Knowledge base-based medical examination item intelligent multi-dimensional standardization method, system, electronic device and storage medium

Abstract

The application discloses a kind of medical examination item intelligent multidimensional standardization method, system, electronic equipment based on knowledge base, storage medium, belong to medical big data information processing technical field.The application constructs four-dimensional medical knowledge covering examination type, examination method, examination site, observation variable and disambiguation, conflict detection, combination word analysis rule, through pre-processing historical matching, hierarchical structure split, context constraint intelligent standardization, multi-rule linkage verification, confidence level grading output, artificial closed loop self-learning architecture, solve the defects that existing cross medical institution examination item standardization technology dimension is single, abbreviation cannot be parsed, semantic ambiguity misjudgment, result logic conflict, cannot iterate.The application can complete the multidimensional standardization processing of medical examination item name, improve standardization accuracy and processing efficiency, adapt to regional medical data convergence, medical association quality control, clinical research data analysis and other various medical big data application scenarios.

Description

Technical Field

[0001] This application belongs to the technical field of medical data processing, specifically relating to a knowledge base-based intelligent multidimensional standardization method, system, electronic device, and storage medium for medical examination items. Background Technology

[0002] In scenarios such as regional medical data collaboration, data aggregation within medical alliances / consortia, and medical performance evaluation, it is necessary to uniformly aggregate and compare examination and testing data from different medical institutions. Medical examination items are important data generated during clinical diagnosis and treatment, including various types such as CT, MRI, ultrasound, X-ray / radiology, endoscopy, pathology, and nuclear medicine. However, significant differences exist in the naming methods used by different medical institutions for the same examination item, making it impossible to directly correlate and compare data across institutions. For example, the same obstetric ultrasound examination may be named in various forms in different institutions, such as "Obstetric Color Doppler Ultrasound (Single-Pregnancy 2D Color Doppler Ultrasound + Umbilical Cord Blood Flow + Placental Maturity)," "Color Doppler Ultrasound (Single-Pregnancy 2D Color Doppler Ultrasound + Umbilical Cord Blood Flow + Placental Maturity)," or "Fetal Umbilical Cord Blood Flow + Placental Maturity Examination (Maternity Insurance Out-of-Pocket Special Payment)," etc.; the same ERCP examination may be originally classified as "Radiology," but actually involves both endoscopic procedures and X-ray fluoroscopy; the same MR examination may be written as "Bile and Pancreatic Duct MR Imaging (MRCP) + Upper Abdominal MR Plain Scan + Enhanced (3.0T)," including multiple dimensions such as method, location, plain scan / enhanced scan, etc. The above-mentioned naming differences and classification biases make it difficult for traditional data processing methods based on string equality or full-text similarity to achieve automatic association of examination items across institutions, which seriously restricts the statistical analysis, quality control management, and clinical research applications after medical data aggregation. Therefore, it is necessary to standardize the vocabulary of each dimension in medical examination items.

[0003] In existing technologies, one standardization scheme uses a full-text search engine and word segmentation technology to break down examination names, combining manual matching experience values ​​to classify matching levels, and employing an intelligent matching-assisted manual matching mode. This type of scheme uses a flat word segmentation matching mode, which cannot recognize the inherent hierarchical structure of examination item names, outputting a single-dimensional standard name mapping. It cannot express the multi-dimensional semantics of complex items, and lacks a context-aware mechanism. The same keyword cannot be mapped differently under different examination types, and it cannot handle the automatic decomposition of combined abbreviations such as CTA and MRCP. Another standardization scheme standardizes medical imaging examination sites by integrating real-time acquired data and historical data, relying on site standardization models and synonym / near-synonym calculations to achieve unified site names. However, this type of scheme only addresses the single dimension of the examination site and does not involve the standardization of other dimensions such as examination type, examination method, and observed variables. Its applicability is limited to image quality control scenarios and cannot be extended to all types of examination items. Furthermore, the synonym calculation does not consider contextual constraints, which poses a risk of ambiguity. Another standardization approach combines medical knowledge graphs and pre-trained models to perform semantic similarity assessment, mapping the original examination item names to unified standard examination items. This type of approach relies on knowledge graphs and deep learning models, which have high costs for model training and graph construction. The output result is still a single standard item mapping, without performing multi-dimensional structured decomposition of the examination item names. For combined abbreviations, it only uses semantic similarity matching, which cannot guarantee that they are decomposed into accurate multi-dimensional standard value combinations.

