyword table records the length of the longest equal prefix and suffix before each position of the inspection result string; when a mismatch occurs, the KMP matching algorithm moves the inspection result string to the right to the appropriate position to continue matching without checking the inspection result string that has been matched; and combines full-text search and regular expression matching to improve matching speed and accuracy. Specifically, the KMP algorithm is a string matching algorithm that uses information after matching failure to minimize the number of matches between the pattern string and the main string to achieve the purpose of fast matching; it is implemented through the next() function, and the function itself contains local matching information of the pattern string.

[0024] More specifically, the KMP matching algorithm is an algorithm technology for quickly skipping the matched part of the string, so that when the match fails, there is no need to re-match from the beginning, so as to achieve the effect of speeding up the matching speed. Further, the length of the longest equal prefix and suffix is ​​to record the maximum length of the front part and the back part of the string at the specified position, so as to determine how many characters can be skipped to start a new match after failure, thereby reducing repeated matching. This length must be calculated for each position of the string. For example: for abcabcabcabe, at the position of abcab, the longest length is 2, that is, the length of ab, and ab is the longest equal substring in the substrings obtained from the front and back numbers. This length at each position also constitutes the comparison table in next. When the string match fails, and the substring that has been successfully matched is abcab, it means that the middle position (here is 2 positions) can be skipped to continue matching, rather than re-matching from the next position as in the traditional algorithm. In this embodiment, KMP is a technology for accelerating query retrieval, which speeds up the efficiency of querying out relevant medical record data compared to traditional solutions.

[0025] More specifically, the TF-IDF value is a word frequency value, which indicates how often a word appears in a text. If a word frequency reaches a certain level, it indicates that the word is highly relevant to the document, thus determining whether it is a keyword.

[0026] More specifically, the Next() function's Next means to determine how to offset and skip unnecessary matches through the position table. Next records an array containing the length of the longest equal prefix and suffix corresponding to each position in the string.

[0027] Furthermore, according to the matching results, the medical records that match the disease-specific keywords are screened out. Specifically, according to the matching results, after the medical records that match the disease-specific keywords are screened out, the following steps are also included: establishing a verification and review mechanism, setting up a manual review step, and the manual review is used to confirm the relevant records to avoid misjudgment or omission.

[0028] The captured medical records are marked or displayed according to usage requirements, and the medical records filtered out from the business data set are marked as special disease records, and marked or classified and displayed in the medical management system to facilitate quick search and understanding of special disease situations. Embodiment 2

[0029] On the basis of the first embodiment, when pre-building a disease-specific database: Collect fracture information: First, collect comprehensive information on fracture types by consulting medical literature, clinical guidelines, expert opinions and other channels.

[0030] Keyword extraction: Keywords are extracted from the collected fracture information to form a fracture keyword list. Keywords include but are not limited to specific fracture names, symptom descriptions (such as pain, swelling), diagnostic terms (such as X-ray showing fracture lines), etc.

[0031] Construct a fracture type-fracture keyword library: extract keywords corresponding to fracture types through a memory database to build a set of fracture type-fracture keyword libraries that can be dynamically updated. This library can be manually or automatically extracted and added based on algorithms similar to TF-IDF to adapt to the continuous development of medical knowledge. TF-IDF algorithm (Term Frequency–Inverse Document Frequency) is a commonly used weighting technology for information retrieval and data mining, which is used to evaluate the importance of words to a document in a document collection.

[0032] The steps for building the disease-specific intelligent relationship system model are: Data preprocessing: Use the big data ETL tool (Kettle) to preprocess the medical business data set, and convert the required in-hospital outpatient, hospitalization, examination and other medical data into a standard data set that meets the screening requirements. The value extraction methods include data cleaning (removing duplicate and invalid data), formatting (unifying data format), etc., to ensure the accuracy and availability of the data. ETL (Extract-Transform-Load) is used to describe the process of extracting (extract), transforming (transform), and loading (load) data from the source end to the destination end. The ETL tool refers to the tool that implements this process. Kettle is used as the ETL tool in this system.

[0033] Keyword matching: Using the established fracture types and keyword library, keyword matching is performed on the preprocessed business data set. The KMP matching algorithm is used for data matching. The core of the algorithm is to preprocess the pattern string (referring to the inspection result string in the data set) and build a KMP table (referring to the fracture keyword library). The KMP table is used to determine the next position of the text pointer when there is no match. KMP records the length of the longest equal prefix and suffix before each position of the pattern string; using this KMP table, the algorithm can move the pattern string to the right to the appropriate position to continue matching when a mismatch occurs, without checking the already matched string. And it combines full-text search, regular expression matching and other efficient methods to improve matching speed and accuracy. Furthermore, the KMP algorithm is an improved string matching algorithm. The core of the KMP algorithm is to use the information after the match fails to minimize the number of matches between the pattern string and the main string to achieve the purpose of fast matching. The core of the specific implementation is realized through a next() function, which itself contains the local matching information of the pattern string.

