A surgical coding method, device, storage medium and equipment
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- WUHAN BRANCH OF ANHUI IFLYTEK MEDICAL INFORMATION TECH CO LTD
- Filing Date
- 2021-12-21
- Publication Date
- 2026-06-02
Smart Images
Figure CN114242223B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of intelligent medical technology, and in particular to a surgical coding method, apparatus, storage medium and device. Background Technology
[0002] With the development of medical technology, the types of diseases are increasing, making the standardization of non-standardized surgical names particularly important for achieving medical information sharing and data integration. In the medical field, surgical coding refers to the standardization of surgical names given by clinicians, providing a reliable basis for hospital statistics, scientific research, teaching, and medical information management.
[0003] Currently, the common method for surgical coding is to use a rule engine combined with manual identification. However, this rule engine plus manual identification coding method is highly subjective, difficult to quantify, has low coding efficiency, requires a lot of human resources, is difficult to promote on a large scale, and is not suitable for big data scenarios in the medical industry.
[0004] Therefore, how to achieve intelligent surgical coding to improve coding efficiency and accuracy is a technical problem that urgently needs to be solved. Summary of the Invention
[0005] The main objective of this application is to provide a surgical coding method, apparatus, storage medium, and device that can improve the accuracy of surgical coding results.
[0006] This application provides a surgical coding method, including:
[0007] Obtain the target surgical name to be encoded; and perform data cleaning on the target surgical name to obtain the cleaned target surgical name;
[0008] The cleaned target surgical name is segmented into words to obtain individual target words; and the preset category to which each target word belongs is determined.
[0009] Based on the preset category to which each target word belongs, candidate standard surgical names are selected from a pre-constructed standard corpus; the standard corpus stores a large number of standard surgical names and their preset category word content;
[0010] The target surgical name is matched with the candidate standard surgical name based on similarity, and the standard surgical name corresponding to the target surgical name is determined based on the matching result, which is used as the encoding result corresponding to the target surgical name.
[0011] In one possible implementation, the step of data cleaning the target surgical name to obtain the cleaned target surgical name includes:
[0012] Perform at least one of the following data cleaning operations on the target surgical name: special character processing, punctuation processing, number standardization processing, common name replacement, and synonym replacement, to obtain the cleaned target surgical name.
[0013] In one possible implementation, the cleaned target surgical name is segmented to obtain various target words, including:
[0014] Using a medical dictionary, the cleaned target surgical name is segmented using the Jieba word segmentation algorithm to obtain individual target words.
[0015] In one possible implementation, the step of selecting candidate standard surgical names from a pre-built standard corpus based on the preset categories to which each target word belongs includes:
[0016] Based on the preset category to which each target word belongs, M candidate standard surgical names are selected from a pre-constructed standard corpus; where M is a positive integer greater than 1.
[0017] Duplicate candidate standard surgical names among the M candidate standard surgical names are merged to obtain N candidate standard surgical names; where N is a positive integer not greater than M and greater than 0.
[0018] In one possible implementation, the step of performing similarity matching between the target surgical name and the candidate standard surgical names, and determining the standard surgical name corresponding to the target surgical name based on the matching result, as the encoding result corresponding to the target surgical name, includes:
[0019] Calculate the text edit distance between the target surgical name and the candidate standard surgical name;
[0020] Calculate the matching score of the candidate standard surgical name based on the text edit distance;
[0021] Based on the matching scores of all the candidate standard surgical names, the standard surgical name corresponding to the target surgical name is determined and used as the encoding result corresponding to the target surgical name.
[0022] In one possible implementation, determining the standard surgical name corresponding to the target surgical name based on the matching scores of all the candidate standard surgical names, as the encoding result corresponding to the target surgical name, includes:
[0023] When the matching scores of all the candidate standard surgical names include a candidate standard surgical name that reaches a preset score, the candidate standard surgical name corresponding to the preset score is determined as the standard surgical name corresponding to the target surgical name, and is used as the encoding result corresponding to the target surgical name.
[0024] In one possible implementation, determining the standard surgical name corresponding to the target surgical name based on the matching scores of all the candidate standard surgical names, as the encoding result corresponding to the target surgical name, includes:
[0025] When the matching scores of all the candidate standard surgical names do not include a candidate standard surgery that has reached the preset score, the order of the target words in the target surgical name is rearranged to obtain the updated target surgical name.
[0026] The updated target surgical name is matched with the candidate standard surgical name to obtain a matching score;
[0027] When all the matching scores include the preset score, the candidate standard surgery corresponding to the preset score is determined as the standard surgery name corresponding to the updated target surgery name, and is used as the encoding result corresponding to the target surgery name;
[0028] Alternatively, if the preset score is not included in all the matching scores, then the candidate standard surgery corresponding to the highest matching score among all the matching scores is determined as the standard surgery name corresponding to the updated target surgery name, and used as the encoding result corresponding to the target surgery name.