[0004] It is evident that existing technologies generally suffer from the following shortcomings: First, standardization methods based on full-text or full-string matching have low accuracy, cannot identify the inherent structure of names, are prone to false positives and false negatives, and have low computational efficiency for full-database matching. Second, mapping methods based on a single dictionary or a single standard item cannot handle composite inspection items, have high dictionary maintenance costs, and cannot cope with newly added item combinations. Third, they lack context-aware standardization mechanisms, making it easy for the same word to generate ambiguity and mislabeling under different inspection types. Fourth, they cannot automatically handle abbreviations and compound words, resulting in incomplete standardization results. Fifth, they lack conflict detection mechanisms, easily outputting standardization results with logical conflicts. A false negative refers to inspection content that is essentially the same but is judged as different, while a false positive refers to two inspection contents that are essentially different but are incorrectly merged into two completely different dimensions. Summary of the Invention

[0005] To address one or more of the aforementioned problems, this application provides a first aspect of a knowledge-based intelligent multidimensional standardization method for medical examination items, a second aspect of a knowledge-based intelligent multidimensional standardization system for medical examination items, a third aspect of an electronic device, and a fourth aspect of a storage medium, for improving the accuracy of medical examination item identification in cross-institutional data standardization linkage.

[0006] The technical solution of this application is as follows.

[0007] The first aspect is a knowledge-based intelligent multidimensional standardization method for medical examination items, including: The original examination item names across multiple medical institutions or information systems are preprocessed and matched against historical lookup tables. If a match is successful, the historical standardized results are directly output; otherwise, subsequent standardization is performed. According to the preset delimiter syntax hierarchy priority, the names of the original check items that failed to match in the historical lookup table are structurally split to obtain a semantic list containing the outer main name and the inner sub-items. Based on a pre-built multi-dimensional medical knowledge base, the examination type is identified in the split semantic list, and the combination of abbreviations in the medical field is split and parsed to supplement and correct the standardized dimensions of the examination type. Using the identified examination type as a context constraint, a subset of the multi-dimensional medical knowledge base corresponding to the examination type is invoked to complete the standardized matching of the three dimensions of examination method, examination site, and observed variables. Based on the pre-set disambiguation rule base, context-differentiated disambiguation processing is performed on polysemous keywords, and mutual exclusion conflict detection and correction are completed for each standardized dimension based on the exclusion word rule base. The confidence level is graded based on the coverage completeness of the standardized matching results. Differential processing is performed on results with different confidence levels, and the formatted output includes four-dimensional standardized results containing inspection type, inspection method, inspection location, and observed variables. The review process includes updating historical comparison tables, a multi-dimensional medical knowledge base, and a disambiguation rule base.

[0008] As one implementation method of the first aspect, preprocessing and historical lookup table matching specifically include: Read the data of the original examination items from the examination item dictionaries of the respective HIS, PACS or RIS medical information systems of multiple medical institutions; Perform data cleaning operations on the original inspection items, including removing invalid special characters, standardizing the case of English characters, and standardizing the delimiter format. The names of the original inspection items after data cleaning are matched precisely with the historical comparison table. If a match is successful, the historical standardized results are returned directly.

[0009] As one implementation of the first aspect, the separator syntax hierarchy priority includes: The first priority is the bracket-type separator, which is used to extract the content within the brackets and distinguish the outer main name from the inner sub-items; The second priority is the plus sign separator, used to split multiple parallel inner sub-items within and / or outside the parentheses; The third priority is slashes, connectors, and coordinating conjunctions, used to separate independent semantic units under the same outer main name.

[0010] As one implementation of the first aspect, the inspection type identification adopts the longest matching strategy and performs word boundary verification, giving priority to matching the synonyms with the longest character length in the multi-dimensional medical knowledge base; If no corresponding examination type is identified for the outer main name and all inner sub-items, the standardized matching of subsequent dimensions will be performed by loading the full multi-dimensional medical knowledge base and marking the corresponding results as low confidence.

[0011] As one implementation of the first aspect, the splitting and parsing of medical abbreviations and compound words is achieved through a compound word recognition rule base; The rule base for recognizing compound words stores a multidimensional mapping relationship between abbreviations and inspection types, inspection methods, and inspection locations, as well as regular expression patterns used to match variant expressions; If the inspection type has been identified, the word splitting and parsing only supplements and corrects the standardized dimensions of the inspection method; if the inspection type has not been identified, the word splitting and parsing supplements the standardized dimensions of the inspection type at the same time.