[0034] Result screening: According to the matching results, medical records matching the fracture keywords are screened out. At the same time, a strict verification and review mechanism is established to support doctors to review and confirm relevant records to avoid misjudgment or omissions. In addition, the system also has intelligent learning and optimization functions, which can continuously adjust and improve the keyword library and algorithm according to actual results to adapt to different application scenarios and needs.

[0035] The captured medical records are marked or displayed according to the usage requirements, including the following: Mark the selected medical records as fracture records, and specially mark or classify them in the medical management system to facilitate doctors to quickly find and understand the patient's fracture situation. Provide a detailed summary of fracture information, including but not limited to related medical records, examination and inspection records, medication records, etc., to provide strong support for doctors to formulate treatment plans. Embodiment 3

[0036] A management system for intelligently grabbing fracture records, used to implement the grabbing method of the management system for intelligently grabbing fracture records of embodiment 1; The system structure of the management system includes: A disease-specific database, which includes a data collection module, a data sorting module, and a disease-specific data acquisition module with visualization function.

[0037] The research subjects are as follows: Patient inclusion criteria (one of criteria 3 and 4 can be met): Standard 1. Age ≥XX years old; Standard 2: Seeking medical treatment at XXXX Hospital from XX / XX / XXXX to XX / XX / XXXX (within N years of the occurrence of OF); Standard 3: Major osteoporotic fracture (MOF) diagnosis (vertebral fracture, hip fracture, proximal humerus fracture and distal radius fracture) is included in the outpatient, emergency, hospital admission and discharge diagnoses; Standard 4: The hospital's imaging examination description or diagnosis contains the preset "OF keywords".

[0038] Dataset definition: The data of osteoporotic fracture patients were collected based on the osteoporotic fracture standard data set. The standard data set includes diagnostic information and imaging information. The collection of diagnostic information uses the standardized and unified clinical medical system terminology (SNOMED), osteoporosis or fragility fracture diagnosis and treatment guidelines, and relevant domestic and foreign terminology standards; the collection of imaging information uses the "OF keywords" formulated by experts from the research center.

[0039] The dataset includes 9 modules and 339 fields in total, which define case demographic information, fracture records, medical records, test results, examination results, treatment plans, surgical treatment, follow-up information, etc., and also includes the Chinese and English names, codes, data element descriptions, value domain definitions and ranges of each data element.

[0040] Data collection and processing: In the data integration stage, the system first uses natural language processing technology and medical expertise to deeply analyze the electronic medical record text, convert unstructured natural language descriptions into structured data elements, and accurately match them with the predefined fields in the database. Subsequently, the system will perform multi-dimensional verification and cleaning on the mapped data, identify and handle missing values, outliers, and duplicate records, to ensure the integrity, accuracy, and consistency of the data, and provide a high-quality data foundation for subsequent analysis and application.

[0041] A comprehensive data security system is built from three aspects: network security, data transmission, and data application, and is committed to ensuring the confidentiality, integrity, and availability of data, as well as the legality and compliance of data use and disclosure. Network security: ① Physical isolation: The database is deployed on the server in the computer room of the hospital information center, and a professional information department team is responsible for security. ② Encrypted transmission: The gateway access layer is forced to use the HTTPS protocol for encrypted transmission, and a high-security certificate is used to prevent data from being stolen during transmission. ③ Closed environment: A closed network environment is built through technical means such as VPN, bastion host, and firewall to effectively block external intrusion and minimize the risk of data leakage. Data transmission: ① Encryption protection: The information department encrypts the data in transmission to effectively prevent the risk of third-party eavesdropping. ② Integrity verification: A unified integrity check is performed on the interface input parameters and stored data, and the verification rules are uniformly declared to ensure the consistency and integrity of the data. ③ Data quality control: For data that does not require mandatory verification but needs to be controlled in quality, the system provides a special data quality calculation method to ensure that the data is accurate and reliable. Data application: The database implements multiple security measures: ① IP access control: Strictly limit database access rights, and only authorize designated computers to access to prevent unauthorized access. ② Refined permission management: Different doctors have different levels of database access rights, such as viewing, editing, and exporting, to ensure the principle of minimizing data access. ③ Data desensitization: The SHA256 hash algorithm is used to irreversibly encrypt and desensitize patient privacy data, and sensitive information is deformed to prevent sensitive information leakage. ④ Operation audit: The system records all operation logs, including login permissions, operation content, etc., to achieve traceability of operation behaviors and ensure that data security responsibilities are assigned to individuals. ⑤ Strict export control: If patient information needs to be exported for scientific research or department management, it must be approved by the department director and the hospital information department must use an administrator account to perform the export operation to prevent data abuse to the greatest extent.

