A data analysis system and method for quality control of the incidence of unreasonable duplication of medical records

Abstract

The present invention discloses a data analysis system and method for quality control of the incidence of unreasonable duplication of medical records, belonging to the technical field of medical record quality control. Based on punctuation features and pathological description feature words, a coding library is established; and electronic medical records of users are collected periodically; based on punctuation features, directed combination pairs are generated; through directed combination pairs, text paragraphs in electronic medical records are identified to generate text paragraph sets; with directed combination pairs as the first guide for unreasonable duplication, the first incidence of unreasonable duplication of electronic medical records is analyzed and calculated; with pathological description feature words as the second guide for unreasonable duplication, the second incidence of unreasonable duplication of electronic medical records is analyzed and calculated; through coordinate points and iterative analysis, different users are clustered for quality control, and quality control clustering results are output; irregular behaviors such as unreasonable copying and pasting of medical records can be clustered and represented, and a unified trade-off idea from fuzzy to concrete is adopted to improve the interconnection and interoperability of quality control.

Description

Technical Field

[0001] The present invention relates to the technical field of medical record quality control, and in particular to a data analysis system and method for quality control of the incidence of unreasonable duplication of medical records. Background Art

[0002] With the advancement of medical information technology, electronic medical record systems have become an essential tool for medical staff. Electronic medical records use relevant electronic equipment to save, manage, transmit and reproduce patients' medical process records, and have the functions of simplicity, clarity, speed and information sharing. Compared with traditional paper medical records, electronic medical records solve the shortcomings of storage, information retrieval, slow writing speed and illegible handwriting.

[0003] However, with the popularization of electronic medical records and the rapid increase in the number of patients visiting the hospital, unreasonable copying and pasting of medical records and other irregular behaviors are becoming more and more common, becoming a key factor affecting the quality of hospital medical records and medical care. In the existing technology, unreasonable copying and pasting of medical records and other behaviors are generally identified through technical means such as natural language processing. When comparing medical records, based on weighing the situation where the chief complaint is exactly the same, the patient's disease process is exactly the same, the treatment outside the hospital is the same, the patient's cause of illness is the same, and even the punctuation marks are the same, it can be identified as erroneous copying and pasting;

[0004] In view of this, when supervising irregular behaviors such as unreasonable copying and pasting of medical records, it is necessary to consider a variety of medical record characteristics. The identification process is cumbersome, and it is difficult to unify the judgment situations, which affects the interconnection and interoperability of quality control. Summary of the invention

[0005] The purpose of the present invention is to provide a data analysis system and method for quality control of the incidence of unreasonable duplication of medical records, so as to solve the problems raised in the above-mentioned background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] A data analysis system for quality control of the incidence of unreasonable duplication of medical records, the system comprises: an archive module, a text processing module, an unreasonable duplication generation analysis module and a quality control analysis module;

[0008] The archive module establishes a coding library based on punctuation features and pathological description feature words; periodically collects the user's electronic medical records to generate medical record data clusters;

[0009] The text processing module generates directed combination pairs based on punctuation mark features; identifies and divides text paragraphs in the electronic medical record through the directed combination pairs to generate a text paragraph set;

[0010] The unreasonable duplication generation analysis module is used to analyze and calculate the first incidence rate of unreasonable duplication of electronic medical records by taking the directed combination pair as the first guide for unreasonable duplication generation; and to analyze and calculate the second incidence rate of unreasonable duplication of electronic medical records by taking the pathological description feature word as the second guide for unreasonable duplication generation;

[0011] The quality control analysis module is used to construct a two-dimensional coordinate system, generate coordinate points in the two-dimensional coordinate system based on the first incidence rate of unreasonable duplication and the second incidence rate of unreasonable duplication; based on the coordinate points, quality control clustering is performed on different users through iterative analysis, and the quality control clustering results are output.

[0012] Further, the archive module includes a feature encoding unit and a periodic data collection unit;

[0013] The feature encoding unit is used to uniformly encode punctuation marks and pathological description feature words, respectively, wherein each punctuation mark and each pathological description feature word has a unique encoding attribute; the rth punctuation mark is denoted as PI r , the fth pathological description feature word is recorded as W f ;

[0014] The periodic data collection unit is used to collect the user's electronic medical records after authorization, uniformly encode the user, and add a time tag to the electronic medical record based on the diagnosis time; the user code is used as the medical record data cluster label to generate the user's medical record data cluster, recorded as MR i ={mr i (t x )|x∈[1,Y]}, where MR i represents the medical record data cluster generated for the user coded as i, mr i (t x ) indicates additional time tag t x The electronic medical record of user i is coded as x, which represents the unified collection cycle number, and Y represents the current collection cycle number.

[0015] Further, the text processing module includes a directed combination pair generation unit and a text paragraph recognition unit;

[0016] The directed combination pair generating unit forms directed combination pairs of punctuation marks based on the number of punctuation mark types, and records any e-th directed combination pair of punctuation marks as CO e :PI r →PI m , where PI mrepresents the e-th punctuation mark, and r=m or r≠m; the directed combination pair is composed of any two punctuation marks, the two punctuation marks in the combination pair are repeated or not, and the arrangement order of any two punctuation marks determines the directed attribute of the combination pair;

[0017] The text paragraph recognition unit recognizes text paragraphs for each electronic medical record by punctuation marks, wherein a text paragraph is formed between any two consecutive punctuation marks in the electronic medical record; based on the directed combination pairs of punctuation marks, a combination pair label is added to the text paragraph, which is recorded as [mr i (t x ):CO e ]={TP a (CO e )|a∈[1,A]},where mr i (t x ):CO e Represents a combination of labels, [mr i (t x ):CO e ] indicates the combination of label mr i (t x ):CO e The set of text paragraphs, TP a (CO e ) represents the directed combination of CO e The ath text paragraph under the category, A represents the electronic medical record mr i (t x ) contains the total number of text paragraphs.