[0029] In one possible implementation, the preset category includes at least one of the following: surgical procedure, location, approach, disease nature, and implant.
[0030] This application also provides a surgical coding device, including:
[0031] An acquisition unit is used to acquire the target surgical name to be encoded; and to perform data cleaning on the target surgical name to obtain the cleaned target surgical name.
[0032] The word segmentation unit is used to segment the cleaned target surgical name into words to obtain individual target words; and to determine the preset category to which each target word belongs.
[0033] The filtering unit is used to filter candidate standard surgical names from a pre-built standard corpus according to the preset category to which each target word belongs; the standard corpus stores a large number of standard surgical names and their preset category contents;
[0034] The matching unit is used to perform fuzzy matching between the target surgical name and the candidate standard surgical name, and determine the standard surgical name corresponding to the target surgical name based on the matching result, which is used as the encoding result corresponding to the target surgical name.
[0035] In one possible implementation, the acquisition unit is specifically used for:
[0036] Perform at least one of the following data cleaning operations on the target surgical name: special character processing, punctuation processing, number standardization processing, common name replacement, and synonym replacement, to obtain the cleaned target surgical name.
[0037] In one possible implementation, the word segmentation unit is specifically used for:
[0038] Using a medical dictionary, the cleaned target surgical name is segmented using the Jieba word segmentation algorithm to obtain individual target words.
[0039] In one possible implementation, the filtering unit includes:
[0040] The filtering subunit is used to filter out M candidate standard surgical names from a pre-constructed standard corpus according to the preset category to which each target word belongs; where M is a positive integer greater than 1.
[0041] The merging subunit is used to merge duplicate candidate standard surgical names among the M candidate standard surgical names to obtain N candidate standard surgical names; where N is a positive integer not greater than M and greater than 0.
[0042] In one possible implementation, the matching unit includes:
[0043] The first calculation subunit is used to calculate the text edit distance between the target surgical name and the candidate standard surgical name;
[0044] The second calculation subunit is used to calculate the matching score of the candidate standard surgical name based on the text edit distance;
[0045] A subunit is defined to determine the standard surgical name corresponding to the target surgical name based on the matching scores of all the candidate standard surgical names, and to use this as the encoding result corresponding to the target surgical name.
[0046] In one possible implementation, the determining subunit is specifically used for:
[0047] When the matching scores of all the candidate standard surgical names include a candidate standard surgical name that reaches a preset score, the candidate standard surgical name corresponding to the preset score is determined as the standard surgical name corresponding to the target surgical name, and is used as the encoding result corresponding to the target surgical name.
[0048] In one possible implementation, the determining subunit includes:
[0049] The sorting subunit is used to rearrange the order of the target words in the target surgery name to obtain an updated target surgery name when the matching scores of all the candidate standard surgery names do not include a candidate standard surgery that has reached a preset score.
[0050] The matching subunit is used to perform similarity matching between the updated target surgical name and the candidate standard surgical name to obtain a matching score;
[0051] A subunit is obtained, which is used to determine the candidate standard surgery corresponding to the preset score as the standard surgery name corresponding to the updated target surgery name when all the matching scores contain the preset score, and use it as the encoding result corresponding to the target surgery name;
[0052] Alternatively, if the preset score is not included in all the matching scores, then the candidate standard surgery corresponding to the highest matching score among all the matching scores is determined as the standard surgery name corresponding to the updated target surgery name, and used as the encoding result corresponding to the target surgery name.
[0053] In one possible implementation, the preset category includes at least one of the following: surgical procedure, location, approach, disease nature, and implant.
[0054] This application also provides a surgical coding device, including: a processor, a memory, and a system bus;
[0055] The processor and the memory are connected via the system bus;
[0056] The memory is used to store one or more programs, the one or more programs including instructions, which, when executed by the processor, cause the processor to perform any of the above-described surgical coding methods.
[0057] This application also provides a computer-readable storage medium storing instructions that, when executed on a terminal device, cause the terminal device to perform any of the above-described surgical coding methods.
[0058] This application also provides a computer program product, which, when run on a terminal device, causes the terminal device to execute any of the above-described surgical coding methods.