[0042] The special disease intelligent relationship system module is interconnected with the special disease database, and the special disease intelligent relationship system module includes a patient identification module, a file construction module, an auxiliary education module, a follow-up management module, a same-screen display module, a special disease management full-process module, a data quality control module, and a retrieval and export module, which are set with the special disease database as the information source.

[0043] In this embodiment, an automatic data collection system for osteoporosis fractures with complete data collection, data collation and visualization functions is established. The data categories include patient electronic medical record information, laboratory information such as blood calcium and bone turnover indicators, imaging information, surgical information, medication records and other modules. A total of 12,754 patients with osteoporosis fractures from XX / XX / XXXX to XX / XX / XXXX were included.

[0044] Patient Identification: The system adopts a three-level screening strategy to gradually narrow the target population. The identification process requires the intervention of fracture management specialists for manual review to ensure the accuracy of management.

[0045] 1. Potential fracture list: The system automatically searches for matching fracture entries from June 1, 2022 to June 1, 2024 (N=12754) Inpatients Outpatients Number, n41698585 Male, n(%)1229 (29.5) 3076 (35.8) Female, n(%)2940 (70.5) 5509 (64.2) Age, mean±SD62.19±10.7761.28±11.53 Initial fracture site Vertebral fracture, n(%)2395 (57.45) 6452 (75.15) Hip fracture, n(%)1121 (26.89) 253 (2.95) Radial fracture Distal humeral fracture, n (%) 374 (8.97) 1162 (13.54) Proximal humeral fracture, n (%) 279 (6.69) 718 (8.36) Data source Fracture diagnosis, n (%) 3273 (78.51) 3079 (35.86) Imaging information, n (%) 896 (21.49) 5506 (64.14) Diagnostic information and imaging description of patients with osteoporotic fractures. If the patient's outpatient, emergency, admission and discharge diagnosis included a major osteoporotic fracture diagnosis or the description of the hospital's imaging examination or the diagnosis contained the preset "OF keywords", they were included in the "potential fracture list", and a total of 15,482 patients were included.

[0046] 2. High-risk population list: For patients on the potential fracture list, the system uses NLP technology to automatically identify and extract key information from the patient's electronic medical record, such as fracture time, location, type, treatment method, etc., to generate a "high-risk population list" to provide fracture management specialists with an overall view of the condition and assist in manual review.

[0047] 3. List of target population: Fracture management specialists manually screened the information provided by the “List of High-risk Populations” and excluded: ① 146 patients who were younger than 50 years old at the time of fracture, ② 724 patients with fractures caused by high-energy violence, ③ 1,295 patients whose fractures had occurred more than 2 years ago, and ④ 563 outpatient and emergency patients with no contact information. Finally, 12,754 patients were included in the management.

[0048] Archive build: For the target population, the system automatically captures relevant information from various systems within the hospital, including medical records, test data, osteoporosis medication records, medical records, etc., cleans, converts and integrates the data, removes redundant information, and builds a structured initial management file for OF patient re-fracture prevention and treatment. Fracture management specialists can quickly understand the patient's condition through the file, and communicate through the contact information of the treating doctor associated with the system to improve the patient's information and provide a basis for subsequent precise intervention.

[0049] For inpatients in the target population, the system automatically captures the patient's ward and bed number information, and combines it with the initial management files of refracture prevention and treatment to form an "inpatient education list". The fracture management specialist can quickly locate the patient's position through the ward bed information provided by the system, and choose the appropriate time to conduct face-to-face education for the patient based on the surgical information provided by the system. After the education is completed, the fracture management specialist needs to fill in the completion status in the system. For non-inpatients in the target population, where face-to-face education cannot be achieved, the system can be combined with the hospital's internal SMS service to accurately push the education content through SMS according to the pre-established education rules.