[0012] As one implementation of the first aspect, the standardized matching of the three dimensions under context constraints specifically includes: Load the multi-dimensional medical knowledge base subsets of the examination method, examination site, and observation variables corresponding to the current examination type, and perform keyword matching only within the subsets constrained by the context. If one of the vocabulary standardizations within the same inner-layer sub-project matches multiple optional standardization dimensions, then the priority is assigned according to the order of inspection method over inspection location, and inspection location over observed variable. If different words within the same inner sub-item are standardized and matched to different standardized dimensions, then the corresponding standardized dimensions are retained.

[0013] As one implementation of the first aspect, the disambiguation rule base includes different standardized dimensions of the same polysemous keyword under different inspection types; Contextual disambiguation processing includes selecting the standardized dimension corresponding to the polysemous keyword based on the determined inspection type, and marking it as pending manual review if there is no corresponding standardized dimension. The exclusion word rule base includes standardized dimensions of mutually exclusive inspection methods, inspection locations, and observed variables under contextual constraints of different inspection types; When mutually exclusive normalized dimensions are detected, the higher-priority normalized dimension is retained according to the preset priority and the mutual exclusion conflict information is recorded.

[0014] Secondly, a knowledge-based intelligent multidimensional standardization system for medical examination items includes: The preprocessing module is used to preprocess the original examination item names across multiple medical institutions or information systems and match them with historical reference tables. If the match is successful, the historical standardized results are directly output; otherwise, subsequent standardization is performed. The structured splitting module is used to structurally split the names of the original check items that failed to match in the historical lookup table according to the preset delimiter syntax hierarchy priority, and obtain a semantic list containing the outer main name and the inner sub-items. The identification module is used to identify the examination type of the split semantic list based on a pre-built multi-dimensional medical knowledge base, and to split and parse medical abbreviations and combinations to supplement and correct the standardized dimensions of the examination type. The standardization module is used to call a subset of the multi-dimensional medical knowledge base corresponding to the examination type as a context constraint, and to complete the standardization matching of the three dimensions of examination method, examination site, and observed variables respectively. The disambiguation inspection module is used to perform context-differentiated disambiguation processing on polysemous keywords based on a preset disambiguation rule library, and at the same time, it completes mutual exclusion conflict detection and correction for each standardized dimension based on the exclusion word rule library. The confidence processing module is used to classify the confidence level based on the coverage and completeness of the standardized matching results, perform differential processing on results with different confidence levels, and output a formatted output containing four-dimensional standardized results including inspection type, inspection method, inspection location, and observed variables. The iteration module is used to review and update the historical comparison table, the multi-dimensional medical knowledge base, and the disambiguation rule base.

[0015] In a third aspect, an electronic device includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor, when executing the computer program, implements the knowledge base-based intelligent multidimensional standardization method for medical examination items as described in the first aspect.

[0016] Fourthly, a storage medium, which is a computer-readable storage medium, stores computer instructions for enabling a computer to implement the knowledge base-based intelligent multidimensional standardization method for medical examination items as described in the first aspect.

[0017] Compared with the prior art, the advantages of this application are as follows: The standardization process achieves higher accuracy and fewer false matches; the output retains multi-dimensional semantics, resulting in stronger usability of the standardized data; the standardization process requires less computation and is more efficient; it can automatically handle combined abbreviations, reducing manual maintenance costs; the standardization results have better logical consistency; and the knowledge base can be continuously updated to adapt to changes in terminology. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating one of the knowledge base-based intelligent multidimensional standardization methods for medical examination items in this application.

[0019] Figure 2 This is a schematic diagram of the structural framework of one of the knowledge base-based intelligent multidimensional standardization systems for medical examination items in this application.

[0020] Figure 3 This is a schematic diagram of the frame of one of the electronic devices in this application. Detailed Implementation

[0021] Referring to the illustrations, the principles of this application are illustrated by way of example implementation in a suitable operating environment. The following description is based on the illustrative specific embodiments of this application and should not be construed as limiting other specific embodiments not detailed herein, such as corresponding adjustments to the order of steps of the method of this application based on the principles of the technical concept.

[0022] It should be noted that this application only standardizes the names of examination items in the medical institution's examination data, and does not involve sensitive medical data such as patient personal identification information, medical record content, and diagnostic conclusions. It complies with the relevant requirements of laws, regulations, and ethical norms regarding de-identification and minimal processing of personal information.