[0018] Further, the unreasonable duplication generation analysis module includes a first guidance analysis unit and a second guidance analysis unit;

[0019] The first guide analysis unit is used to generate the first guide for unreasonable duplication based on the directed combination pair, and to count the electronic medical record mr i (t x ) has a directed combination pair CO e The number of i (t x ):CO e ), analyze and calculate electronic medical records mr between different users i (t x ) is calculated as follows:

[0020]

[0021] In the formula, FIR irepresents the first incidence rate of unreasonable copying of the user numbered i, E represents the total number of directed combination pairs, and I represents the total number of users;

[0022] The second guide analysis unit is used to generate a second guide based on the pathological description feature word as an unreasonable copy, and to count the text paragraph set [mr i (t x ):CO e ]Inner text paragraph TP a (CO e ) contains pathological description feature word W f The number of f |TP a (CO e )|[mr i (t x ):CO e ]}, analyze and calculate the electronic medical record mr between different users i (t x ) is calculated as follows:

[0023]

[0024] In the formula, SIR i represents the second incidence rate of unreasonable copying of user numbered i, and F represents the total number of pathological description feature words.

[0025] Further, the quality control analysis module includes a quality control sample characterization unit and an iterative analysis unit;

[0026] The quality control sample characterization unit is used to establish a two-dimensional coordinate system, the horizontal axis independent variable of the two-dimensional coordinate system is the first incidence rate of unreasonable replication, and the vertical axis dependent variable of the two-dimensional coordinate system is the second incidence rate of unreasonable replication. Based on the first incidence rate of unreasonable replication and the second incidence rate of unreasonable replication of the user numbered i, the user numbered i is punctuated in the two-dimensional coordinate system, and the coordinate point is recorded as i: (FIR i , SIR i );

[0027] The iterative analysis unit is used to take the coordinate point i: (FIR i , SIR i ) is the quality control clustering center, and the users are clustered for quality control. The quality control clustering degree is calculated. The formula is as follows:

[0028]

[0029] In the formula, CD(i, j) represents the quality control clustering degree between user numbered i and user numbered j, FIR jDenote the first incidence rate of unreasonable replication of the user numbered j as SIR j Denote the second incidence rate of unreasonable replication of the user numbered j, where i≠j;

[0030] Let i = i + 1 and perform the first-layer iteration. Calculate the quality control clustering degree CD(i + 1, j) between the user numbered i + 1 and the user numbered j, where i + 1≠j. If CD(i + 1, j) < CD(i, j), then cluster the user numbered i and the user numbered j to obtain the clustering set of the user numbered i, denoted as C(i);

[0031] Let j = j + 1 and then perform the second-layer iteration. Calculate the quality control clustering degree CD(i, j + 1) between the user numbered i and the user numbered j + 1, where i≠j + 1, and return to the first-layer iteration to calculate the quality control clustering degree CD(i + 1, j + 1) between the user numbered i + 1 and the user numbered j + 1. If CD(i + 1, j + 1) < CD(i, j + 1), then cluster the user numbered i and the user numbered j + 1 and record the user numbered j + 1 into the clustering set C(i);

[0032] After all users except the user numbered i participate in the iterative analysis, output the clustering set C(i);

[0033] Until all users participate in the clustering analysis, output all clustering sets;

[0034] Sort each clustering set in descending order according to the number of users contained in the clustering set and send the sorting result to the staff.

[0035] A data analysis method for quality control of the incidence rate of unreasonable medical record replication. This method includes the following steps:

[0036] Step S100: Based on punctuation mark features and pathological description feature words, establish a coding library; periodically collect the electronic medical records of users to generate medical record data clusters;

[0037] Step S200: Generate directed combination pairs based on punctuation mark features; through the directed combination pairs, identify and divide the text paragraphs in the electronic medical records to generate a text paragraph set;

[0038] Step S300: Take the directed combination pairs as the first guide for the generation of unreasonable replication, analyze and calculate the first incidence rate of unreasonable replication of the electronic medical records; take the pathological description feature words as the second guide for the generation of unreasonable replication, analyze and calculate the second incidence rate of unreasonable replication of the electronic medical records;

[0039] Step S400: construct a two-dimensional coordinate system, generate coordinate points in the two-dimensional coordinate system based on the first incidence rate of unreasonable duplication and the second incidence rate of unreasonable duplication; based on the coordinate points, perform quality control clustering on different users through iterative analysis, and output the quality control clustering results.

[0040] Furthermore, the specific implementation process of step S100 includes:

[0041] Step S101: Uniformly encode punctuation marks and pathological description feature words respectively, wherein each punctuation mark and each pathological description feature word respectively has a unique encoding attribute; the rth punctuation mark is recorded as PI r , the fth pathological description feature word is recorded as W f ;

[0042] Step S102: After the user's authorization, collect the user's electronic medical records, uniformly code the user, and add a time tag to the electronic medical record based on the diagnosis time; use the user code as the medical record data cluster label to generate the user's medical record data cluster, recorded as MR i ={mr i (t x )|x∈[1,Y]}, where MR i represents the medical record data cluster generated for the user coded as i, mr i (t x ) indicates additional time tag t x The electronic medical record of user i is coded as x, which represents the unified collection cycle number, and Y represents the current collection cycle number.