[0050] Follow-up management: After the patient completes the patient education, the system automatically includes them in the management list and generates a personalized follow-up plan according to the preset rules. ① Initial follow-up plan: The first follow-up is scheduled one month after surgery, and then every three months until the fracture occurs for two years. ② Dynamic adjustment: If the patient re-fractures during the follow-up process, the system will automatically regenerate the follow-up plan. Fracture management specialists can also manually adjust the follow-up plan according to the actual situation of the patient to enhance flexibility. According to the follow-up plan, the system will automatically send a text message reminder to the patient one week before each follow-up. After the patient completes the follow-up, the system automatically captures the follow-up record, searches for OF-related information, updates the patient's fracture management file, and ensures the timeliness of the information. In addition, the system has built-in commonly used OF-related follow-up questionnaires for standardized data collection. At the same time, the system supports custom questionnaires to meet scientific research needs. All questionnaire results will be saved in a structured manner by the system so that researchers can directly call for data analysis without tedious data sorting.

[0051] On-screen display: The system is deeply integrated with the hospital's existing outpatient / inpatient medical record system. When a managed patient visits the hospital, the patient page of the outpatient / inpatient operating system will intelligently display a striking logo to remind the doctor that the patient is in the management process of osteoporotic fractures. The doctor only needs to specify a module and can access the patient's complete re-fracture prevention and treatment management file on the current page with one click, including key information such as the diagnosis, treatment plan, bone density data, osteoporosis-related test results, and osteoporosis medication records of previous fractures, providing comprehensive information support for the doctor's diagnosis and treatment decisions, improving work efficiency, and ensuring the continuity of diagnosis and treatment.

[0052] Refracture Tips: In order to achieve real-time monitoring of patient status, the system automatically updates patient data in HIS, PACS, LIS and other systems every day. Once a new fracture record is detected, the system will immediately start intelligent analysis, integrate the patient's basic information, previous fracture conditions and the source of the new fracture record, generate a "possible re-fracture list", and promptly notify the fracture management specialist in a striking manner. The list provides specialists with a quick judgment basis to help them promptly confirm whether the patient has re-fractured and quickly take appropriate intervention measures to effectively avoid delays in treatment. Among the 12,754 patients managed by the system, a total of 1,951 re-fracture events were identified, involving 1,158 patients, with a cumulative incidence of 15.30%. Among them, there were 1,142 vertebral re-fractures (8.95%), 372 hip re-fractures (2.92%), 298 distal radius fractures (2.34%), and 139 distal humeral fractures (1.09%). It is noteworthy that among the 1951 refracture events, 575 did not have a recorded fracture diagnosis but were captured by the system's recognition of the "OF keywords" in the imaging diagnosis and description, of which 419 were vertebral fractures, accounting for 72.87%.

[0053] Data quality control: The system automatically collects patient data and generates multi-dimensional quality control charts in real time, covering key indicators such as patient gender ratio, age distribution, bone density and bone turnover index detection rate, hospitalization diagnosis and osteoporosis treatment ratio, etc., which fully reflects the diagnosis and treatment status of patients with brittle fractures. All quality control data can be accurately reported to each patient's attending physician and treatment team leader. Department managers can use the system to monitor the changing trends of various quality control indicators in real time, promptly discover deviations in the diagnosis and treatment process, and take effective measures for continuous improvement, provide targeted rectification suggestions to relevant doctors, continuously optimize the diagnosis and treatment process, improve the quality of diagnosis and treatment, and ensure patient safety. Among the 12,754 patients included in the system, the diagnosis rate of fracture in hospitalized patients was 78.51% (3,273 cases), the diagnosis rate of osteoporosis was 62.29% (2,597 cases), the bone density examination rate was 49.56% (2,066 cases), the bone turnover index examination rate was 76.28% (3,180 cases), and the osteoporosis treatment rate was 72.06% (3,004 cases), of which the basic treatment rate was 28.50% (1,188 cases), the anti-bone resorption drug treatment rate was 36.97% (1,537 cases), and the osteoproliferative drug treatment rate was 6. 69% (279 cases); the fracture diagnosis rate of outpatients was 35.86% (3079 cases), the osteoporosis diagnosis rate was 25.14% (2158 cases), the bone density examination rate was 9.70% (833 cases), the bone turnover index examination rate was 8.12% (697 cases), and the osteoporosis treatment rate was 52.96% (4547 cases), of which the basic treatment rate was 36.60% (3142 cases), the anti-bone resorption drug treatment rate was 15.187% (1303 cases), and the osteoproliferative drug treatment rate was 1.19% (102 cases). The chi-square test was used for statistical analysis, and the results showed that the fracture diagnosis rate, osteoporosis diagnosis rate, bone density examination rate, bone turnover index examination rate, osteoporosis treatment rate, anti-bone resorption drug treatment rate and osteoproliferative drug treatment rate of inpatients were significantly higher than those of outpatients (P<0.05); the basic treatment rate was significantly lower than that of outpatients (P<0.05).