[0023] like Figure 1 As shown in the figure, this application discloses an intelligent multidimensional standardization method for medical examination items based on a knowledge base, which includes the following steps.

[0024] S1. Preprocess the original examination item names across multiple medical institutions or information systems and match them against historical reference tables. If the match is successful, output the historical standardized results directly; otherwise, perform subsequent standardization.

[0025] In this embodiment, preprocessing and historical lookup table matching specifically include: S11. Read the data of the original examination items from the examination item dictionary of the respective HIS, PACS or RIS medical information systems of multiple medical institutions; S12. Perform data cleaning operations on the original inspection items, including removing invalid special characters, unifying the case of English characters, and standardizing the delimiter format. S13. Perform exact string matching between the names of the original inspection items after data cleaning and the historical comparison table. If the match is successful, return the historical standardized results directly.

[0026] In this embodiment, optional standardized separator formats include: unifying full-width brackets and half-width brackets into half-width brackets, merging multiple consecutive spaces into a single space, and unifying the Chinese plus sign "+" into the English plus sign "+", etc.

[0027] In this embodiment, the historical lookup table records the mapping relationship between the names of the original inspection items and the confirmed four-dimensional standardized results.

[0028] S2. According to the preset delimiter syntax hierarchy priority, the names of the original check items that failed to match in the historical lookup table are structurally split to obtain a semantic list containing the outer main name and the inner sub-items.

[0029] In this embodiment, the priority of the delimiter syntax hierarchy includes: The first priority is the bracket-type separator, which is used to extract the content within the brackets and distinguish the outer main name from the inner sub-items; The second priority is the plus sign separator, used to split multiple parallel inner sub-items within and / or outside the parentheses; The third priority is slashes, connectors, and coordinating conjunctions, which are used to split independent semantic units (corresponding to inspection methods, inspection locations, and observation variables) under the same outer main name, so as to facilitate the standardization of each independent semantic unit.

[0030] S3. Based on a pre-built multi-dimensional medical knowledge base, the examination type is identified in the split semantic list, and the abbreviations in the medical field are split and parsed to supplement and correct the standardized dimensions of the examination types.

[0031] In this embodiment, the inspection type identification adopts the longest matching strategy and performs word boundary verification, prioritizing the matching of the synonym with the longest character length in the multi-dimensional medical knowledge base.

[0032] In this embodiment, the splitting and parsing of medical abbreviations is achieved through a rule base for recognizing abbreviations. The rule base stores the multidimensional mapping relationship between abbreviations and examination types, examination methods, and examination sites, as well as regular expression patterns used to match variant expressions. If an examination type has been identified, the splitting and parsing of abbreviations only supplements and corrects the standardized dimension of the examination method. If no examination type has been identified, the splitting and parsing of abbreviations supplements the standardized dimension of the examination type.

[0033] In this embodiment, the subset of the multidimensional medical knowledge base corresponding to the examination type includes standard types and their synonym sets; the standard types include fields such as CT, MR, ultrasound, X-ray / radiology, endoscopy, pathology, and nuclear medicine; for example, "X-ray / radiology" refers to routine radiological examinations that use the principle of X-ray imaging, including X-ray photography, fluoroscopy, X-ray angiography, etc., but does not include CT, MR, and nuclear medicine examinations. "X-ray / radiology" is a standard examination type entry in the subset of the multidimensional medical knowledge base corresponding to the examination type, existing alongside CT and MR, so as to independently standardize the common "radiology" examination items in the medical institution's examination item dictionary.

[0034] In this embodiment, the abbreviation combination word splitting and parsing is performed after the check type identification; if the check type has been identified, the check method is supplemented and corrected only using the combination word rules; if the check type has not been identified, the check type is supplemented using the type field in the combination word rules.

[0035] In this embodiment, the abbreviation combination word splitting and parsing results are merged with the subsequent standardized matching results; if the type field in the combination word rule is consistent with the identified check type, only the standardized dimension of the check method is output to avoid duplication; if the check method output by the combination word rule is the same as the standardized matching result of the context constraint, the duplicate is removed to avoid the same check method being output twice.

[0036] S4. Using the identified examination type as a context constraint, call the corresponding multi-dimensional medical knowledge base subset for the examination type to complete the standardized matching of the three dimensions of examination method, examination site, and observed variables.