[0059] This application provides a surgical coding method, apparatus, storage medium, and device. First, the target surgical name to be encoded is obtained. Then, the target surgical name is cleaned to obtain a cleaned target surgical name. Next, the cleaned target surgical name is segmented to obtain individual target words. The preset category to which each target word belongs is determined. Then, based on the preset category to which each target word belongs, candidate standard surgical names are selected from a pre-constructed standard corpus. The standard corpus stores a large number of standard surgical names and their corresponding preset categories. The target surgical name can then be matched with the candidate standard surgical names for similarity. Based on the matching result, the standard surgical name corresponding to the target surgical name is determined as the encoding result for the target surgical name. Because this application selects candidate standard surgical names that are more similar to the preset category to which the target surgical name belongs from a pre-constructed standard corpus, and then matches these with the target surgical name to be encoded for similarity, a more accurate encoding result is determined, thereby improving the accuracy of the surgical coding result. Attached Figure Description
[0060] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0061] Figure 1 A flowchart illustrating a surgical coding method provided in an embodiment of this application;
[0062] Figure 2 A classification diagram of data cleaning provided for embodiments of this application;
[0063] Figure 3 This is an example diagram illustrating the overall implementation of the surgical coding method provided in the embodiments of this application;
[0064] Figure 4 This is a schematic diagram of the composition of a surgical coding device provided in an embodiment of this application. Detailed Implementation
[0065] With the rapid development of information technology, regional medical interconnection and information sharing and data integration have become the main development trend of smart medical applications. Among them, it is particularly important to standardize the increasing number of non-standard surgical names so as to provide a reliable basis for hospital statistics, scientific research, teaching and medical information management and to achieve standardized surgical coding.
[0066] However, current surgical coding methods typically employ a combination of rule engines and manual identification. This rule-engine-plus-manual-identification approach is highly subjective, difficult to quantify, inefficient, and requires significant human resources, making it unsuitable for large-scale deployment and big data scenarios in the medical industry. Therefore, achieving intelligent surgical coding to improve efficiency and accuracy is a pressing technical challenge.
[0067] To address the aforementioned deficiencies, this application provides a surgical coding method. First, the target surgical name to be encoded is obtained; then, the target surgical name is cleaned to obtain a cleaned target surgical name; next, the cleaned target surgical name is segmented into words to obtain individual target words; and the preset category to which each target word belongs is determined. Then, based on the preset category to which each target word belongs, candidate standard surgical names are selected from a pre-constructed standard corpus. The standard corpus stores a large number of standard surgical names and their corresponding preset categories, allowing for similarity matching between the target surgical name and the candidate standard surgical names. Based on the matching results, the standard surgical name corresponding to the target surgical name is determined as the encoding result for the target surgical name. Therefore, because this embodiment selects candidate standard surgical names from a pre-constructed standard corpus that are more similar to the preset category to which the target surgical name belongs, and then performs similarity matching between these candidate and target surgical names to determine a more accurate encoding result, the accuracy of the surgical coding result is improved.
[0068] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0069] First Embodiment
[0070] See Figure 1 This is a flowchart illustrating a surgical coding method provided in this embodiment. The method includes the following steps:
[0071] S101: Obtain the target surgical name to be encoded; and perform data cleaning on the target surgical name to obtain the cleaned target surgical name.
[0072] In this embodiment, any surgical name encoded using this embodiment is defined as the target surgical name. It should be noted that this embodiment does not limit the language of the target surgical name; for example, the target surgical name can be a Chinese surgical name or an English surgical name, etc. This embodiment also does not limit the length of the target surgical name; for example, the target surgical name can be composed of multiple word segments, etc.
[0073] To improve the accuracy of the target surgical name encoding results, this embodiment first cleans the data after obtaining the target surgical name to be encoded, resulting in a cleaned target surgical name. This cleaned target surgical name contains more standardized text content that meets the naming requirements of standard surgical names, and is then used to execute the subsequent step S102.
[0074] One optional implementation method is that the process of "cleaning the target surgical name to obtain the cleaned target surgical name" in step S101 may include: performing at least one of the following data cleaning operations on the target surgical name: special character processing, punctuation processing, number unification processing, common name replacement, and synonym replacement, to obtain the cleaned target surgical name.
[0075] Specifically, in this implementation, to improve the accuracy of the encoding results obtained when subsequently encoding and matching the target surgical name, such as... Figure 2 As shown, the target surgical procedure requires the following eight cleaning steps:
[0076] (1) Perform special character processing on the target surgical name, that is, encode the target surgical name in UTF8 and remove characters that do not exist in the regular character encoding. For example, invisible characters such as carriage return and newline tab^M in the target surgical name can be deleted.
[0077] (2) Punctuation processing of the target surgical name. This is because there is a distinction between full-width and half-width characters in the input method, which may cause the two to be mixed together in the target surgical name. These characters do not have obvious meaning characteristics, but they will cause anomalies for computer matching. Therefore, this application proposes to use ASCII and Unicode for conversion to realize punctuation processing of the target surgical name.
[0078] (3) Standardize the numbers in the target surgical name. This is because the numbers in the input method are represented in various ways, such as Arabic numerals, Chinese numerals, Roman numerals, etc. In order to facilitate unified encoding, the numbers in the target surgical name need to be uniformly replaced according to the preset type, such as unifying the numbers in the target surgical name to Arabic numerals, etc.
