An automated coding method based on standard terminology for clinical data

Abstract

The present invention relates to the field of medical data processing, and in particular to an automated coding method based on clinical data standard terminology, comprising: determining a text reference state according to a comparison result of a text keyword recognition reference value and an unrecognized character interval reference value in a target text with a preset threshold value; determining a text analysis method according to the text reference state; in a text elimination analysis, determining whether to assign a standardized code to the target text according to a text matching state; in a text spelling analysis, determining replaceable keywords, performing a medical data classification value detection on the replaced target text, and determining whether the replaced target text is a valid text according to the medical data classification value and a preset medical data classification value range; replacing the initial target text with a valid text and assigning a standardized code; the text analysis method of the present invention can effectively analyze the target text, thereby improving text processing efficiency.

Description

Technical Field

[0001] The present invention relates to the field of medical data processing, and in particular to an automatic coding method based on clinical data standard terminology. Background Art

[0002] With the gradual advancement of medical informatization and the rapid development of data analysis technology, in order to effectively promote scientific and refined management of hospitals, control medical expenses and improve service quality, coders classify and encode the unstructured clinical data text recorded by clinicians in accordance with the national disease classification standard ICD (International Classification of Diseases), so that clinical data can be applied to various data analysis occasions. However, due to the large number of medical records generated in real time, the domestic coding volume is huge, and clinical data texts sometimes have problems such as illogicality, spelling errors, and semantic ambiguity, which brings many challenges to the automated coding process. Therefore, how to improve the problem of poor coding efficiency caused by clinical data text problems during the coding process is an urgent problem for technicians to solve.

[0003] Chinese patent publication number CN105069124A discloses an automated International Classification of Diseases coding method. The method includes: inputting Chinese disease diagnosis information; performing natural language processing on the Chinese disease diagnosis information to obtain one or more names to be coded; based on the standard terminology library and the extended terminology library, searching for standard terms or extended terms that match the names to be coded, and determining the codes of the successfully matched standard terms or extended terms as the codes of the names to be coded. It can be seen that the above technical solution can automatically complete ICD coding, but there are the following problems: the problem of poor coding efficiency caused by poor quality of clinical data text cannot be effectively solved. Summary of the invention

[0004] To this end, the present invention provides an automated coding method based on standard terminology for clinical data, so as to overcome the problem in the prior art that poor coding efficiency due to poor text quality of clinical data cannot be effectively solved.

[0005] To achieve the above object, the present invention provides an automated coding method based on standard clinical data terminology, comprising:

[0006] Determine the text reference state according to the comparison result of the text keyword recognition reference value and the unrecognized character interval reference value in the target text with the preset threshold value;

[0007] Determine the text analysis method based on the text reference status;

[0008] In the text elimination analysis, the text matching status is determined according to the matching results of the target text and the target text after elimination with the pre-stored ICD text, and whether to assign a standardized code to the target text is determined according to the text matching status;

[0009] In the text spelling analysis, the historical usage frequency of each unrecognized character in the unrecognized word in the target text is detected and replaceable keywords are determined under different replacement conditions. The medical data classification value is detected for the target text after the replacement, and it is determined whether the target text after the replacement is a valid text based on the medical data classification value and the preset medical data classification value range;

[0010] Replace the initial target text with the valid text and assign a standardized code;

[0011] Among them, the replaceable keywords determined under different replacement conditions are different, including recording the historical keywords with recognizable characters in the unrecognized words as the first replaceable keywords, or recording the historical text with the greatest similarity using the associated keywords corresponding to the unrecognized words in the historical text library as the reference text, and extracting the keywords in the reference text that do not exist in the target text as the second replaceable keywords.

[0012] Furthermore, a text analysis method is determined according to the text reference status, and the text analysis method includes text elimination analysis or text spelling analysis.

