Medical record data evaluation method and device, electronic equipment and storage medium

The method evaluates medical records by calculating similarities and keyword weights to address the limitations of existing quality control, providing accurate assessment of medical record quality considering content and individual differences.

CN119092026APending Publication Date: 2024-12-06BEIJING JIAHE HAISEN HEALTH TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410865789.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-28
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

Existing medical record quality control methods focus on form and structure, neglecting the substantive content and individual differences, leading to ineffective quality control and inaccurate reflection of patient conditions.

Method used

A method and device that evaluate medical records by calculating similarities with historical records, determining disease phenotypes, extracting keywords, and assigning weights based on probability, to assess the quality of medical records through keyword similarity scores.

Benefits of technology

Enables accurate evaluation of medical record quality by considering content logic and individual differences, ensuring effective quality control and precise reflection of patient conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119092026A_ABST
    Figure CN119092026A_ABST
Patent Text Reader

Abstract

The invention discloses a medical record data evaluation method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining a to-be-evaluated medical record; calculating the similarity between the to-be-evaluated medical record and each target historical medical record based on the feature information in the to-be-evaluated medical record and the target historical medical records; wherein the diagnosis results of the target historical medical record and the to-be-evaluated medical record are the same; based on the similarity between the to-be-evaluated medical record and the target historical medical record of each disease phenotype, determining the disease phenotype to which the to-be-evaluated medical record belongs; extracting keywords from the to-be-evaluated medical record; based on the weight of each keyword, calculating the similarity between the keyword of the to-be-assessed medical record and the keyword corresponding to the disease phenotype to which the to-be-assessed medical record belongs, and obtaining an assessment score of the to-be-assessed medical record; wherein the keyword corresponding to one disease phenotype is extracted from a plurality of historical medical records belonging to the disease phenotype in advance; the weight of the keyword corresponding to the disease phenotype is the probability that the keyword appears in the medical record belonging to the disease phenotype.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of data evaluation, and in particular to a medical record data evaluation method and device, electronic equipment, and storage medium. Background Art

[0002] Medical records objectively, completely and continuously record the changes in the patient's condition and the process of diagnosis and treatment. They are the basic data for scientific diagnosis and treatment in clinical practice, are also archives of medical science, and can be used to measure the level of doctors, etc. Therefore, medical record quality control is a very important task.

[0003] At present, the quality control of medical records mainly focuses on the form and structure of medical records. That is, it mainly uses pre-set rules or templates, and based on pre-set rules or templates, checks the structure of medical records, the integrity of the content, and the standardization of the writing format, so as to provide timely feedback on format problems such as missing content or non-standard descriptions.

[0004] However, the existing methods mainly control quality from the form and structure, without considering the content in the medical records, that is, they cannot control the logic of the content, the method of diagnosis and treatment, etc., while the substantive content of the medical records is the key to ensuring the quality of medical records, so the existing methods cannot achieve effective quality control of medical records. In addition, the existing methods are based on unified standard rules or templates for quality control, without considering individual differences, resulting in the inability to accurately reflect the individual's disease and treatment conditions, thus affecting the accurate control of medical record quality. Summary of the invention

[0005] Based on the above-mentioned deficiencies of the prior art, the present application provides a medical record data evaluation method and device, electronic equipment, and storage medium to solve the problem that the prior art cannot effectively perform medical record quality control.

[0006] In order to achieve the above objectives, this application provides the following technical solutions:

[0007] The first aspect of the present application provides a medical record data evaluation method, comprising:

[0008] Obtain medical records for evaluation;

[0009] Based on the characteristic information in the medical record to be evaluated and each target historical medical record, the similarity between the medical record to be evaluated and each target historical medical record is calculated; wherein the target historical medical record refers to a historical medical record in a medical record database having the same diagnosis result as the medical record to be evaluated;

[0010] Determining the disease phenotype to which the medical record to be evaluated belongs based on the similarity between the medical record to be evaluated and each of the target historical medical records belonging to each disease phenotype;

[0011] Extracting each keyword of the medical record to be evaluated from the medical record to be evaluated;

[0012] Based on the weight of each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, the similarity between each keyword of the medical record to be evaluated and each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs is calculated to obtain an evaluation score of the medical record to be evaluated; wherein, one type of each keyword corresponding to the disease phenotype is composed of keywords pre-extracted from various historical medical records belonging to the disease phenotype in the medical record database; one type of weight of the keyword corresponding to the disease phenotype is the probability of the keyword appearing in a medical record when a medical record belongs to the disease phenotype.