[0037] In this embodiment, if no corresponding examination type is identified in the outer layer main name and all inner layer sub-items in step S3, then the full multi-dimensional medical knowledge base is loaded to perform standardized matching.

[0038] In this embodiment, the standardized matching of the three dimensions under context constraints specifically includes: S41. Load the multi-dimensional medical knowledge base subsets of the examination method, examination site, and observation variables corresponding to the current examination type, and perform keyword matching only within the subsets after context constraints. S42. If one of the terms in the same inner sub-item matches multiple optional standardized dimensions, the term is assigned according to the priority of the inspection method over the inspection location, and the inspection location over the observed variable. For example, if a term matches both the inspection method and the inspection location, the term is assigned to the inspection method if the inspection method is prioritized over the inspection location, and if it matches both the inspection location and the observed variable, the term is assigned to the inspection location if the inspection location is prioritized over the observed variable. S43. If different terms in the same inner sub-item are standardized and matched to different standardized dimensions, the corresponding standardized dimensions are retained. For example, in the sub-item "two-dimensional color ultrasound", "two-dimensional" is mapped to the examination method (two-dimensional grayscale ultrasound), and "color ultrasound" is mapped to the examination method (color Doppler). The two belong to the same examination method and their standardized dimensions do not conflict. If the inner sub-item also contains the term "fetus" which is mapped to the examination site, then "two-dimensional" and "color ultrasound" are different terms from "fetus" and are matched to different standardized dimensions, so they are retained at the same time.

[0039] In this embodiment, the priority allocation of standardized matching is determined based on the regular naming rules of the inspection items: the inspection method usually directly determines the technical means of inspection, the inspection location describes the object on which the inspection method is applied, and the observed variables are used to describe the indicators or states that the inspection method focuses on. Therefore, the priority allocation is carried out in the order of inspection method, inspection location, and observed variables.

[0040] S5. Based on the preset disambiguation rule base, perform context-differentiated disambiguation processing on polysemous keywords, and at the same time, based on the exclusion word rule base, complete the mutual exclusion conflict detection and correction of each standardized dimension.

[0041] In this embodiment, the disambiguation rule base includes different standardized dimensions of the same polysemous keyword under different inspection types; the context-differentiated disambiguation processing includes selecting the standardized dimension corresponding to the polysemous keyword according to the determined inspection type, and marking it as pending manual review if there is no corresponding standardized dimension; the exclusion word rule base includes mutually exclusive standardized dimensions of inspection methods, inspection locations, and observation variables under different inspection type context constraints; when mutually exclusive standardized dimensions are detected, the higher priority standardized dimension is retained according to the preset priority and the mutual exclusion conflict information is recorded.

[0042] S6. Based on the coverage and completeness of the standardized matching results, the confidence level is classified, and different processing is performed on the results of different confidence levels. The formatted output includes four-dimensional standardized results containing inspection type, inspection method, inspection location, and observed variables.

[0043] In this embodiment, if the corresponding examination type is not identified in the outer layer main name and all inner layer sub-items in step S3, and the standardized matching is performed by loading the full multi-dimensional medical knowledge base in step S4, the corresponding standardized matching result is marked as low confidence.

[0044] In this embodiment, one optional confidence level is as follows: when the four-dimensional results are covered and there are no conflicts, the confidence level is high; when only the three-dimensional results are covered, the confidence level is medium; and when the one- to two-dimensional results are covered, the confidence level is low.

[0045] In this embodiment, an optional example of formatted output is as follows: the inspection type, inspection method / details, inspection location, and observation variables are all separated by the "|" symbol.

[0046] The standardized results are stored in a standard database, which is used to support subsequent data indexing, retrieval, statistical analysis, and cross-institutional data comparison.

[0047] S7. Review and iterate the historical comparison table, multi-dimensional medical knowledge base, and disambiguation rule base.

[0048] Optionally, in this embodiment, the review is conducted manually. Results with medium confidence can be confirmed in batches, while results with low confidence and those that cannot be mapped are reviewed one by one. Correctly confirmed standardized results are automatically updated to the historical comparison table. Newly identified standardized content is then extracted from the manually confirmed results and automatically added to the corresponding synonym set in the multidimensional medical knowledge base. When the same abbreviation combination is broken down into the same combination in a preset number of manual confirmations (configurable based on data size and accuracy requirements), a combination word recognition rule is automatically generated. When the same keyword is manually confirmed as different standardized dimensions under different examination types, and the cumulative confirmations under each examination type are not less than a preset confirmation threshold, a disambiguation rule is automatically generated, and the difference between the manual review results and the automatic standardized results is recorded for periodic evaluation and adjustment of the confidence level threshold. The manually confirmed standardized results and the updated knowledge base and rule base are used to optimize subsequent standardized tasks without re-triggering the currently completed standardized process.