[0013] Furthermore, the text reference state is determined based on a text keyword recognition reference value and an unrecognized character interval reference value, and the text reference state includes that the text keyword recognition reference value is less than a preset recognition reference value and the unrecognized character interval reference value is greater than a preset unrecognized character interval reference value, or that the text keyword recognition reference value is less than the preset recognition reference value and the unrecognized character interval reference value is less than or equal to the preset unrecognized character interval reference value.

[0014] Furthermore, the text elimination analysis includes:

[0015] Eliminate unrecognized characters in the target text;

[0016] The target text and the eliminated target text are matched with each pre-stored ICD text for similarity;

[0017] The text matching status is detected, and whether to assign a standardized encoding to the target text is determined according to the text matching status.

[0018] Furthermore, the text matching status includes:

[0019] In the first preset text matching state, the target text and the target text after elimination are both successfully matched with the same pre-stored ICD text and only successfully matched with the same pre-stored ICD text;

[0020] A second preset text matching state, the target text and the target text after elimination are matched successfully with different pre-stored ICD texts, or the target text and / or the target text after elimination are not matched successfully;

[0021] Among them, in the first preset text matching state, a standardized code is assigned to the target text.

[0022] Furthermore, the text spelling analysis includes:

[0023] The historical usage frequency of each unrecognized character in the unrecognized word in the target text is detected, and if the first replacement condition is met, the unrecognized character with the smallest historical usage frequency is recorded as the character to be replaced;

[0024] Extract each historical keyword in the historical keyword library, record the historical keyword with recognizable characters in the unrecognized word as the first replaceable keyword, detect the spelling similarity between the characters to be replaced and the replaceable characters in the first replaceable keyword, and replace the unrecognized word with the first replaceable keyword corresponding to the replaceable character with the greatest spelling similarity;

[0025] The first replacement condition is that there is only a single unrecognized character in the unrecognized word and the historical usage frequency is less than a preset historical usage frequency.

[0026] Further, under the second replacement condition, the associated keywords of the unrecognized word are extracted, the historical text with the greatest similarity in the use of the associated keywords in the historical text library is recorded as the reference text, the keywords that do not exist in the target text are extracted from the reference text and recorded as the second replaceable keywords, the spelling similarity between the unrecognized word in the target text and each second replaceable keyword is detected, and the unrecognized word in the target text is replaced with the second replaceable keyword with the greatest spelling similarity;

[0027] The second replacement condition is that there are at least two unrecognized characters in the unrecognized word whose historical usage frequencies are less than a preset historical usage frequency.

[0028] Further, a medical data classification value detection is performed on the target text after the first replaceable keyword is replaced or the target text after the second replaceable keyword is replaced;

[0029] Medical data classification value detection includes:

[0030] Detect the pre-stored ICD text that matches successfully and has the greatest similarity to the target text after the replacement is completed;

[0031] Detecting the medical data classification value of the medical data corresponding to the target text after the replacement and the medical data corresponding to the pre-stored ICD text;

[0032] When the medical data classification value is within a preset medical data classification value range, the target text after replacement is determined to be a valid text, the initial target text is replaced with the valid text and assigned a standardized code.

[0033] Furthermore, the medical data classification value is determined based on the number of identical items, the degree of difference of identical item data, and the data collection interval.

[0034] Furthermore, the target text is a clinical medical diagnosis text uploaded by an online editor.

[0035] Compared with the prior art, the beneficial effect of the present invention lies in that the text reference state is determined according to the text keyword recognition reference value and the unrecognized character interval reference value in the technical scheme of the present invention, and the recognition difficulty of the target text and the distribution state of the unrecognized characters are effectively reflected through the text reference state, so that the subsequently selected text analysis method is more in line with the actual work scenario, and the text analysis method can effectively analyze the target text, thereby improving the text processing efficiency.

[0036] Furthermore, the present invention uses text elimination analysis to pre-match the target text and the target text after elimination with each pre-stored ICD text for similarity, and when the target text and the target text after elimination are both successfully matched with the same pre-stored ICD text and only successfully matched with the same pre-stored ICD text, a standardized code is assigned to the target text, which can avoid the problem of long data analysis time required due to the direct and unified use of text spelling analysis, thereby improving the data analysis speed of the present invention.