[0079] (4) Replace the target surgical name with a common name. This is because doctors often record common names such as medical terms and medical names in the form of abbreviations or acronyms (e.g., URL, DBS) when writing. In order to facilitate subsequent coding, it is necessary to replace the target surgical name with a common name so that it can be processed by word segmentation in subsequent steps.
[0080] (5) Replace the target surgical name with synonyms. This is because doctors often record the same disease name with different words due to factors such as different regions and disciplines. For example, they may use "aneurysm" and "vascular malformation" to refer to the same disease, or "leukemia" and "blood cancer" to refer to the same disease. In order to facilitate subsequent coding, it is necessary to replace the target surgical name with synonyms. For example, replace "aneurysm" and "vascular malformation" with "aneurysm", and replace "leukemia" and "blood cancer" with "leukemia", so that word segmentation and category labeling can be performed in subsequent steps after replacement.
[0081] S102: Perform word segmentation on the cleaned target surgical name to obtain individual target words; and determine the preset category to which each target word belongs.
[0082] In this embodiment, the target surgical name to be encoded is obtained in step S101 and its data is cleaned. After obtaining the cleaned target surgical name, in order to quickly and accurately determine the encoding result of the target surgical name, the cleaned target surgical name can be further segmented to obtain the words contained in the cleaned target surgical name. Here, each word obtained by segmentation is defined as a target word, and the preset category to which each target word belongs is determined for subsequent step S103. The preset category is consistent with the preset categories contained in the pre-constructed standard corpus, and may specifically include at least one of surgical procedure, location, approach, disease nature, and implant. For a detailed description, please refer to the description of the standard corpus in subsequent step S103.
[0083] It should be noted that, in order to improve the accuracy of word segmentation, an optional implementation method is that the process of "segmenting the cleaned target surgical name to obtain each target word" in step S102 may include: using a medical dictionary, segmenting the cleaned target surgical name using the Jieba word segmentation algorithm to obtain each target word.
[0084] In this implementation, since the target surgical name belongs to a specific medical field, in order to improve the accuracy of word segmentation, a medical dictionary can be used as the basis for word segmentation. The enhanced Jieba word segmentation method is adopted to segment the cleaned target surgical name to obtain each target word. The specific word segmentation process is the same as the existing process and will not be described in detail here.
[0085] Among them, the Jieba word segmentation algorithm is one of the more complete word segmentation technologies currently available. It can achieve efficient word graph scanning based on a prefix dictionary and generate a directed acyclic graph (DAG) composed of all possible word formation scenarios of the characters in a sentence. It uses dynamic programming to find the maximum probability path and finds the maximum segmentation combination based on word frequency. For out-of-vocabulary words, it uses an HMM model based on the word formation capability of characters and employs the Viterbi algorithm.
[0086] For example: Figure 3 As shown, assuming the target surgical procedure is named "cerebellar aneurysm resection," after segmenting the word using a medical dictionary and the Jieba word segmentation algorithm, the resulting target words are "cerebellum," "aneurysm," and "resection." These three target words can then be categorized into preset categories: "cerebellum" belongs to "location," "aneurysm" belongs to "disease nature," and "resection" belongs to "surgical procedure." Figure 3 As shown in step ①.
[0087] S103: Based on the preset category to which each target word belongs, candidate standard surgical names are selected from the pre-constructed standard corpus; the standard corpus stores a large number of standard surgical names and their preset categories.
[0088] In this embodiment, to improve the accuracy of surgical coding results, a standard corpus can be pre-constructed. This standard corpus stores a large number of standard surgical names and their corresponding preset categories. That is, for each disease, this application will utilize the professional knowledge of doctors and combine it with the latest national and Beijing clinical standards of ICD-9-CM3 to store only one standard surgical name for the disease in the standard corpus. The standard surgical name will be segmented, and the resulting segments will be assigned to the corresponding preset categories (including surgical procedure, location, approach, disease nature, implant, and others).
[0089] For example: Figure 3As shown, taking the three existing standard surgical names "superior cerebellar artery aneurysm clipping," "cerebellar lesion resection," and "cerebral debridement" as examples, the three standard surgical names are segmented into words, and the resulting words are assigned to corresponding preset categories as follows: "clamping," "lesion resection," and "debridement" are assigned to the preset category "surgical procedure"; "cerebellum," "cerebellum," and "cerebrum" are assigned to the preset category "location"; and "aneurysm" is assigned to the preset category "disease nature," such as... Figure 3 As shown in step 0 of the diagram.