[0043] Furthermore, the specific implementation process of step S200 includes:

[0044] Step S201: Based on the number of types of punctuation marks, a directed combination pair of punctuation marks is formed, and any e-th directed combination pair of punctuation marks is recorded as CO e :PI r →PI m , where PI m represents the e-th punctuation mark, and r=m or r≠m. The directed combination pair is composed of any two punctuation marks, the two punctuation marks in the combination pair are repeated or not, and the arrangement order of any two punctuation marks determines the directed attribute of the combination pair;

[0045] Step S202: Identify text paragraphs for each electronic medical record by punctuation marks, wherein a text paragraph is formed between any two consecutive punctuation marks in the electronic medical record; based on the directed combination pairs of punctuation marks, add a combination pair label to the text paragraph, denoted as [mr i (t x ):COe ]={TP a (CO e )|a∈[1,A]},where mr i (t x ):CO e Represents a combination of labels, [mr i (t x ):CO e ] indicates the combination of label mr i (t x ):CO e The set of text paragraphs, TP a (CO e ) represents the directed combination of CO e The ath text paragraph under the category, A represents the electronic medical record mr i (t x ) contains the total number of text paragraphs.

[0046] Furthermore, the specific implementation process of step S300 includes:

[0047] Step S301: Taking the directed combination pair as the first guide for the unreasonable duplication, the electronic medical record mr is counted i (t x ) has a directed combination pair CO e The number of i (t x ):CO e ), analyze and calculate electronic medical records mr between different users i (t x ) is calculated as follows:

[0048]

[0049] In the formula, FIR i represents the first incidence rate of unreasonable copying of the user numbered i, E represents the total number of directed combination pairs, and I represents the total number of users;

[0050] Step S302: Taking the pathological description feature words as the second guide for unreasonable duplication, statistically analyze the text paragraph set [mr i (t x ):CO e ]Inner text paragraph TP a (CO e ) contains pathological description feature word W f The number of f |TP a (CO e )|[mr i(t x ):CO e ]}, analyze and calculate the electronic medical record mr between different users i (t x ) is calculated as follows:

[0051]

[0052] In the formula, SIR i represents the second incidence rate of unreasonable copying of the user numbered i, and F represents the total number of pathological description feature words;

[0053] According to the above method, the present invention abandons the traditional natural language processing method of unreasonable copying and pasting of medical records and other irregular behaviors based on multiple medical record characteristics, and instead uses text paragraphs composed of punctuation marks to blur the distribution of pathological description feature words; through the directed combination pairs composed of punctuation marks, the first incidence rate of unreasonable copying of electronic medical records between different users is quantified, thereby forming the unreasonable copying and pasting of medical records and other irregular behavior characteristics of different users from the user's perspective, and the larger the first incidence rate, the more obvious the formal copying feature of the entire medical record; at the same time, the pathological description feature words are used to visualize the unreasonable copying and pasting of medical records and other irregular behavior characteristics of different users in each text paragraph, and the larger the second incidence rate, the more obvious the formal copying feature of the medical record fragment.

[0054] Furthermore, the specific implementation process of step S400 includes:

[0055] Step S401: Establish a two-dimensional coordinate system, the horizontal axis independent variable of the two-dimensional coordinate system is the first incidence rate of unreasonable copying, and the vertical axis dependent variable of the two-dimensional coordinate system is the second incidence rate of unreasonable copying. Based on the first incidence rate of unreasonable copying and the second incidence rate of unreasonable copying of the user numbered i, the user numbered i is punctuated in the two-dimensional coordinate system, and the coordinate point is recorded as i: (FIR i , SIR i );

[0056] According to the above method, the present invention can express the characteristics of unreasonable copying and pasting of medical records and other irregular behaviors of different users from the dual dimensions of article to paragraph, and characterize the coordinate points of the users, so that the data processing architecture from fuzzy to concrete can be adopted, so that the characteristics of unreasonable copying and pasting of medical records and other irregular behaviors of different users based on the user as the starting point can be accurately clustered and represented, which is convenient for the rapid tracing and behavior process restoration of unreasonable copying and pasting of medical records and other irregular behaviors, and the contexts under unreasonable copying and pasting of medical records and other irregular behaviors can be uniformly weighed, and the cluster representation can improve the interconnection and interoperability of quality control;

[0057] Step S402: Using the coordinate point i: (FIR i , SIR i ) as the quality control clustering center, perform quality control clustering on users and calculate the quality control clustering degree. The formula is as follows:

[0058]

[0059] In the formula, CD(i, j) represents the quality control clustering degree between the user numbered i and the user numbered j. FIR j represents the first incidence rate of unreasonable replication of the user numbered j, and SIR j represents the second incidence rate of unreasonable replication of the user numbered j, where i ≠ j;

[0060] Let i = i + 1 and perform the first-layer iteration. Calculate the quality control clustering degree CD(i + 1, j) between the user numbered i + 1 and the user numbered j, where i + 1 ≠ j. If CD(i + 1, j) < CD(i, j), then cluster the user numbered i and the user numbered j to obtain the clustering set of the user numbered i, denoted as C(i);

[0061] Let j = j + 1 and then perform the second-layer iteration. Calculate the quality control clustering degree CD(i, j + 1) between the user numbered i and the user numbered j + 1, where i ≠ j + 1, and return to the first-layer iteration to calculate the quality control clustering degree CD(i + 1, j + 1) between the user numbered i + 1 and the user numbered j + 1. If CD(i + 1, j + 1) < CD(i, j + 1), then cluster the user numbered i and the user numbered j + 1 and record the user numbered j + 1 into the clustering set C(i);

[0062] After all users except the user numbered i participate in the iterative analysis, output the clustering set C(i);

[0063] Until all users participate in the clustering analysis, output all clustering sets;

[0064] Sort each clustering set in descending order according to the number of users included in the clustering set, and send the sorting result to the staff.