[0049] Compared with the prior art, the advantages of this application are as follows: The standardization accuracy is higher and there are fewer false matches. By first identifying the examination type and then identifying the other standardization dimensions under the contextual constraints of the examination type, cross-type keyword mismatches can be avoided. For example, "enhanced" is mapped to "enhanced" under the contextual constraints of the CT / MR examination type and to "contrast ultrasound" under the contextual constraints of the ultrasound examination type, which effectively reduces standardization errors caused by missing context. At the same time, the contextual constraints of the examination type reduce invalid string comparisons during the standardization process. Compared with the flat word segmentation matching of the prior art, this application can identify the inherent hierarchical structure of composite examination items and significantly reduce false positive and false negative results. The output retains multi-dimensional semantics, and the standardized data has stronger usability. The complex inspection items are broken down into four independent dimensions: inspection type, inspection method, inspection location, and observation variable. Each dimension supports the expression of multiple standardized dimensions, fully retaining the semantic information of the original items. The output results can be directly stored and dimension indexes can be built, supporting fast retrieval by dimension, grouped statistics, and cross-institutional data comparison, solving the problem that a single mapping cannot express complex semantics. The standardization process involves less computation and is more efficient in data processing. The constraints of the examination type narrow the matching range of examination methods, examination sites, and observation variables from the entire database to the subset corresponding to the current type. The computation of standardization matching for a single dimension is greatly reduced, and the processing time can be shortened for large-scale examination item dictionary batch standardization tasks that span multiple medical institutions or information systems. It can automatically process combined abbreviations, reducing manual maintenance costs; through the decomposition and parsing of combined words, it can automatically identify and map abbreviations in fields such as CTA, MRA, MRCP, and ERCP to the dimensions of the corresponding inspection type and inspection method. When the decomposition results of the same abbreviation are confirmed to be consistent after multiple reviews, combined word rules can be generated and added to the rule library, eliminating the need for manual maintenance of mapping relationships one by one. The standardized results have better logical consistency; the exclusion word conflict detection can automatically identify and handle mutually exclusive standardized dimensions, such as the mutual exclusion relationship between "plain scan" and "enhanced scan" in CT / MR, and make retention decisions according to preset priorities to avoid outputting standardized results with logical conflicts, thereby improving the reliability of downstream data analysis and quality control applications. The knowledge base is continuously updated to adapt to changes in terminology; the historical comparison table, multi-dimensional medical knowledge base and disambiguation rule base are iteratively updated based on the review feedback results. New terms and rules are gradually accumulated during use, thereby reducing long-term maintenance costs and possessing continuous self-optimization capabilities.

[0050] like Figure 2 As shown in the figure, this application discloses an intelligent multidimensional standardization system for medical examination items based on a knowledge base, including: The preprocessing module is used to preprocess the original examination item names across multiple medical institutions or information systems and match them with historical reference tables. If the match is successful, the historical standardized results are directly output; otherwise, subsequent standardization is performed. The structured splitting module is used to structurally split the names of the original check items that failed to match in the historical lookup table according to the preset delimiter syntax hierarchy priority, and obtain a semantic list containing the outer main name and the inner sub-items. The identification module is used to identify the examination type of the split semantic list based on a pre-built multi-dimensional medical knowledge base, and to split and parse medical abbreviations and combinations to supplement and correct the standardized dimensions of the examination type. The standardization module is used to call a subset of the multi-dimensional medical knowledge base corresponding to the examination type as a context constraint, and to complete the standardization matching of the three dimensions of examination method, examination site, and observed variables respectively. The disambiguation inspection module is used to perform context-differentiated disambiguation processing on polysemous keywords based on a preset disambiguation rule library, and at the same time, it completes mutual exclusion conflict detection and correction for each standardized dimension based on the exclusion word rule library. The confidence processing module is used to classify the confidence level based on the coverage and completeness of the standardized matching results, perform differential processing on results with different confidence levels, and output a formatted output containing four-dimensional standardized results including inspection type, inspection method, inspection location, and observed variables. The iteration module is used to review and update the historical comparison table, the multi-dimensional medical knowledge base, and the disambiguation rule base.