[0090] Alternatively, the ICD-9-CM3 surgical naming convention can be referenced. Each preset category is identified by a characteristic dimension of the data. Thus, for standard surgical names in the standard corpus, "location" and "operative procedure" are relatively common data characteristic dimensions. When storing, "location" and "operative procedure" can be used as dimensions to store standard surgical names and their corresponding standard codes.
[0091] For example, the standard corpus can store: "cerebellum": [["superior cerebellar aneurysm clipping", "39.5108", "National Clinical Version"], ["cerebellar lesion resection", "01.5900x043", "National Clinical Version"], ["cerebellar puncture", "01.0900x005", "National Clinical Version"]].
[0092] Based on this, after determining the preset category to which each target word in the target surgical name belongs through step S102, standard surgical names that are the same as the preset category to which each target word in the target surgical name belongs can be matched from the pre-built standard corpus as candidate standard surgical names for subsequent step S104.
[0093] For example: Based on the above examples, such as Figure 3 As shown, assuming the target surgical procedure is "cerebellar aneurysm resection," after determining the preset categories of the three target terms it contains—"cerebellum" belonging to "location," "aneurysm" belonging to "disease nature," and "resection" belonging to "surgical procedure"—it can further match the standard surgical names corresponding to "resection," "cerebellum," and "aneurysm" from the three preset categories of "surgical procedure," "location," and "disease nature" in the pre-constructed standard corpus. Specifically, in "location," "cerebellum" corresponds to "1. Supracerebellar aneurysm clipping" and "2. Cerebellar lesion resection," and in "disease nature," "aneurysm" corresponds to "1. Supracerebellar aneurysm clipping," as shown below. Figure 3 As shown in step ② of the process.
[0094] In one possible implementation of this application embodiment, in order to improve encoding efficiency, the implementation process of "selecting candidate standard surgical names from a pre-built standard corpus according to the preset category to which each target word belongs" in step S103 may specifically include: firstly, selecting M candidate standard surgical names from a pre-built standard corpus according to the preset category to which each target word belongs; where M is a positive integer greater than 1; then, merging the duplicate candidate standard surgical names among these M candidate standard surgical names to obtain N non-duplicate candidate standard surgical names; where N is a positive integer not greater than M and greater than 0.
[0095] For example: Based on the above examples, such as Figure 3 As shown, assuming the target surgical procedure is named "cerebellar aneurysm resection," and the candidate standard surgical names matched for "cerebellum" in the "location" section of the pre-constructed standard corpus are "1. Superior cerebellar aneurysm clipping" and "2. Cerebellar lesion resection," and the candidate standard surgical name matched for "aneurysm" in the "disease nature" section is also "1. Superior cerebellar aneurysm clipping," then the two duplicate "1. Superior cerebellar aneurysm clipping" can be merged to obtain two non-duplicate candidate standard surgical names: "1. Superior cerebellar aneurysm clipping" and "2. Cerebellar lesion resection," which are then used to execute subsequent step S104, such as... Figure 3 As shown in step ③.
[0096] S104: Perform similarity matching between the target surgical name and the candidate standard surgical names, and determine the standard surgical name corresponding to the target surgical name based on the matching results, which will be used as the encoding result corresponding to the target surgical name.
[0097] It should be noted that after determining the candidate standard surgical name through step S103, existing or future similarity calculation methods (such as the fuzzy matching algorithm of FuzzyWuzzy) can be used to perform similarity matching calculation between the target surgical name and the candidate standard surgical name. Based on the obtained matching results, the standard surgical name most similar to the target surgical name is determined as the encoding result corresponding to the target surgical name.
[0098] In one possible implementation of this application embodiment, the implementation process of step S104 may specifically include the following steps AC:
[0099] Step A: Calculate the text edit distance between the target surgical name and the candidate standard surgical name.
[0100] In this implementation, text edit distance is used to compare the similarity between the target surgical name and the candidate standard surgical name. Specifically, it refers to the minimum number of editing operations required to transform one string into the other. A larger text edit distance indicates greater dissimilarity between the two strings. Permitted editing operations include replacing one character with another, inserting a character, and deleting a character.
[0101] For example, the steps to calculate the text edit distance between the strings "kitttin" and "sitting" may include:
[0102] 1) -> kitttin->sitttin, perform one replacement operation, replacing k with s, and record the number of replacements as 1.
[0103] 2) `sitttin->sittin` performs a deletion operation, deleting one `t`, and records the deletion count as 1.
[0104] 3) `sittin->sitting`: Perform an insertion operation, add `g`, and record the insertion count as 1.
[0105] After the above editing operations, the text edit distance between the two strings "kitttin" and "sitting" is 1 for replacement, 1 for deletion, and 1 for insertion.
[0106] Step B: Calculate the matching score of the candidate standard surgical names based on the text edit distance.