[0065] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: in a data analysis system and method for quality control of the incidence of unreasonable duplication of medical records provided by the present invention, a coding library is established based on punctuation features and pathological description feature words; and electronic medical records of users are collected periodically; directed combination pairs are generated based on punctuation features; text paragraphs in electronic medical records are identified through directed combination pairs to generate text paragraph sets; the first incidence of unreasonable duplication of electronic medical records is analyzed and calculated using the directed combination pairs as the first guide for unreasonable duplication; the second incidence of unreasonable duplication of electronic medical records is analyzed and calculated using the pathological description feature words as the second guide for unreasonable duplication; quality control clustering is performed on different users through coordinate points and iterative analysis, and quality control clustering results are output; irregular behaviors such as unreasonable copying and pasting of medical records can be clustered and represented, and a unified trade-off idea from fuzzy to concrete is adopted to improve the interconnection and interoperability of quality control. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0067] Figure 1 It is a structural schematic diagram of a data analysis system for quality control of the incidence of unreasonable duplication of medical records according to the present invention;

[0068] Figure 2 It is a schematic diagram of the steps of a data analysis method for quality control of the incidence of unreasonable duplication of medical records according to the present invention. DETAILED DESCRIPTION

[0069] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0070] See also Figure 1 , in the first embodiment: a data analysis system for quality control of the incidence of unreasonable duplication of medical records is provided, the system comprising: an archive library module, a text processing module, an unreasonable duplication generation analysis module and a quality control analysis module;

[0071] The archive module establishes a coding library based on punctuation features and pathological description feature words; periodically collects users' electronic medical records and generates medical record data clusters;

[0072] Preferably, the archive module includes a feature encoding unit and a periodic data collection unit;

[0073] The feature coding unit is used to uniformly code punctuation marks and pathological description feature words, wherein each punctuation mark and each pathological description feature word has a unique coding attribute; the rth punctuation mark is denoted as PI r , the fth pathological description feature word is recorded as W f ;

[0074] The periodic data collection unit is used to collect the user's electronic medical records after authorization, uniformly encode the user, and add a time tag to the electronic medical record based on the diagnosis time; the user code is used as the medical record data cluster label to generate the user's medical record data cluster, recorded as MR i ={mr i (t x )|x∈[1,Y]}, where MR i represents the medical record data cluster generated for the user coded as i, mr i (t x ) indicates additional time tag t x The electronic medical record of the user coded as i, x represents the unified collection cycle number, and Y represents the current collection cycle number;

[0075] The text processing module generates directed combination pairs based on punctuation mark features; through the directed combination pairs, the text paragraphs in the electronic medical record are identified and divided to generate a text paragraph set;

[0076] Preferably, the text processing module includes a directed combination pair generation unit and a text paragraph recognition unit;

[0077] The directed combination pair generating unit forms directed combination pairs of punctuation marks based on the number of punctuation marks, and any directed combination pair of punctuation marks of the eth type is recorded as CO e :PI r →PI m , where PI m represents the e-th punctuation mark, and r=m or r≠m; a directed combination pair is composed of any two punctuation marks, the two punctuation marks in the combination pair may be repeated or not, and the order of arrangement of any two punctuation marks determines the directed attribute of the combination pair;

[0078] The text paragraph recognition unit recognizes the text paragraphs of each electronic medical record through punctuation marks, wherein a text paragraph is formed between any two consecutive punctuation marks in the electronic medical record; based on the directed combination pairs of punctuation marks, a combination pair label is added to the text paragraph, which is recorded as [mr i (t x ):CO e ]={TP a (CO e )|a∈[1,A]},where mri (t x ):CO e Represents a combination of labels, [mr i (t x ): cO e ] indicates the combination of label mr i (t x ):CO e The set of text paragraphs, TP a (CO e ) represents the directed combination of CO e The ath text paragraph under the category, A represents the electronic medical record mr i (t x ) contains the total number of text paragraphs;

[0079] The unreasonable duplication generation analysis module is used to analyze and calculate the first incidence rate of unreasonable duplication of electronic medical records by taking the directed combination pair as the first guide for unreasonable duplication generation; and to analyze and calculate the second incidence rate of unreasonable duplication of electronic medical records by taking the pathological description feature word as the second guide for unreasonable duplication generation;

[0080] Preferably, the unreasonable duplication generation analysis module includes a first guide analysis unit and a second guide analysis unit;

[0081] The first guide analysis unit is used to generate the first guide for unreasonable duplication with directed combination pairs, and to count the electronic medical record mr i (t x ) has a directed combination pair CO e The number of i (t x ):CO e ), analyze and calculate electronic medical records mr between different users i (t x ) is calculated as follows:

[0082]