[0051] The system described in the above embodiments is used to execute the corresponding knowledge base-based intelligent multidimensional standardization method for medical examination items in the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0052] like Figure 3 As shown, based on the same application concept, and corresponding to any of the above embodiments, this application also discloses an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the above-mentioned intelligent multidimensional standardization method for medical examination items based on a knowledge base.

[0053] Specifically, the device includes a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected within the device via the bus 1050.

[0054] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), GPU (Graphics Processing Unit), or one or more integrated circuits, to implement relevant programs and achieve the technical solutions provided in the embodiments of this specification.

[0055] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 1020 and called by the processor 1010. The input / output interface 1030 is used to connect input / output modules to realize information input and output. Input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices can include keyboards, mice, touch screens, microphones, various sensors, etc., and output devices can include displays, projectors, speakers, vibrators, indicator lights, etc.

[0056] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB (Universal Serial Bus), network cable, etc.) or wireless means (such as mobile network, WIFI (Wireless Fidelity), Bluetooth, etc.).

[0057] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0058] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0059] The electronic devices described above are used to implement the corresponding knowledge-based intelligent multidimensional standardization method for medical examination items in any of the foregoing embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0060] Based on the same application concept, and corresponding to any of the above embodiments, this application also discloses a computer-readable storage medium that stores computer instructions for enabling a computer to implement the above-described intelligent multidimensional standardization method for medical examination items based on a knowledge base.

[0061] The computer-readable storage medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transfer medium, which can be used to store information accessible by a computing device. The computer instructions stored in the storage medium of the above embodiments are used to enable the computer to implement the intelligent multi-dimensional standardization method for medical examination items based on a knowledge base as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0062] The above description is merely a preferred embodiment and the technical principles employed in this application. This application is not limited to the specific embodiments or combinations thereof, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of this application. Therefore, although this application has been described in detail through the above embodiments, this application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of this application, the scope of which is determined by the scope of the claims.

Claims

1. A knowledge-based intelligent multidimensional standardization method for medical examination items, characterized in that, include: The original examination item names across multiple medical institutions or information systems are preprocessed and matched against historical lookup tables. If a match is successful, the historical standardized results are directly output; otherwise, subsequent standardization is performed. According to the preset delimiter syntax hierarchy priority, the names of the original check items that failed to match in the historical lookup table are structurally split to obtain a semantic list containing the outer main name and the inner sub-items. Based on a pre-built multi-dimensional medical knowledge base, the examination type is identified in the split semantic list, and the combination of abbreviations in the medical field is split and parsed to supplement and correct the standardized dimensions of the examination type. Using the identified examination type as a context constraint, a subset of the multi-dimensional medical knowledge base corresponding to the examination type is invoked to complete the standardized matching of the three dimensions of examination method, examination site, and observed variables respectively. Based on the preset disambiguation rule base, context-differentiated disambiguation processing is performed on polysemous keywords, and mutual exclusion conflict detection and correction of each of the standardized dimensions is completed based on the exclusion word rule base. The confidence level is graded based on the coverage completeness of the standardized matching results. Differential processing is performed on the results with different confidence levels, and a four-dimensional standardized result containing the inspection type, the inspection method, the inspection location, and the observed variable is output in a formatted manner. The historical comparison table, the multi-dimensional medical knowledge base, and the disambiguation rule base are reviewed and updated iteratively.

2. The intelligent multidimensional standardization method for medical examination items based on a knowledge base according to claim 1, characterized in that, The preprocessing and historical lookup table matching specifically include: Read the data of the original examination items from the examination item dictionaries of the respective HIS, PACS or RIS medical information systems of multiple medical institutions; Perform data cleaning operations on the original inspection items, including removing invalid special characters, standardizing the case of English characters, and standardizing the delimiter format. The names of the original inspection items after the data cleaning operation are matched precisely with the historical comparison table. If the match is successful, the historical standardized result is returned directly.

3. The intelligent multidimensional standardization method for medical examination items based on a knowledge base according to claim 1, characterized in that, The syntax hierarchy priority of the delimiters includes: The first priority is the bracket-type separator, which is used to extract the content within the brackets and distinguish the outer main name from the inner sub-item; The second priority is the plus sign separator, used to split multiple parallel inner sub-items within and / or outside the parentheses; The third priority is slashes, connectors, and coordinating conjunctions, used to separate independent semantic units under the same outer main name.