[0037] Furthermore, the present invention performs detection based on the historical usage frequency of each unrecognized character in the unrecognized word in the target text, and under the first replacement condition, detects the spelling similarity between the character to be replaced and the replaceable character in the first replaceable keyword, thereby reducing the impact of spelling errors on encoding efficiency, being able to quickly identify and correct spelling errors, thereby improving encoding efficiency and making the automated encoding process more efficient.

[0038] Furthermore, the present invention extracts related keywords of unrecognized words under the second replacement condition. To address the problem of a large number of unrecognized characters in unrecognized words, the unrecognized words are effectively replaced by detecting the spelling similarity between the unrecognized words in the target text and each second replaceable keyword, thereby further improving the accuracy of the automated coding method and enhancing the overall level of medical informationization.

[0039] Furthermore, the present invention performs medical data classification value detection on the target text after the first replaceable keyword is replaced or the target text after the second replaceable keyword is replaced, and when the medical data classification value is within a preset medical data classification value range, the target text after replacement is determined to be a valid text, and the medical data classification value is used to determine whether the target text is associated with the relevant disease information corresponding to the code to be assigned, so as to achieve accurate judgment of whether the target text is valid, thereby improving the encoding accuracy of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 is a schematic diagram of an automated coding method based on standard clinical data terminology of the present invention;

[0041] Figure 2 A flow chart of the present invention for determining a text analysis method according to a text reference state;

[0042] Figure 3 A flowchart of the present invention for determining whether to assign standardized coding to a target text according to a text matching state;

[0043] Figure 4 The present invention is a flow chart for determining replacement conditions according to the historical usage frequency of each unrecognized character in an unrecognized word. DETAILED DESCRIPTION

[0044] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0045] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0046] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0047] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0048] See also Figures 1 to 4 As shown, the present invention provides an automated coding method based on standard clinical data terminology, comprising:

[0049] Determine the text reference state according to the comparison result of the text keyword recognition reference value and the unrecognized character interval reference value in the target text with the preset threshold value;

[0050] Determine the text analysis method based on the text reference status;

[0051] In the text elimination analysis, the text matching status is determined according to the matching results of the target text and the target text after elimination with the pre-stored ICD text, and whether to assign a standardized code to the target text is determined according to the text matching status;

[0052] In the text spelling analysis, the historical usage frequency of each unrecognized character in the unrecognized word in the target text is detected and replaceable keywords are determined under different replacement conditions. The medical data classification value is detected for the target text after the replacement, and it is determined whether the target text after the replacement is a valid text based on the medical data classification value and the preset medical data classification value range;

[0053] Replace the initial target text with the valid text and assign a standardized code;

[0054] Among them, the replaceable keywords determined under different replacement conditions are different, including recording the historical keywords with recognizable characters in the unrecognized words as the first replaceable keywords, or recording the historical text with the greatest similarity using the associated keywords corresponding to the unrecognized words in the historical text library as the reference text, and extracting the keywords in the reference text that do not exist in the target text as the second replaceable keywords.

[0055] The application scenario of the method described in the present invention is the standardized coding of terminology of clinical medical diagnosis texts in a medical data center. The target text is the clinical medical diagnosis text edited and uploaded online by users of the medical data center. The clinical medical diagnosis text includes the name of the disease, symptom description, physical sign parameters and medication records; wherein the physical sign parameters are the names of physical characteristics related to the disease and the corresponding physical characteristic parameters recorded during the diagnosis process and the medical care process. For example, the name of the disease is gestational hypertension, and the physical sign parameters include blood pressure and specific blood pressure values ​​and urine protein and specific urine protein amounts. The name of the disease is hyperthyroidism, and the corresponding physical sign parameters include thyroid stimulating hormone and specific thyroid stimulating hormone values ​​and free thyroxine and specific free thyroxine values. The medication records include the names of the drugs used and the corresponding dosages. This is content that is easy for people in this field to understand, and is not described herein. Without going into details, the historical text library and the historical keyword library in the present invention are both storage modules for storing historical texts, wherein the historical texts are clinical medical diagnosis texts uploaded historically, and the historical keyword library is used to store historical keywords, wherein the historical keywords are identifiable keywords and the corresponding number of occurrences that appear in the clinical medical diagnosis texts uploaded historically by the medical data center. In addition, the method described in the present invention also requires the application of a standard library, which is used to store a number of standard keywords, a number of pre-stored ICD texts, and standard codes corresponding to the pre-stored ICD texts. The standard keywords are clinical medical diagnosis terms set by users for text recognition based on their medical experience and deep learning models, and the pre-stored ICD texts are clinical medical diagnosis texts set by users in the historical process that accurately correspond to the standard codes. The above are all common technical means in this field and will not be described here.