[0083] In the formula, FIR i represents the first incidence rate of unreasonable copying of the user numbered i, E represents the total number of directed combination pairs, and I represents the total number of users;

[0084] The second guide analysis unit is used to generate the second guide with the pathological description feature words as the unreasonable copy, and to count the text paragraph set [mr i (t x ):CO e ]Inner text paragraph TP a (CO e ) contains pathological description feature word Wf The number of f |TP a (CO e )|[mr i (t x ):CO e ]}, analyze and calculate the electronic medical record mr between different users i (t x ) is calculated as follows:

[0085]

[0086] In the formula, SIR i represents the second incidence rate of unreasonable copying of the user numbered i, and F represents the total number of pathological description feature words;

[0087] The quality control analysis module is used to construct a two-dimensional coordinate system, generate coordinate points in the two-dimensional coordinate system based on the first incidence rate of unreasonable duplication and the second incidence rate of unreasonable duplication; based on the coordinate points, perform quality control clustering on different users through iterative analysis, and output the quality control clustering results;

[0088] Preferably, the quality control analysis module includes a quality control sample characterization unit and an iterative analysis unit;

[0089] The quality control sample characterization unit is used to establish a two-dimensional coordinate system. The horizontal axis independent variable of the two-dimensional coordinate system is the first incidence rate of unreasonable replication, and the vertical axis dependent variable of the two-dimensional coordinate system is the second incidence rate of unreasonable replication. Based on the first incidence rate and the second incidence rate of unreasonable replication of the user numbered i, the user numbered i is punctuated in the two-dimensional coordinate system, and the coordinate point is recorded as i: (FIR i , SIR i );

[0090] Iterative analysis unit, used to take coordinate point i: (FIR i , SIR i ) is the quality control clustering center, and the users are clustered for quality control. The quality control clustering degree is calculated. The formula is as follows:

[0091]

[0092] In the formula, CD(i, j) represents the quality control clustering degree between user numbered i and user numbered j, FIR j represents the first occurrence rate of unreasonable copying of user number j, SIR j represents the second occurrence rate of unreasonable copying of user numbered j, i≠j;

[0093] Let \(i = i + 1\) and perform the first - layer iteration. Calculate the quality - control clustering degree \(CD(i + 1,j)\) between the user numbered \(i + 1\) and the user numbered \(j\), where \(i+1\neq j\). If \(CD(i + 1,j)<CD(i,j)\), then cluster the user numbered \(i\) and the user numbered \(j\) to obtain the clustering set of the user numbered \(i\), denoted as \(C(i)\).

[0094] Let \(j = j + 1\), and then perform the second - layer iteration. Calculate the quality - control clustering degree \(CD(i,j + 1)\) between the user numbered \(i\) and the user numbered \(j + 1\), where \(i\neq j + 1\), and return to the first - layer iteration to calculate the quality - control clustering degree \(CD(i + 1,j + 1)\) between the user numbered \(i + 1\) and the user numbered \(j + 1\). If \(CD(i + 1,j + 1)<CD(i,j + 1)\), then cluster the user numbered \(i\) and the user numbered \(j + 1\), and record the user numbered \(j + 1\) into the clustering set \(C(i)\).

[0095] When all users except the user numbered \(i\) have participated in the iterative analysis, output the clustering set \(C(i)\).

[0096] Until all users have participated in the clustering analysis, output all clustering sets.

[0097] Sort each clustering set in descending order according to the number of users included in the clustering set, and send the sorting result to the staff.

[0098] Please refer to Figure 2 , in the second embodiment: Provide a data analysis method for controlling the incidence of unreasonable medical record copying. The method includes the following steps:

[0099] Step S100: Based on punctuation - mark features and pathological - description feature words, establish a coding library; periodically collect users' electronic medical records to generate medical - record data clusters;

[0100] Exemplarily, uniformly encode punctuation marks and pathological - description feature words respectively. Among them, each punctuation mark and each pathological - description feature word have unique coding attributes; denote the \(r\) - th punctuation mark as \(PI\) r , and denote the \(f\) - th pathological - description feature word as \(W\) f ;

[0101] After obtaining user authorization, collect users' electronic medical records, uniformly encode users, and attach a time tag to the electronic medical records based on the diagnosis time; use the user code as the label of the medical - record data cluster to generate the medical - record data cluster of the user, denoted as \(MR\) i =\(\{mr\) i (t x )|x\in[1,Y]\}\), where \(MR\) i represents the medical - record data cluster corresponding to the user encoded as \(i\), \(mr\)i (t x ) indicates additional time tag t x The electronic medical record of the user coded as i, x represents the unified collection cycle number, and Y represents the current collection cycle number;

[0102] Step S200: generating directed combination pairs based on punctuation mark features; identifying and dividing text paragraphs in the electronic medical record through the directed combination pairs to generate a text paragraph set;

[0103] For example, based on the number of types of punctuation marks, directed combination pairs of punctuation marks are formed, and any e-th directed combination pair of punctuation marks is denoted as CO e :PI r →PI m , where PI m represents the e-th punctuation mark, and r=m or r≠m; a directed combination pair is composed of any two punctuation marks, the two punctuation marks in the combination pair may be repeated or not, and the order of arrangement of any two punctuation marks determines the directed attribute of the combination pair;