[0107] After obtaining the text edit distance between the target surgical name and each candidate standard surgical name in step A, the matching score of each candidate standard surgical name can be calculated. The specific calculation formula is as follows:
[0108]
[0109] Where Score_Reg_i represents the matching score of the i-th candidate standard surgical name; N0 represents the total number of characters in the i-th candidate standard surgical name; DELETE represents the number of deletions included in the text edit distance; REPLACE represents the number of replacements included in the text edit distance; and NSERT represents the number of insertions included in the text edit distance.
[0110] Step C: Based on the matching scores of all candidate standard surgical names, determine the standard surgical name corresponding to the target surgical name, which will be used as the encoding result corresponding to the target surgical name.
[0111] In this implementation, after calculating the matching scores of all candidate standard surgical names, these matching scores can be further processed to determine the standard surgical name corresponding to the target surgical name based on the processing results, which is then used as the encoding result corresponding to the target surgical name.
[0112] Specifically, one possible implementation is that when the matching scores of all candidate standard surgical names include candidate standard surgical names that reach a preset score (the specific value can be set according to the actual situation, and this application does not limit it, for example, the highest matching score of 105 corresponding to the fuzzy matching algorithm can be selected), the candidate standard surgical name corresponding to the preset score can be determined as the standard surgical name corresponding to the target surgical name, and used as the encoding result corresponding to the target surgical name.
[0113] Alternatively, another possible implementation is that when none of the candidate standard surgical names in the matching scores reach a preset score, the order of the target words in the target surgical name can be rearranged to obtain an updated target surgical name. Then, following the execution process of steps A and B above, the updated target surgical name is used to replace the original target surgical name, and a similarity match is performed between the updated target surgical name and the candidate standard surgical names to obtain a matching score. Furthermore, if all matching scores include the preset score, the candidate standard surgical name corresponding to the preset score can be determined as the standard surgical name corresponding to the updated target surgical name and used as the encoding result corresponding to the target surgical name; or, if none of the matching scores include the preset score, the candidate standard surgical name corresponding to the highest matching score among all matching scores is determined as the standard surgical name corresponding to the updated target surgical name and used as the encoding result corresponding to the target surgical name.
[0114] For example: Based on the above examples, such as Figure 3 As shown, assuming the target surgical procedure is named "cerebellar aneurysm resection," and two candidate standard surgical names are selected from a pre-built standard corpus, namely "1. Superior cerebellar artery aneurysm clipping" and "2. Cerebellar lesion resection," then the matching scores for "1. Superior cerebellar artery aneurysm clipping" and "2. Cerebellar lesion resection" can be calculated using the above steps AC, resulting in matching scores of 71 and 67 respectively. Figure 3 As shown in step ④, the candidate standard surgical name "superior cerebellar artery aneurysm clipping" corresponding to the highest score of 71 can be used as the standard surgical name corresponding to the target surgical name "cerebellar aneurysm resection", that is, as the coding result corresponding to the target surgical name "cerebellar aneurysm resection".
[0115] In summary, the surgical coding method provided in this embodiment first obtains the target surgical name to be encoded; then, it cleans the target surgical name to obtain a cleaned target surgical name; next, it performs word segmentation on the cleaned target surgical name to obtain individual target words; and then determines the preset category to which each target word belongs. Following this, based on the preset category to which each target word belongs, candidate standard surgical names are selected from a pre-constructed standard corpus. The standard corpus stores a large number of standard surgical names and their corresponding preset categories, allowing for similarity matching between the target surgical name and the candidate standard surgical names. Based on the matching results, the standard surgical name corresponding to the target surgical name is determined as the encoding result for the target surgical name. Therefore, since this embodiment selects candidate standard surgical names from a pre-constructed standard corpus that are more similar to the preset category to which the target surgical name belongs, and then performs similarity matching between these candidate and target surgical names to determine a more accurate encoding result, the accuracy of the surgical coding result is improved.
[0116] Second Embodiment
[0117] This embodiment will introduce a surgical coding device; please refer to the above method embodiment for related content.
[0118] See Figure 4 This is a schematic diagram of the composition of a surgical coding device provided in this embodiment. The device 400 includes:
[0119] The acquisition unit 401 is used to acquire the target surgical name to be encoded; and to perform data cleaning on the target surgical name to obtain the cleaned target surgical name.
[0120] The word segmentation unit 402 is used to segment the cleaned target surgical name into words to obtain individual target words; and to determine the preset category to which each target word belongs.
[0121] The filtering unit 403 is used to filter candidate standard surgical names from a pre-built standard corpus according to the preset category to which each target word belongs; the standard corpus stores a large number of standard surgical names and their preset category contents;
[0122] The matching unit 404 is used to perform fuzzy matching between the target surgical name and the candidate standard surgical name, and determine the standard surgical name corresponding to the target surgical name based on the matching result, as the encoding result corresponding to the target surgical name.