[0056] Specifically, the text analysis method is determined according to the text reference status, and the text analysis method includes text elimination analysis or text spelling analysis.

[0057] Specifically, the text reference state is determined based on a text keyword recognition reference value and an unrecognized character interval reference value, and the text reference state includes that the text keyword recognition reference value is less than a preset recognition reference value and the unrecognized character interval reference value is greater than a preset unrecognized character interval reference value, or that the text keyword recognition reference value is less than a preset recognition reference value and the unrecognized character interval reference value is less than or equal to a preset unrecognized character interval reference value.

[0058] The unrecognized characters are individual characters in an unrecognized word.

[0059] The text reference state includes a first preset text reference state and a second preset text reference state, the first preset text reference state being that the text keyword recognition reference value is less than the preset recognition reference value and the unrecognized character interval reference value is greater than the preset unrecognized character interval reference value;

[0060] The second preset text reference state is that the text keyword recognition reference value is less than the preset recognition reference value and the unrecognized character interval reference value is less than or equal to the preset unrecognized character interval reference value;

[0061] In the first preset text reference state, the text analysis method is text elimination analysis;

[0062] In the second preset text reference state, the text analysis method is text spelling analysis.

[0063] If the text keyword recognition reference value is greater than or equal to the preset recognition reference value, the target text is determined to be a valid text.

[0064] Text keyword recognition reference value = number of recognizable keywords in target text / (number of recognizable keywords in target text + number of unrecognized words). If a text entry in the target text is the same as a standard keyword in the standard library, the text entry is recorded as a recognizable keyword. If a text entry in the target text is different from all the standard keywords in the standard library, the text entry is recorded as an unrecognized word. The unrecognized word and at least one recognizable keyword are in the same unit sentence. The unit sentence is a single sentence separated by punctuation marks at both ends of the sentence. The confirmation method of the unrecognized character spacing reference value is to detect the maximum spacing distance corresponding to each unrecognized word in the target text, and record the maximum spacing distance with a larger value as the unrecognized character spacing reference value. The spacing distance is the distance between two unrecognized words. The number of characters in between, the preset recognition reference value and the preset unrecognized character spacing reference value can be set by the user according to the actual application scenario. The greater the user's demand for text recognition accuracy, the larger the value of the preset recognition reference value and the smaller the value of the preset unrecognized character spacing reference value. It can be understood by those skilled in the art that the larger the keyword recognition reference value, the greater the text recognition accuracy, and the smaller the unrecognized character spacing reference value, the smaller the impact of deleting unrecognized words on the overall semantic fluency. A preset recognition reference value and a preset unrecognized character spacing reference value are provided. The preset recognition reference value is 80% of the number of recognizable keywords + the number of unrecognized words in the target text, and the preset unrecognized character spacing reference value is 20% of the number of characters in the target text.

[0065] Specifically, the text elimination analysis includes:

[0066] Eliminate unrecognized characters in the target text;

[0067] The target text and the eliminated target text are matched with each pre-stored ICD text for similarity;

[0068] The text matching status is detected, and whether to assign a standardized encoding to the target text is determined according to the text matching status.