[0104] Through punctuation marks, text paragraph recognition is performed on each electronic medical record, where a text paragraph is formed between any two consecutive punctuation marks in the electronic medical record; based on the directed combination pairs of punctuation marks, a combination pair label is added to the text paragraph, denoted as [mr i (t x ):CO e ]={TP a (CO e )|a∈[1,A]},where mr i (t x ):CO e Represents a combination of labels, [mr i (t x ):CO e ] indicates the combination of label mr i (t x ):CO e The set of text paragraphs, TP a (CO e ) represents the directed combination of CO e The ath text paragraph under the category, A represents the electronic medical record mr i (t x ) contains the total number of text paragraphs;

[0105] Step S300: Analyze and calculate the first incidence rate of unreasonable duplication of electronic medical records by taking the directed combination pair as the first guide generated by unreasonable duplication; analyze and calculate the second incidence rate of unreasonable duplication of electronic medical records by taking the pathological description feature word as the second guide generated by unreasonable duplication;

[0106] For example, the first direction of the unreasonable duplication is generated by the directed combination pair, and the electronic medical record mr i (t x ) has a directed combination pair CO e The number of i (t x ):CO e ), analyze and calculate electronic medical records mr between different users i (t x ) is calculated as follows:

[0107]

[0108] In the formula, FIR i represents the first incidence rate of unreasonable copying of the user numbered i, E represents the total number of directed combination pairs, and I represents the total number of users;

[0109] Taking pathological description feature words as the second guide for unreasonable duplication, statistical text paragraph collection [mr i (t x ):CO e ]Inner text paragraph TP a (CO e ) contains pathological description feature word W f The number of f |TP a (CO e )|[mr i (t x ):CO e ]}, analyze and calculate the electronic medical record mr between different users i (t x ) is calculated as follows:

[0110]

[0111] In the formula, SIR i represents the second incidence rate of unreasonable copying of the user numbered i, and F represents the total number of pathological description feature words;

[0112] Step S400: constructing a two-dimensional coordinate system, generating coordinate points in the two-dimensional coordinate system based on the first incidence rate of unreasonable duplication and the second incidence rate of unreasonable duplication; performing quality control clustering on different users through iterative analysis based on the coordinate points, and outputting the quality control clustering results;

[0113] Exemplarily, a two-dimensional coordinate system is established. The independent variable of the abscissa of the two-dimensional coordinate system is the first incidence rate of unreasonable replication, and the dependent variable of the ordinate of the two-dimensional coordinate system is the second incidence rate of unreasonable replication. Then, based on the first incidence rate of unreasonable replication and the second incidence rate of unreasonable replication of the user numbered i, the user numbered i is marked in the two-dimensional coordinate system, and the coordinate point is denoted as i: (FIR i , SIR i );

[0114] Taking the coordinate point i: (FIR i , SIR i ) as the quality control clustering center, the users are subjected to quality control clustering, and the quality control clustering degree is calculated. The formula is as follows:

[0115]

[0116] In the formula, CD(i, j) represents the quality control clustering degree between the user numbered i and the user numbered j. FIR j represents the first incidence rate of unreasonable replication of the user numbered j, and SIR j represents the second incidence rate of unreasonable replication of the user numbered j, where i ≠ j;

[0117] Let i = i + 1, perform the first-layer iteration, calculate the quality control clustering degree CD(i + 1, j) between the user numbered i + 1 and the user numbered j, where i + 1 ≠ j. If CD(i + 1, j) < CD(i, j), then cluster the user numbered i and the user numbered j to obtain the clustering set of the user numbered i, denoted as C(i);

[0118] Let j = j + 1, then perform the second-layer iteration, calculate the quality control clustering degree CD(i, j + 1) between the user numbered i and the user numbered j + 1, where i ≠ j + 1, and return to the first-layer iteration to calculate the quality control clustering degree CD(i + 1, j + 1) between the user numbered i + 1 and the user numbered j + 1. If CD(i + 1, j + 1) < CD(i, j + 1), then cluster the user numbered i and the user numbered j + 1, and record the user numbered j + 1 into the clustering set C(i);

[0119] When all users except the user numbered i participate in the iterative analysis, output the clustering set C(i);

[0120] Until all users participate in the clustering analysis, output all clustering sets;

[0121] Sort the clustering sets in descending order according to the number of users included in the clustering sets, and send the sorting result to the staff.