[0054] Data retrieval and export: The system has built-in data retrieval and export functions, providing researchers with convenient ways to obtain and analyze data. The system provides retrieval methods such as quick retrieval and advanced retrieval. The retrieval fields are connected to the data set fields. Researchers can freely select data fields according to research needs. After review by the department director and the hospital information department, the included patient data will be exported in the form of Excel tables. Researchers can also use the cohort export function to directly export the data of a cohort of patients in the database for statistical analysis through the advanced retrieval function provided by the database.

[0055] Working principle: The present invention aims to break through the limitations of traditional management models and realize accurate and intelligent management of osteoporotic fractures. The research results show that the system has significant advantages in improving patient identification efficiency, optimizing management processes, and improving diagnosis and treatment quality: The system has successfully built a high-quality disease-specific database. By establishing a unified data interface, the system has broken down the barriers between multi-source heterogeneous data such as HIS, PACS, and LIS, achieved the interconnection of information within the hospital, and provided a high-quality, structured data foundation for the management and research of osteoporotic fractures.

[0056] The system showed high sensitivity to re-fracture events and could effectively identify potential re-fracture risks.

[0057] The system provides functions such as precise positioning, personalized education, intelligent follow-up and multi-dimensional data quality control, which provide strong support for optimizing the OF patient management process and improving the quality of diagnosis and treatment.

[0058] This system is a closed-loop system operated within the hospital. Data collection, processing, and circulation are all carried out within the hospital's data system, complying with relevant national laws and regulations on patient information security. Embodiment 4

[0059] A management system for intelligently capturing fracture records is used to implement the capture method of the management system for intelligently capturing fracture records in embodiment 1; the management system is based on the JavaEE platform, adopts front-end and back-end separation and microservice architecture, and realizes high performance and easy-to-expand functions.

[0060] The application framework of the management system includes: interconnected front-end modules, gateway modules, back-end modules, and technical platform modules.

[0061] The front-end module includes an interconnected operating platform layer, a technical tool layer, and an application implementation layer; the operating platform layer is used to provide an encapsulated operating environment, the technical tool layer is used to provide basic tools and frameworks, and the application implementation layer is used to implement business functions; wherein, the operating environment includes a PC and a Web browser; the basic tools and frameworks include providing logs and visual interactions; the business functions include state models, functional components, and page routing; specifically, the front-end module adopts the MVVM model to separate views and models, thereby improving code reusability and testability; at the same time, shell technologies such as CefSharp, Electron, and Chromium are used to achieve cross-platform interface consistency.

[0062] The backend module includes an interconnected interface adaptation layer, a business logic layer, and a technical implementation layer; the interface adaptation layer is used to adapt the return results, protocols, and formats of different requests, the business logic layer is used to process the core business logic, perform business orchestration, and task scheduling, and the technical implementation layer is used to implement database connection, remote service calls, and configuration reading; specifically, the backend builds microservices based on Spring Cloud to achieve high availability and scalability. The gateway layer uses Nginx and SpringCloud Gateway to provide: resource and request distribution, authority authentication, traffic cleaning, blacklist and other security controls, monitoring functions such as logging, caching, and current limiting, and dynamic routing rules.

[0063] The above specific implementation methods are specific support for the scheme ideas proposed in the present invention, and cannot be used to limit the protection scope of the present invention. Any equivalent changes or equivalent modifications made on the basis of this technical scheme in accordance with the technical ideas proposed in the present invention still fall within the scope of protection of the technical scheme of the present invention.

Claims

1. A method for intelligently capturing fracture records in a management system, characterized in that: The following steps are involved: Pre-constructing a disease-specific database, the disease-specific database including disease-specific types and disease-specific keyword libraries, the establishment of the disease-specific types and disease-specific keyword libraries including: collecting disease-specific type information, extracting keywords from the collected disease-specific type information, and constructing a disease-specific database including disease-specific types and disease-specific keyword libraries by collecting disease-specific type information and the extracted keywords; Use the disease-specific database as an information source to build a disease-specific intelligent relationship system model; The steps of constructing the disease-specific intelligent relationship system model include: Extract medical records matching the keywords in the disease-specific database from the medical business data set; The captured medical records are marked or displayed according to usage requirements, the medical records filtered out from the medical business data set are marked as special disease records, and marked or classified and displayed in the medical management system.