[0069] Specifically, the text matching status includes:

[0070] In the first preset text matching state, the target text and the target text after elimination are both successfully matched with the same pre-stored ICD text and only successfully matched with the same pre-stored ICD text;

[0071] The second preset text matching state is that the target text and the target text after elimination are successfully matched with different pre-stored ICD texts, or the target text and / or the target text after elimination are not successfully matched with any pre-stored ICD text;

[0072] Among them, in the first preset text matching state, a standardized code is assigned to the target text.

[0073] Similarity can be confirmed by:

[0074] Remove stop words and punctuation marks. The stop words are set by the user. The user can determine the words that have no practical meaning in text processing based on historical records and actual experience and set them as stop words, including but not limited to "maybe", "ah" and "suggestion". ;

[0075] Convert the text into a measurable numerical form. The present invention uses TF-IDF (Term Frequency-Inverse Document Frequency) technology to convert the text;

[0076] Use cosine similarity to calculate the cosine value of the angle between two text vectors to measure their directional similarity;

[0077] The pre-similarity is recorded as the similarity of the text required to be obtained, and the value range of the similarity is [0,1].

[0078] If the similarity between the two texts is greater than 0.8, it is determined that the two texts are matched successfully;

[0079] If the similarity between the two texts is less than or equal to 0.8, it is determined that the match between the two texts fails.

[0080] Specifically, the text spelling analysis includes:

[0081] The historical usage frequency of each unrecognized character in the unrecognized word in the target text is detected, and if the first replacement condition is met, the unrecognized character with the smallest historical usage frequency is recorded as the character to be replaced;

[0082] Extract each historical keyword in the historical keyword library, record the historical keyword with recognizable characters in the unrecognized word as the first replaceable keyword, detect the spelling similarity between the characters to be replaced and the replaceable characters in the first replaceable keyword, and replace the unrecognized word with the first replaceable keyword corresponding to the replaceable character with the greatest spelling similarity;

[0083] The first replacement condition is that each unrecognized word has only a single unrecognized character whose historical usage frequency is less than a preset historical usage frequency.

[0084] The historical usage frequency is the number of times the unrecognized character appears in all historical texts. The preset historical usage frequency can be set according to the actual application scenario. The greater the user's demand for text recognition encoding accuracy, the greater the value of the preset historical usage frequency. A value of the preset historical usage frequency is provided, and the preset historical usage frequency is 50% of the total number of historical texts. The spelling similarity between the character to be replaced and the replaceable character = the number of the same initials and finals of the character to be replaced and the replaceable character / the number of initials and finals of the replaceable character.

[0085] Specifically, under the second replacement condition, the associated keywords of the unrecognized words are extracted, the historical text with the greatest similarity of the associated keywords in the historical text library is recorded as the reference text, the keywords that do not exist in the target text are extracted from the reference text and recorded as the second replaceable keywords, the spelling similarity between the unrecognized words in the target text and each second replaceable keyword is detected, and the unrecognized words in the target text are replaced with the second replaceable keywords with the greatest spelling similarity;

[0086] The second replacement condition is that there are at least two unrecognized characters in the unrecognized word whose historical usage frequencies are less than a preset historical usage frequency.

[0087] Specifically, a medical data classification value detection is performed on the target text after the first replaceable keyword is replaced or the target text after the second replaceable keyword is replaced;

[0088] Medical data classification value detection includes:

[0089] Detect the pre-stored ICD text that matches successfully and has the greatest similarity to the target text after the replacement is completed;

[0090] Detecting the medical data classification value of the medical data corresponding to the target text after the replacement and the medical data corresponding to the pre-stored ICD text;

[0091] When the medical data classification value is within a preset medical data classification value range, the target text after replacement is determined to be a valid text, the initial target text is replaced with the valid text and assigned a standardized code.

[0092] Specifically, the medical data classification value is determined based on the number of identical items, the degree of difference of identical item data, and the data collection interval.