[0123] In one implementation of this embodiment, the acquisition unit 401 is specifically used for:
[0124] Perform at least one of the following data cleaning operations on the target surgical name: special character processing, punctuation processing, number standardization processing, common name replacement, and synonym replacement, to obtain the cleaned target surgical name.
[0125] In one implementation of this embodiment, the word segmentation unit 402 is specifically used for:
[0126] Using a medical dictionary, the cleaned target surgical name is segmented using the Jieba word segmentation algorithm to obtain individual target words.
[0127] In one implementation of this embodiment, the filtering unit 403 includes:
[0128] The filtering subunit is used to filter out M candidate standard surgical names from a pre-constructed standard corpus according to the preset category to which each target word belongs; where M is a positive integer greater than 1.
[0129] The merging subunit is used to merge duplicate candidate standard surgical names among the M candidate standard surgical names to obtain N candidate standard surgical names; where N is a positive integer not greater than M and greater than 0.
[0130] In one implementation of this embodiment, the matching unit 404 includes:
[0131] The first calculation subunit is used to calculate the text edit distance between the target surgical name and the candidate standard surgical name;
[0132] The second calculation subunit is used to calculate the matching score of the candidate standard surgical name based on the text edit distance;
[0133] A subunit is defined to determine the standard surgical name corresponding to the target surgical name based on the matching scores of all the candidate standard surgical names, and to use this as the encoding result corresponding to the target surgical name.
[0134] In one implementation of this embodiment, the determining subunit is specifically used for:
[0135] When the matching scores of all the candidate standard surgical names include a candidate standard surgical name that reaches a preset score, the candidate standard surgical name corresponding to the preset score is determined as the standard surgical name corresponding to the target surgical name, and is used as the encoding result corresponding to the target surgical name.
[0136] In one implementation of this embodiment, the determining subunit includes:
[0137] The sorting subunit is used to rearrange the order of the target words in the target surgery name to obtain an updated target surgery name when the matching scores of all the candidate standard surgery names do not include a candidate standard surgery that has reached a preset score.
[0138] The matching subunit is used to perform similarity matching between the updated target surgical name and the candidate standard surgical name to obtain a matching score;
[0139] A subunit is obtained, which is used to determine the candidate standard surgery corresponding to the preset score as the standard surgery name corresponding to the updated target surgery name when all the matching scores contain the preset score, and use it as the encoding result corresponding to the target surgery name;
[0140] Alternatively, if the preset score is not included in all the matching scores, then the candidate standard surgery corresponding to the highest matching score among all the matching scores is determined as the standard surgery name corresponding to the updated target surgery name, and used as the encoding result corresponding to the target surgery name.
[0141] In one implementation of this embodiment, the preset category includes at least one of the following: surgical procedure, location, approach, disease nature, and implant.
[0142] Furthermore, embodiments of this application also provide a surgical coding device, including: a processor, a memory, and a system bus;
[0143] The processor and the memory are connected via the system bus;
[0144] The memory is used to store one or more programs, the one or more programs including instructions that, when executed by the processor, cause the processor to perform any of the above-described surgical coding methods.
[0145] Furthermore, embodiments of this application also provide a computer-readable storage medium storing instructions that, when executed on a terminal device, cause the terminal device to perform any of the above-described surgical coding methods.
[0146] Furthermore, this application embodiment also provides a computer program product, which, when run on a terminal device, causes the terminal device to execute any of the above-described surgical coding methods.
[0147] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that all or part of the steps in the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network communication device such as a media gateway, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.
[0148] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.
[0149] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0150] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A surgical coding method, characterized in that, include: Obtain the target surgical name to be encoded; and perform data cleaning on the target surgical name to obtain the cleaned target surgical name; The cleaned target surgical name is segmented into words to obtain individual target words; and the preset category to which each target word belongs is determined; the preset category includes surgical procedure, location, approach, disease nature, and implant. Based on the preset category to which each target word belongs, candidate standard surgical names are selected from a pre-constructed standard corpus; the standard corpus stores a large number of standard surgical names and their preset category word content; Calculate the text edit distance between the target surgical name and the candidate standard surgical name; the text edit distance is used to compare the similarity between the two strings corresponding to the target surgical name and the candidate standard surgical name. Based on the text edit distance, the matching score of the candidate standard surgical name is calculated; wherein, the matching score of the i-th candidate standard surgical name is calculated as follows: the total number of characters in the i-th candidate standard surgical name is subtracted from the sum of the minimum number of deletions, minimum number of replacements, and minimum number of insertions required to convert the i-th candidate standard surgical name into the target surgical name, and the difference is obtained by dividing the difference by the total number of characters in the i-th candidate standard surgical name; The candidate standard surgery that achieves the preset score is determined as the standard surgery name corresponding to the target surgery name. Alternatively, when the matching score does not reach the preset score, the order of the target words in the target surgical name is rearranged to obtain an updated target surgical name. Then, the text edit distance between the updated target surgical name and the candidate standard surgical name is calculated to obtain an updated text edit distance. Based on the updated text edit distance, the updated matching score of the candidate standard surgical name is calculated to determine the standard surgical name corresponding to the updated target surgical name, which is used as the encoding result corresponding to the target surgical name.