2. The method for intelligently capturing fracture records of a management system according to claim 1, characterized in that: The extracted keywords corresponding to the fracture types are stored in the memory database, and used as a disease-specific database that can dynamically update disease-specific types and disease-specific keyword libraries; The update of the disease-specific database includes: extracting keywords that collect disease-specific information based on the TF-IDF algorithm and adding them to the disease-specific database.

3. The method for intelligently capturing fracture records of a management system according to claim 2, characterized in that: The establishment of the medical business data set includes: using a big data ETL tool to pre-process the medical business data to form medical data, and the medical data is converted into a medical business data set that meets the screening requirements through a data conversion method; Capturing medical records that match special disease keywords in a special disease database from a medical business data set includes: performing keyword matching on a preprocessed medical business data set through an established special disease type and a special disease keyword library to obtain a matching result; and screening out medical records that match special disease keywords in the special disease keyword library based on the matching result.

4. The method for intelligently capturing fracture records of a management system according to claim 3, characterized in that: The data conversion method includes: data cleaning and data formatting; The data formatting includes unifying the data format of the medical data after data cleaning; The data cleaning includes removing duplicate medical data and invalid medical data.

5. The method for intelligently capturing fracture records of a management system according to claim 4, characterized in that: The keyword matches The method includes the following steps: using the KMP matching algorithm to match the medical business data set with the disease-specific database; The KMP matching algorithm includes building a KMP keyword table as a disease-specific keyword library based on the examination result character string in the medical business data set. The KMP keyword table is used to determine the next position of the text pointer when there is no match; the KMP keyword table records the length of the longest equal prefix and suffix before each position of the examination result character string; When a mismatch occurs, the KMP matching algorithm moves the inspection result string to the right to an appropriate position to continue matching without checking the inspection result string that has already matched; and combines full-text search and regular expression matching.

6. The method for intelligently capturing fracture records of a management system according to claim 5, characterized in that: According to the matching results of keyword matching, after screening out the medical records that match the keywords of the specific disease, the following steps are also included: Establish a verification and audit mechanism and set up manual review steps. Manual review is used to confirm relevant records to avoid misjudgment or omissions; And / or, regularly update the disease-specific database and disease-specific intelligent relationship system model to adapt to different application scenarios and needs.

7. The method for intelligently capturing fracture records of a management system according to claim 6, characterized in that: The keywords include: disease names, symptom descriptions, and diagnostic terms; The information collection sources for collecting disease-specific information include: medical literature, medical clinical guidelines, and medical expert opinions.

8. The method for intelligently capturing fracture records of a management system according to claim 6, characterized in that: The TF-IDF algorithm includes: After the TF-IDF value is obtained using the TF-IDF algorithm, it is compared with the set threshold. If it exceeds the set value, it is considered to be a disease-specific keyword and included in the disease-specific keyword library; And / or, ETL algorithm is used to describe the process of extracting, transforming, and loading data from the source to the destination.

9. The method for intelligently capturing fracture records of a management system according to claim 8, characterized in that: The KMP algorithm is a string matching algorithm that uses information after matching failures to minimize the number of matches between the pattern string and the main string to achieve fast matching. It is implemented through the next() function, which itself contains local matching information of the pattern string.

10. A management system for intelligently capturing fracture records, characterized in that: A method for implementing a management system for intelligently capturing fracture records as described in any one of claims 1 to 9; The system structure of the management system includes: A disease-specific database, which includes a data collection module, a data sorting module, and a disease-specific data acquisition module with visualization function; The special disease intelligent relationship system module is interconnected with the special disease database, and the special disease intelligent relationship system module includes a patient identification module set up with the special disease database as the information source, a file construction module, an auxiliary education module, a follow-up management module, a same-screen display module, a special disease management full-process module, a data quality control module, and a retrieval and export module.

Citation Information

Patent Citations

  • Construction method of general medical special disease data system

    CN108538395A

  • Medical decision system based on medical big data and artificial intelligence

    CN114842977A

  • Electronic medical record duplicate checking rate estimation method based on data stream matching

    CN115983233A

  • Multi-source heterogeneous system fracture population data intelligent reading method, system and device

    CN116486974A

  • Doctor-patient video communication intelligent allocation method and system based on remote consultation scene

    CN118317047A