4. The intelligent multidimensional standardization method for medical examination items based on a knowledge base according to claim 1, characterized in that, The inspection type identification adopts the longest matching strategy and performs word boundary verification, giving priority to matching the synonyms with the longest character length in the multi-dimensional medical knowledge base; If no corresponding examination type is identified for the outer main name and all inner sub-items, the standardized matching of subsequent dimensions will be performed by loading the full multi-dimensional medical knowledge base and marking the corresponding results as low confidence.

5. The intelligent multidimensional standardization method for medical examination items based on a knowledge base according to claim 1, characterized in that, The decomposition and parsing of medical abbreviations and compound words is achieved through a compound word recognition rule base. The combined word recognition rule base stores a multidimensional mapping relationship between abbreviations and inspection types, inspection methods, and inspection locations, as well as regular expression patterns used to match variant expressions; If the inspection type has been identified, the word splitting and parsing only supplements and corrects the standardized dimensions of the inspection method; if the inspection type has not been identified, the word splitting and parsing supplements the standardized dimensions of the inspection type at the same time.

6. The intelligent multidimensional standardization method for medical examination items based on a knowledge base according to claim 1, characterized in that, The standardized matching of the three dimensions under the context constraints specifically includes: Load the multi-dimensional medical knowledge base subsets of the examination method, examination site, and observation variables corresponding to the current examination type, and perform keyword matching only within the subsets constrained by the context. If one of the vocabulary standardizations within the same inner-layer sub-project matches multiple optional standardization dimensions, then the priority is assigned according to the order of inspection method over inspection location, and inspection location over observed variable. If different words within the same inner sub-item are standardized and matched to different standardized dimensions, then the corresponding standardized dimensions are retained.

7. The intelligent multidimensional standardization method for medical examination items based on a knowledge base according to claim 1, characterized in that, The disambiguation rule base includes different standardized dimensions of the same polysemous keyword under different inspection types; The contextual differentiation disambiguation process includes selecting the standardized dimension corresponding to the polysemous keyword based on the determined inspection type, and marking it as pending manual review if there is no corresponding standardized dimension. The exclusion word rule base includes standardized dimensions of mutually exclusive inspection methods, inspection locations, and observed variables under contextual constraints of different inspection types; When mutually exclusive normalized dimensions are detected, the higher-priority normalized dimension is retained according to the preset priority and the mutual exclusion conflict information is recorded.

8. A knowledge-based intelligent multidimensional standardization system for medical examination items, characterized in that, include: The preprocessing module is used to preprocess the original examination item names across multiple medical institutions or information systems and match them with historical reference tables. If the match is successful, the historical standardized results are directly output; otherwise, subsequent standardization is performed. The structured splitting module is used to structurally split the names of the original check items that failed to match the historical lookup table according to the preset delimiter syntax hierarchy priority, so as to obtain a semantic list containing the outer main name and the inner sub-items. The identification module is used to identify the examination type of the split semantic list based on a pre-built multi-dimensional medical knowledge base, and to split and parse medical abbreviations and combinations to supplement and correct the standardized dimensions of the examination type. The standardization module is used to call a subset of the multi-dimensional medical knowledge base corresponding to the examination type as a context constraint, and to complete the standardization matching of the three dimensions of examination method, examination site, and observation variables respectively. The disambiguation inspection module is used to perform context-differentiated disambiguation processing on polysemous keywords based on a preset disambiguation rule library, and at the same time, to complete the mutual exclusion conflict detection and correction of each of the standardized dimensions based on the exclusion word rule library. The confidence processing module is used to classify the confidence level according to the coverage completeness of the standardized matching results, perform differential processing on different confidence level results, and output a formatted output containing a four-dimensional standardized result including the inspection type, the inspection method, the inspection location, and the observed variable. The iteration module is used to review and iteratively update the historical comparison table, the multi-dimensional medical knowledge base, and the disambiguation rule base.

9. An electronic device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the intelligent multidimensional standardization method for medical examination items based on a knowledge base as described in any one of claims 1-7.

10. A storage medium, which is a computer-readable storage medium, characterized in that, The system stores computer instructions for instructing a computer to implement the knowledge-based intelligent multidimensional standardization method for medical examination items as described in any one of claims 1-7.

Images