2. The method according to claim 1, characterized in that, The step of cleaning the target surgical name to obtain the cleaned target surgical name includes: Perform at least one of the following data cleaning operations on the target surgical name: special character processing, punctuation processing, number standardization processing, common name replacement, and synonym replacement, to obtain the cleaned target surgical name.
3. The method according to claim 1, characterized in that, The process involves segmenting the cleaned target surgical name into words to obtain individual target words, including: Using a medical dictionary, the cleaned target surgical name is segmented using the Jieba word segmentation algorithm to obtain individual target words.
4. The method according to claim 1, characterized in that, The step of selecting candidate standard surgical names from a pre-built standard corpus based on the preset categories to which each target word belongs includes: Based on the preset category to which each target word belongs, M candidate standard surgical names are selected from a pre-constructed standard corpus; where M is a positive integer greater than 1. Duplicate candidate standard surgical names among the M candidate standard surgical names are merged to obtain N candidate standard surgical names; where N is a positive integer not greater than M and greater than 0.
5. The method according to claim 1, characterized in that, When the matching score does not reach the preset score, the order of the target words in the target surgical name is rearranged, the text edit distance between the updated target surgical name and the candidate standard surgical name is calculated, and the updated text edit distance is obtained. Then, based on the updated text edit distance, the updated matching score of the candidate standard surgical name is calculated to determine the standard surgical name corresponding to the updated target surgical name, which is used as the encoding result corresponding to the target surgical name, including: The candidate standard surgery corresponding to the updated matching score that reaches the preset score is determined as the standard surgery name corresponding to the updated target surgery name; or, when the preset score is not included in all the updated matching scores, the candidate standard surgery corresponding to the highest matching score among all the updated matching scores is determined as the standard surgery name corresponding to the updated target surgery name, and used as the encoding result corresponding to the target surgery name.
6. A surgical coding device, characterized in that, include: The acquisition unit is used to acquire the name of the target surgery to be encoded. The target surgery name is then cleaned to obtain the cleaned target surgery name. The word segmentation unit is used to segment the cleaned target surgical name into words to obtain individual target words. And determine the preset category to which each target term belongs; the preset category includes surgical procedure, location, approach, disease nature, and implant; The filtering unit is used to filter candidate standard surgical names from a pre-built standard corpus according to the preset category to which each target word belongs; The standard corpus stores a large number of standard surgical names and their corresponding preset categories; The matching unit is used to perform fuzzy matching between the target surgical name and the candidate standard surgical name, and determine the standard surgical name corresponding to the target surgical name based on the matching result, as the encoding result corresponding to the target surgical name; The matching unit includes: The first calculation subunit is used to calculate the text edit distance between the target surgical name and the candidate standard surgical name; the text edit distance is used to compare the similarity between the two strings corresponding to the target surgical name and the candidate standard surgical name. The second calculation subunit is used to calculate the matching score of the candidate standard surgical name based on the text edit distance; wherein, the matching score of the i-th candidate standard surgical name is calculated as follows: the total number of characters in the i-th candidate standard surgical name is subtracted from the sum of the minimum number of deletions, minimum number of replacements and minimum number of insertions required to convert the i-th candidate standard surgical name into the target surgical name, and the difference is obtained by dividing the difference by the total number of characters in the i-th candidate standard surgical name; A subunit is defined to determine the candidate standard surgery corresponding to the matching score that reaches the preset score as the standard surgery name corresponding to the target surgery name; or, when the matching score does not reach the preset score, the order of the target words in the target surgery name is rearranged to obtain the updated target surgery name, and then the updated text edit distance between the updated target surgery name and the candidate standard surgery name is calculated to obtain the updated text edit distance. Then, based on the updated text edit distance, the updated matching score of the candidate standard surgery name is calculated to determine the standard surgery name corresponding to the updated target surgery name, which is used as the encoding result corresponding to the target surgery name.
7. A surgical coding device, characterized in that, include: Processor, memory, system bus; The processor and the memory are connected via the system bus; The memory is used to store one or more programs, the one or more programs including instructions that, when executed by the processor, cause the processor to perform the method according to any one of claims 1-5.
8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed on a terminal device, cause the terminal device to perform the method described in any one of claims 1-5.