[0013] Optionally, in the above-mentioned medical record data evaluation method, before obtaining the medical record to be evaluated, the method further includes:

[0014] Acquire multiple historical medical records to form the medical record database;

[0015] Preprocessing each of the historical medical records;

[0016] Based on the diagnosis results and core processing items of each of the historical medical records, each of the historical medical records is divided into a plurality of the disease phenotypes, and the disease phenotype to which each of the historical medical records belongs is obtained.

[0017] Optionally, in the above-mentioned medical record data evaluation method, before obtaining the medical record to be evaluated, the method further includes:

[0018] Respectively extracting each keyword of each of the historical medical records in the medical record database;

[0019] Collecting and removing duplicates of the keywords of the historical medical records belonging to the same disease phenotype, respectively, to obtain keywords corresponding to each disease phenotype;

[0020] For each keyword corresponding to each disease phenotype, the conditional probability of the keyword appearing in the selected historical medical record under the condition that a historical medical record of the disease phenotype is randomly selected from the medical record database is calculated to obtain the weight of the keyword corresponding to the disease phenotype.

[0021] Optionally, in the above-mentioned medical record data evaluation method, the calculating the similarity between the medical record to be evaluated and each target historical medical record based on the feature information in the medical record to be evaluated and each target historical medical record includes:

[0022] Extracting each feature information of the medical record to be evaluated and each feature information of each target historical medical record from the medical record to be evaluated and each target historical medical record respectively;

[0023] For each target historical medical record, the number of identical feature information between the medical record to be evaluated and the target historical medical record is divided by the total number of all feature information between the medical record to be evaluated and the target historical medical record to obtain the similarity between the medical record to be evaluated and the target historical medical record.

[0024] Optionally, in the above-mentioned medical record data evaluation method, determining the disease phenotype to which the medical record to be evaluated belongs based on the similarity between the medical record to be evaluated and each of the target historical medical records belonging to each disease phenotype includes:

[0025] Calculating the average of the similarities between each of the target historical medical records belonging to each of the disease phenotypes and the medical records to be evaluated;

[0026] The disease phenotype corresponding to the maximum value of the mean value of the similarity with the medical record to be evaluated is determined as the disease phenotype to which the medical record to be evaluated belongs.

[0027] Optionally, in the above-mentioned medical record data evaluation method, the similarity between each keyword of the medical record to be evaluated and each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs is calculated based on the weight of each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, so as to obtain the evaluation score of the medical record to be evaluated, including:

[0028] Vectorizing the set of keywords of the medical record to be evaluated and the set of keywords corresponding to the target disease phenotype respectively to obtain the keyword vector of the medical record to be evaluated and the keyword vector of the target disease phenotype; wherein the target disease phenotype refers to the disease phenotype to which the medical record to be evaluated belongs;

[0029] Using the weights of the keywords corresponding to the target disease phenotype, weighted calculations are performed on the keyword vector of the medical record to be evaluated and the numerical values ​​corresponding to the keywords in the keyword vector of the target disease phenotype to obtain the weighted keyword vector of the medical record to be evaluated and the weighted keyword vector of the target disease phenotype;

[0030] The cosine similarity between the weighted keyword vector of the medical record to be evaluated and the weighted keyword vector of the target disease phenotype is calculated to obtain an evaluation score of the medical record to be evaluated.