[0093] All recognizable keywords except unrecognized words in the target text are recorded as related keywords. The method for confirming the similarity of the use of related keywords between the target text and the target text is to detect the number of recognizable keywords in the historical text that are the same as the related keywords of the target text, recorded as M1, and the number of recognizable keywords in the historical text is recorded as M2. The similarity of the use of related keywords = M1 / (M1+M2). The classification value of medical data is H, which is determined according to the number of identical items, the difference of identical item data and the data collection interval. The calculation formula of H is:

[0094] H=1 / (H1 / H10×α+H2 / H20×β+H3 / H30×γ);

[0095] Among them, H1 is the number of identical items, H2 is the degree of data difference of identical items, H3 is the data collection interval, α is the first weight, β is the second weight, and γ is the third weight;

[0096] The number of identical items is the number of identical drug names and physical characteristic names in the physical sign parameters and medication records in the target text after the replacement and the physical sign parameters and medication records in the corresponding pre-stored ICD text, and H10 is the total number of drug names and physical characteristic names in the physical sign parameters and medication records in the target text after the replacement;

[0097] The method for confirming the difference degree of identical data items is to detect the range difference of the physical feature parameters corresponding to each identical physical feature name of the target text after replacement and the corresponding pre-stored ICD text. The difference degree of identical data items = the range difference with the largest value / the reference value of the identical item. For the range difference of the physical feature parameters corresponding to a single identical physical feature name, it is the absolute value of the difference between the maximum values ​​of the physical feature parameters corresponding to the physical feature names corresponding to the two texts. The reference value of the identical item is the maximum value of the physical feature parameters corresponding to the physical feature name corresponding to the target text. H20 is the conventional coefficient. H20 is the average value of the difference degree of identical data items in the coded records of the same type of physical feature names in the historical records that meet the user's recognition requirements.

[0098] The data collection interval is the time interval between the pre-stored ICD text and the initial target text collection. H30 is the preset time interval. The value of the preset time interval can be determined by the user according to the frequency and time of technical updates of the medical data center in the actual application scenario. It is understandable that clinical medical methods will change over time, which will lead to changes in the naming of diseases and the use of medications.

[0099] The values ​​of the first weight, the second weight and the third weight in the medical data classification value can be set by the user according to the actual application scenario. It can be understood that the user can learn from the historical records through the deep learning convolutional neural network to analyze the impact of the number of the same items, the difference in the data of the same items and the data collection interval on the text similarity, and then determine the value of each weight, and provide a first weight, a second weight and a third weight. The values ​​of the first weight, the second weight and the third weight are 40%, 40% and 20% respectively.

[0100] Specifically, the target text is the clinical medical diagnosis text uploaded by the online editor.

[0101] It is worth noting that in the present invention, only the pre-stored ICD text corresponding to the target text is determined, and the code of the disease name of the pre-stored ICD text is pre-coded to the target text or the valid text after the replacement is completed. The final code is determined by the user.

[0102] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0103] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An automated coding method based on standard clinical data terminology, characterized in that: include: Determine the text reference state according to the comparison result of the text keyword recognition reference value and the unrecognized character interval reference value in the target text with the preset threshold value; Determine the text analysis method based on the text reference status; In the text elimination analysis, the text matching status is determined according to the matching results of the target text and the target text after elimination with the pre-stored ICD text, and whether to assign a standardized code to the target text is determined according to the text matching status; In the text spelling analysis, the historical usage frequency of each unrecognized character in the unrecognized word in the target text is detected and replaceable keywords are determined under different replacement conditions. The medical data classification value is detected for the target text after the replacement, and it is determined whether the target text after the replacement is a valid text based on the medical data classification value and the preset medical data classification value range; Replace the initial target text with the valid text and assign a standardized code; Different replaceable keywords are determined under different replacement conditions, including recording the historical keywords with recognizable characters in the unrecognized word as the first replaceable keyword, or recording the historical text with the greatest similarity of the associated keywords corresponding to the unrecognized word in the historical text library as the reference text, and extracting the keywords that do not exist in the target text from the reference text as the second replaceable keyword; The text reference state is determined according to a text keyword recognition reference value and an unrecognized character interval reference value, the text reference state comprising a first preset text reference state in which the text keyword recognition reference value is less than a preset recognition reference value and the unrecognized character interval reference value is greater than a preset unrecognized character interval reference value, or a second preset text reference state in which the text keyword recognition reference value is less than the preset recognition reference value and the unrecognized character interval reference value is less than or equal to a preset unrecognized character interval reference value; In the first preset text reference state, the text analysis method is text elimination analysis; In the second preset text reference state, the text analysis method is text spelling analysis; The replacement conditions include a first replacement condition that there is only a single unrecognized character in the unrecognized word whose historical usage frequency is less than a preset historical usage frequency, and a second replacement condition that there are at least two unrecognized characters in the unrecognized word whose historical usage frequencies are less than a preset historical usage frequency; Under the first replacement condition, the unrecognized word is replaced with the first replaceable keyword corresponding to the replaceable character with the greatest spelling similarity; Under the second replacement condition, the unrecognized word in the target text is replaced by the second replaceable keyword with the greatest spelling similarity.