[0122] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0123] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. 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. A data analysis method for quality control of the incidence of unreasonable duplication of medical records, characterized in that: The method includes the following steps: Step S100: Establish an encoding library based on punctuation features and pathological description feature words; periodically collect the electronic medical records of users to generate medical record data clusters; Step S200: Generate directed combination pairs based on punctuation features; through the directed combination pairs, identify and divide the text paragraphs in the electronic medical records to generate a text paragraph set; Step S300: Use the directed combination pairs as the first guide for unreasonable duplication, analyze and calculate the first incidence rate of unreasonable duplication of electronic medical records; use the pathological description feature words as the second guide for unreasonable duplication, analyze and calculate the second incidence rate of unreasonable duplication of electronic medical records; Step S400: Construct a two-dimensional coordinate system, generate coordinate points in the two-dimensional coordinate system based on the first incidence rate of unreasonable duplication and the second incidence rate of unreasonable duplication; based on the coordinate points, perform quality control clustering on different users through iterative analysis, and output the quality control clustering results; The specific implementation process of step S100 includes: Step S101: Uniformly encode punctuation marks and pathological description feature words respectively, wherein each punctuation mark and each pathological description feature word respectively has a unique encoding attribute; the rth punctuation mark is recorded as PI r , the fth pathological description feature word is recorded as W f ; Step S102: After the user's authorization, collect the user's electronic medical records, uniformly code the user, and add a time tag to the electronic medical record based on the diagnosis time; use the user code as the medical record data cluster label to generate the user's medical record data cluster, recorded as MR i ={mr i (t x )|x∈[1,Y]}, where MR i represents the medical record data cluster generated for the user coded as i, mr i (t x ) indicates additional time tag t x The electronic medical record of the user coded as i, x represents the unified collection cycle number, and Y represents the current collection cycle number; The specific implementation process of step S200 includes: Step S201: Based on the number of types of punctuation marks, a directed combination pair of punctuation marks is formed, and any e-th directed combination pair of punctuation marks is recorded as CO e :PI r →PI m , where PI m represents the e-th punctuation mark, and r=m or r≠m; the directed combination pair is composed of any two punctuation marks, the two punctuation marks in the combination pair are repeated or not, and the arrangement order of any two punctuation marks determines the directed attribute of the combination pair; Step S202: Identify text paragraphs for each electronic medical record by punctuation marks, wherein a text paragraph is formed between any two consecutive punctuation marks in the electronic medical record; based on the directed combination pairs of punctuation marks, add a combination pair label to the text paragraph, denoted as [mr i (t x ):CO e ]={TP a (CO e )|a∈[1,A]},where mr i (t x ):CO e Represents a combination of labels, [mr i (t x ):CO e ] indicates the combination of label mr i (t x ):CO e The set of text paragraphs, TP a (CO e ) represents the directed combination of CO e The ath text paragraph under the category, A represents the electronic medical record mr i (t x ) contains the total number of text paragraphs; The specific implementation process of step S300 includes: Step S301: Taking the directed combination pair as the first guide for the unreasonable duplication, the electronic medical record mr is counted i (t x ) has a directed combination pair CO e The number of i (t x ):CO e ), analyze and calculate electronic medical records mr between different users i (t x ) is calculated as follows: In the formula, FIR i represents the first incidence rate of unreasonable copying of the user numbered i, E represents the total number of directed combination pairs, and I represents the total number of users; Step S302: Taking the pathological description feature words as the second guide for unreasonable duplication, statistically analyze the text paragraph set [mr i (t x ):CO e ]Inner text paragraph TP a (CO e ) contains pathological description feature word W f The number of f |TP a (CO e )|[mr i (t x ):CO e ]}, analyze and calculate the electronic medical record mr between different users i (t x ) is calculated as follows: In the formula, SIR i represents the second incidence rate of unreasonable copying of the user numbered i, and F represents the total number of pathological description feature words; The specific implementation process of step S400 includes: Step S401: Establish a two-dimensional coordinate system, the horizontal axis independent variable of the two-dimensional coordinate system is the first incidence rate of unreasonable copying, and the vertical axis dependent variable of the two-dimensional coordinate system is the second incidence rate of unreasonable copying. Based on the first incidence rate of unreasonable copying and the second incidence rate of unreasonable copying of the user numbered i, the user numbered i is punctuated in the two-dimensional coordinate system, and the coordinate point is recorded as i: (FIR i , SIR i ); Step S402: Take coordinate point i: (FIR i , SIR i ) is the quality control clustering center, and the users are clustered for quality control. The quality control clustering degree is calculated. The formula is as follows: In the formula, CD(i, j) represents the quality control clustering degree between user numbered i and user numbered j, FIR j represents the first occurrence rate of unreasonable copying of user number j, SIR j represents the second occurrence rate of unreasonable copying of user numbered j, i≠j; Let i = i + 1, perform the first-layer iteration, calculate the quality control clustering degree CD(i + 1, j) between the user numbered i + 1 and the user numbered j, i + 1 ≠ j. If CD(i + 1, j) < CD(i, j), then cluster the user numbered i and the user numbered j to obtain the clustering set of the user numbered i, denoted as C(i); Let j = j + 1, then perform the second-layer iteration, calculate the quality control clustering degree CD(i, j + 1) between the user numbered i and the user numbered j + 1, i ≠ j + 1, and return to the first-layer iteration to calculate the quality control clustering degree CD(i + 1, j + 1) between the user numbered i + 1 and the user numbered j + 1. If CD(i + 1, j + 1) < CD(i, j + 1), then cluster the user numbered i and the user numbered j + 1, and record the user numbered j + 1 into the clustering set C(i); When all users except the user numbered i have participated in the iterative analysis, output the clustering set C(i); Until all users have participated in the clustering analysis, output all clustering sets; Sort each clustering set in descending order according to the number of users included in the clustering set, and send the sorting result to the staff.