[0031] A second aspect of the present application provides a medical record data evaluation device, comprising:

[0032] A first acquisition unit, used for acquiring the medical records to be evaluated;

[0033] A similarity calculation unit, configured to calculate the similarity between the medical record to be evaluated and each of the target historical medical records based on the feature information in the medical record to be evaluated and each of the target historical medical records; wherein the target historical medical record refers to a historical medical record in a medical record database having the same diagnosis result as the medical record to be evaluated;

[0034] A phenotype determination unit, configured to determine the disease phenotype to which the medical record to be evaluated belongs based on the similarity between the medical record to be evaluated and each of the target historical medical records belonging to each disease phenotype;

[0035] A first extraction unit, configured to extract each keyword of the medical record to be evaluated from the medical record to be evaluated;

[0036] A scoring calculation unit is used to calculate the similarity between each keyword of the medical record to be evaluated and each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, based on the weight of each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, so as to obtain an evaluation score of the medical record to be evaluated; wherein, one type of each keyword corresponding to the disease phenotype is composed of keywords pre-extracted from each historical medical record belonging to the disease phenotype in the medical record database; one type of weight of the keyword corresponding to the disease phenotype is the probability of the keyword appearing in a medical record when a medical record belongs to the disease phenotype.

[0037] Optionally, in the above-mentioned medical record data evaluation device, it also includes:

[0038] A second acquisition unit, used for acquiring a plurality of historical medical records to form the medical record database;

[0039] A preprocessing unit, used for preprocessing each of the historical medical records;

[0040] The phenotype classification unit is used to classify each of the historical medical records into a plurality of disease phenotypes based on the diagnosis results and core processing items of each of the historical medical records, so as to obtain the disease phenotype to which each of the historical medical records belongs.

[0041] Optionally, in the above-mentioned medical record data evaluation device, it also includes:

[0042] A second extraction unit is used to extract each keyword of each of the historical medical records in the medical record database;

[0043] A merging unit, used to respectively aggregate and remove duplicate keywords of the historical medical records belonging to the same disease phenotype, to obtain keywords corresponding to each disease phenotype;

[0044] The weight calculation unit is used to calculate, for each keyword corresponding to each disease phenotype, the conditional probability of the keyword appearing in the selected historical medical record under the condition that a historical medical record of the disease phenotype is randomly selected from the medical record database, so as to obtain the weight of the keyword corresponding to the disease phenotype.

[0045] Optionally, in the above-mentioned medical record data evaluation device, the similarity calculation unit includes:

[0046] A third extraction unit is used to extract each feature information of the medical record to be evaluated and each feature information of each target historical medical record from the medical record to be evaluated and each target historical medical record respectively;

[0047] The first calculation unit is used to divide the number of identical feature information between the medical record to be evaluated and the target historical medical record by the total number of all feature information between the medical record to be evaluated and the target historical medical record, respectively, to obtain the similarity between the medical record to be evaluated and the target historical medical record.

[0048] Optionally, in the above-mentioned medical record data evaluation device, the phenotype determination unit includes:

[0049] A mean calculation unit, used to calculate the mean of the similarities between each of the target historical medical records belonging to each of the disease phenotypes and the medical records to be evaluated;

[0050] The phenotype determination subunit is used to determine the disease phenotype corresponding to the maximum value of the mean value of the similarity of the medical record to be evaluated as the disease phenotype to which the medical record to be evaluated belongs.

[0051] Optionally, in the above-mentioned medical record data evaluation device, the score calculation unit includes:

[0052] A vector processing unit, used to vectorize the set of each keyword of the medical record to be evaluated and the set of each keyword corresponding to the target disease phenotype, respectively, to obtain the keyword vector of the medical record to be evaluated and the keyword vector of the target disease phenotype; wherein the target disease phenotype refers to the disease phenotype to which the medical record to be evaluated belongs;

[0053] A weighting unit, used to perform weighted calculations on the keyword vector of the medical record to be evaluated and the numerical values ​​corresponding to the keywords in the keyword vector of the target disease phenotype using the weights of the keywords corresponding to the target disease phenotype, to obtain the weighted keyword vector of the medical record to be evaluated and the weighted keyword vector of the target disease phenotype;

[0054] The second calculation unit is used to calculate the cosine similarity between the weighted keyword vector of the medical record to be evaluated and the weighted keyword vector of the target disease phenotype to obtain the evaluation score of the medical record to be evaluated.

[0055] A third aspect of the present application provides an electronic device, including:

[0056] Memory and processor;

[0057] Wherein, the memory is used to store programs;

[0058] The processor is used to execute the program, and when the program is executed, it is specifically used to implement the medical record data evaluation method as described in any one of the above.