2. The automated coding method based on standard clinical data terminology according to claim 1, characterized in that: The text removal analysis includes: Eliminate unrecognized characters in the target text; The target text and the eliminated target text are matched with each pre-stored ICD text for similarity; The text matching status is detected, and whether to assign a standardized encoding to the target text is determined according to the text matching status.

3. The automated coding method based on standard clinical data terminology according to claim 2, characterized in that: The text matching status includes: In the first preset text matching state, the target text and the target text after elimination are both successfully matched with the same pre-stored ICD text and only successfully matched with the same pre-stored ICD text; A second preset text matching state, the target text and the target text after elimination are matched successfully with different pre-stored ICD texts, or the target text and / or the target text after elimination are not matched successfully; Among them, in the first preset text matching state, a standardized code is assigned to the target text.

4. The automated coding method based on standard clinical data terminology according to claim 1, characterized in that: The text spelling analysis includes: The historical usage frequency of each unrecognized character in the unrecognized word in the target text is detected, and if the first replacement condition is met, the unrecognized character with the smallest historical usage frequency is recorded as the character to be replaced; Extract each historical keyword in the historical keyword library, record the historical keyword with recognizable characters in the unrecognized word as the first replaceable keyword, detect the spelling similarity between the characters to be replaced and the replaceable characters in the first replaceable keyword, and replace the unrecognized word with the first replaceable keyword corresponding to the replaceable character with the greatest spelling similarity; The first replacement condition is that there is only a single unrecognized character in the unrecognized word and the historical usage frequency is less than a preset historical usage frequency.

5. The automated coding method based on standard clinical data terminology according to claim 4, characterized in that: Under the second replacement condition, extract the associated keywords of the unrecognized words, record the historical text with the greatest similarity of the associated keywords in the historical text library as the reference text, extract the keywords that do not exist in the target text from the reference text as the second replaceable keywords, detect the spelling similarity between the unrecognized words in the target text and each second replaceable keyword, and replace the unrecognized words in the target text with the second replaceable keywords with the greatest spelling similarity; The second replacement condition is that there are at least two unrecognized characters in the unrecognized word whose historical usage frequencies are less than a preset historical usage frequency.

6. The automated coding method based on standard clinical data terminology according to claim 5, characterized in that: Performing medical data classification value detection on the target text after the first replaceable keyword is replaced or the target text after the second replaceable keyword is replaced; Medical data classification value detection includes: Detect the pre-stored ICD text that matches successfully and has the greatest similarity to the target text after the replacement is completed; Detecting the medical data classification value of the medical data corresponding to the target text after the replacement and the medical data corresponding to the pre-stored ICD text; When the medical data classification value is within a preset medical data classification value range, the target text after replacement is determined to be a valid text, the initial target text is replaced with the valid text and assigned a standardized code.

7. The automated coding method based on standard clinical data terminology according to claim 4, characterized in that: The target text is clinical medical diagnosis text uploaded by online editors.

Images