2. A data analysis system for quality control of the incidence of unreasonable duplication of medical records, characterized in that: The system includes: an archive library module, a text processing module, an unreasonable duplication generation analysis module, and a quality control analysis module; The archive library module establishes an encoding library based on punctuation features and pathological description feature words; periodically collects the electronic medical records of users to generate medical record data clusters; The text processing module generates directed combination pairs based on punctuation features; through the directed combination pairs, identifies and divides the text paragraphs in the electronic medical records to generate a text paragraph set; The unreasonable replication generation analysis module is used to analyze and calculate the first incidence rate of unreasonable replication of electronic medical records with directed combination pairs as the first guidance for the generation of unreasonable replication; and analyze and calculate the second incidence rate of unreasonable replication of electronic medical records with pathological description feature words as the second guidance for the generation of unreasonable replication. The quality control analysis module is used to construct a two-dimensional coordinate system, generate coordinate points in the two-dimensional coordinate system based on the first incidence rate of unreasonable replication and the second incidence rate of unreasonable replication; and perform quality control clustering on different users by means of iterative analysis based on the coordinate points, and output the quality control clustering results. The archive module includes a feature encoding unit and a periodic data acquisition unit. The feature encoding unit is used to uniformly encode punctuation marks and pathological description feature words, respectively, wherein each punctuation mark and each pathological description feature word has a unique encoding attribute; the rth punctuation mark is denoted as PI r , the fth pathological description feature word is recorded as W f ; The periodic data collection unit is used to collect the user's electronic medical records after authorization, uniformly encode the user, and add a time tag to the electronic medical record based on the diagnosis time; the user code is used as the medical record data cluster label to generate the user's medical record data cluster, recorded as MR i ={mr i (t x )|x∈[1,Y]}, where MR i represents the medical record data cluster generated for the user coded as i, mr i (t x ) indicates additional time tag t x The electronic medical record of the user coded as i, x represents the unified collection cycle number, and Y represents the current collection cycle number; The text processing module includes a directed combination pair generation unit and a text paragraph recognition unit. The directed combination pair generating unit forms directed combination pairs of punctuation marks based on the number of punctuation mark types, and records any e-th directed combination pair of punctuation marks as CO e :PI r →PI m , where PI m represents the e-th punctuation mark, and r=m or r≠m; the directed combination pair is composed of any two punctuation marks, the two punctuation marks in the combination pair are repeated or not, and the arrangement order of any two punctuation marks determines the directed attribute of the combination pair; The text paragraph recognition unit recognizes text paragraphs for each electronic medical record by punctuation marks, wherein a text paragraph is formed between any two consecutive punctuation marks in the electronic medical record; based on the directed combination pairs of punctuation marks, a combination pair label is added to the text paragraph, which is recorded as [mr i (t x ):CO e ]={TP a (CO e )|a∈[1,A]},where mr i (t x ):CO e Represents a combination of labels, [mr i (t x ):CO e ] indicates the combination of label mr i (t x ):CO e The set of text paragraphs, TP a (CO e ) represents the directed combination of CO e The ath text paragraph under the category, A represents the electronic medical record mr i (t x ) contains the total number of text paragraphs; The unreasonable replication generation analysis module includes a first guidance analysis unit and a second guidance analysis unit. The first guide analysis unit is used to generate the first guide for unreasonable duplication based on the directed combination pair, and to count the electronic medical record mr i (t x ) has a directed combination pair CO e The number of i (t x ):CO e ), analyze and calculate electronic medical records mr between different users i (t x ) is calculated as follows: In the formula, FIR i represents the first incidence rate of unreasonable copying of the user numbered i, E represents the total number of directed combination pairs, and I represents the total number of users; The second guide analysis unit is used to generate a second guide based on the pathological description feature word as an unreasonable copy, and to count the text paragraph set [mr i (t x ):CO e ]Inner text paragraph TP a (CO e ) contains pathological description feature word W f The number of f |TP a (CO e )|[mr i (t x ):CO e ]}, analyze and calculate the electronic medical record mr between different users i (t x ) is calculated as follows: In the formula, SIR i represents the second incidence rate of unreasonable copying of the user numbered i, and F represents the total number of pathological description feature words; The quality control analysis module includes a quality control sample characterization unit and an iterative analysis unit. The quality control sample characterization unit is used to establish a two-dimensional coordinate system, the horizontal axis independent variable of the two-dimensional coordinate system is the first incidence rate of unreasonable replication, and the vertical axis dependent variable of the two-dimensional coordinate system is the second incidence rate of unreasonable replication. Based on the first incidence rate of unreasonable replication and the second incidence rate of unreasonable replication of the user numbered i, the user numbered i is punctuated in the two-dimensional coordinate system, and the coordinate point is recorded as i: (FIR i , SIR i ); The iterative analysis unit is used to take the coordinate point i: (FIR i , SIR i ) is the quality control clustering center, and the users are clustered for quality control. The quality control clustering degree is calculated. The formula is as follows: In the formula, CD(i, j) represents the quality control clustering degree between user numbered i and user numbered j, FIR j represents the first occurrence rate of unreasonable copying of user number j, SIR j represents the second occurrence rate of unreasonable copying of user numbered j, i≠j; Let i = i + 1, perform the first-layer iteration, calculate the quality control clustering degree CD(i + 1, j) between the user numbered i + 1 and the user numbered j, where i + 1 ≠ j. If CD(i + 1, j) < CD(i, j), then cluster the user numbered i and the user numbered j to obtain the clustering set of the user numbered i, denoted as C(i). Let j = j + 1, then perform the second-layer iteration, calculate the quality control clustering degree CD(i, j + 1) between the user numbered i and the user numbered j + 1, where i ≠ j + 1, and return to the first-layer iteration to calculate the quality control clustering degree CD(i + 1, j + 1) between the user numbered i + 1 and the user numbered j + 1. If CD(i + 1, j + 1) < CD(i, j + 1), then cluster the user numbered i and the user numbered j + 1, and record the user numbered j + 1 into the clustering set C(i). When all users except the user numbered i have participated in the iterative analysis, output the clustering set C(i). Until all users have participated in the clustering analysis, output all clustering sets. Sort each clustering set in descending order according to the number of users included in the clustering set, and send the sorting result to the staff.

Images