[0059] A fourth aspect of the present application provides a computer storage medium for storing a computer program, which, when executed, is used to implement a medical record data evaluation method as described in any one of the above.

[0060] The present application provides a medical record data evaluation method, which pre-divides the disease phenotypes to which each historical medical record in the medical record database belongs, extracts keywords from each historical medical record, forms keywords corresponding to each disease phenotype, and calculates the weight of each keyword. Among them, the weight of a keyword corresponding to a disease phenotype is the probability of the keyword appearing in the medical record when a medical record belongs to the disease phenotype, which is used to reflect the importance of each keyword to the medical record of the disease phenotype. When the medical record needs to be evaluated, the medical record to be evaluated is obtained. Then, based on the characteristic information in the medical record to be evaluated and each target historical medical record, the similarity between the medical record to be evaluated and each target historical medical record is calculated, and based on the similarity between the medical record to be evaluated and each target historical medical record belonging to each disease phenotype, the disease phenotype to which the medical record to be evaluated belongs is determined. Then, each keyword of the medical record to be evaluated is extracted from the medical record to be evaluated, and based on the weight of each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, the similarity between each keyword of the medical record to be evaluated and each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs is calculated to obtain the evaluation score of the medical record to be evaluated. Since the keywords corresponding to the disease phenotype are the key contents that must be included in the medical records of a disease phenotype, the similarity between them and the keywords of the medical records to be evaluated can be calculated to determine the quality of the content of the medical records to be evaluated, so that the evaluation is no longer just based on the surface form, so the quality of the medical records can be effectively evaluated accurately. And because the evaluation is achieved by calculating the score through the similarity calculation of keywords, differences are allowed, that is, individual differences are allowed, so as to ensure accurate control of the quality of medical records. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0062] Figure 1 A flowchart of a medical record data evaluation method provided in an embodiment of the present application;

[0063] Figure 2 A flowchart of a method for calculating the similarity between a medical record to be evaluated and a target historical medical record provided in an embodiment of the present application;

[0064] Figure 3 A flowchart of a method for determining weights of keywords corresponding to a disease phenotype provided in an embodiment of the present application;

[0065] Figure 4 A flowchart of a method for calculating an evaluation score of a medical record to be evaluated provided in an embodiment of the present application;

[0066] Figure 5 A schematic diagram of the architecture of a medical record data evaluation device provided in an embodiment of the present application;

[0067] Figure 6 A schematic diagram of the architecture of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0068] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0069] In this application, relational terms such as first and second, etc. are used only 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 "comprise", "include" 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. In the absence of further restrictions, an element defined by the statement "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0070] The present application embodiment provides a medical record data evaluation method, such as Figure 1 As shown, the specific steps include:

[0071] S101. Obtain the medical records to be evaluated.

[0072] Optionally, the medical record to be evaluated can be a complete medical record. It can also be a sub-medical record of a complete medical record of a multi-diagnosis and treatment node disease, that is, a medical record of one of the diagnosis and treatment nodes. Among them, multi-diagnosis and treatment node diseases refer to complex diseases that require multiple evaluations and treatments, such as type 2 diabetes, tumor diseases, etc. These diseases often involve multiple organs or systems, and the situation is relatively complicated. Each diagnosis and treatment node may have its own specific tests, treatment plans and evaluation criteria.

[0073] S102: Calculate the similarity between the medical record to be evaluated and each target historical medical record based on the feature information in the medical record to be evaluated and each target historical medical record.

[0074] It should be noted that in the embodiment of the present application, the quality of the medical record to be evaluated is evaluated by comparing it with the content of the historical medical record, and only historical medical records of the same type can reflect whether the logic of the content of the medical record to be evaluated is consistent with the disease of the diagnosis result. Therefore, it is necessary to determine the historical medical records similar to the medical record to be evaluated in order to determine the type of the medical record to be evaluated. Since only the historical medical records that are consistent with the diagnosis results of the medical record to be evaluated are of reference value, the target historical medical record refers to the historical medical record in the medical record database that has the same diagnosis results as the medical record to be evaluated, and the target historical medical record has predetermined the disease phenotype to which it belongs. Therefore, based on the diagnosis results, multiple target historical medical records can be screened out from the medical record database, and then their similarity with the medical record to be evaluated can be calculated.

[0075] Optionally, feature information of a specified field may be extracted from the medical record to be evaluated and each target historical medical record, such as information such as diagnosis results, physical signs, and examination results. Then, based on calculating the similarity between the medical record to be evaluated and the target historical medical record in feature information, the similarity between the medical record to be evaluated and each target historical medical record is determined.

[0076] Optionally, in another embodiment of the present application, a specific implementation of step S102 is as follows: Figure 2 Shown include:

[0077] S201. Extract each feature information of the medical record to be evaluated and each feature information of each target historical medical record from the medical record to be evaluated and each target historical medical record respectively.

[0078] It should be noted that in the embodiment of the present application, the similarity between two medical records is determined by calculating the overlap of feature information, so it is necessary to extract the feature information in the medical records first.

[0079] S202. For each target historical medical record, divide the number of identical feature information between the medical record to be evaluated and the target historical medical record by the total number of all feature information between the medical record to be evaluated and the target historical medical record to obtain the similarity between the medical record to be evaluated and the target historical medical record.

[0080] Specifically, the characteristic information set of the medical record to be evaluated and the target historical medical record can be intersected to obtain the characteristic information common to both, and then the amount of characteristic information shared by both can be determined. Similarly, the characteristic information set of the medical record to be evaluated and the target historical medical record can be unioned to obtain the total characteristic information of both, and then the total amount of characteristic information shared by both can be determined.

[0081] S103. Determine the disease phenotype to which the medical record to be evaluated belongs based on the similarity between the medical record to be evaluated and each target historical medical record belonging to each disease phenotype.

[0082] Among them, the disease phenotype can refer to the type of the same treatment plan under the same diagnosis. Therefore, evaluating the medical records to be evaluated based on historical medical records with the same disease phenotype can more accurately evaluate the quality of the logic of the content in the treatment process.

[0083] It should be noted that the disease phenotype to which each target historical medical record belongs has been predetermined, so the disease phenotype to which the medical record to be evaluated belongs can be determined by referring to the disease phenotype to which the target historical medical record similar to the medical record to be evaluated belongs.

[0084] Optionally, in another embodiment of the present application, a specific implementation of step S103 includes:

[0085] The mean of the similarities between each target historical medical record and the medical record to be evaluated belonging to each disease phenotype is calculated, and the disease phenotype corresponding to the maximum value of the mean of the similarities with the medical record to be evaluated is determined as the disease phenotype to which the medical record to be evaluated belongs.

[0086] Considering that the similarity between a single target historical medical record and the medical record to be evaluated may be accidental, in the embodiment of the present application, the disease phenotype to which the target historical medical record most similar to the medical record to be evaluated belongs is not directly used as the disease phenotype to which the medical record to be evaluated belongs. Instead, for each disease phenotype, the mean of the similarity between each target historical medical record under the disease phenotype and the medical record to be evaluated is calculated. Finally, the disease phenotype with the largest mean similarity is determined as the disease phenotype to which the medical record to be evaluated belongs.

[0087] Since it is necessary to determine in advance the disease phenotype to which each historical medical record belongs, optionally, in another embodiment of the present application, before executing step S101, it further includes determining the disease phenotype to which each historical medical record belongs. Figure 2 As shown, a method for determining a disease phenotype to which a historical medical record belongs provided in an embodiment of the present application includes:

[0088] S201. Acquire multiple historical medical records to form a medical record database.

[0089] S202. Pre-process each historical medical record.

[0090] Optionally, preprocessing the historical medical records may include standardizing the data therein to unify the disease diagnosis name, symptom description, physical sign record, and test data, etc. In addition, duplicate and erroneous data may be removed, missing values ​​may be filled, and necessary data cleaning and formatting may also be performed.

[0091] S203. Based on the diagnosis results and core processing items of each historical medical record, each historical medical record is divided into a plurality of disease phenotypes to obtain the disease phenotype to which each historical medical record belongs.

[0092] It should be noted that the core treatment items are the main treatment items used to treat the disease, so the core treatment items and diagnosis results are first determined from the information of each historical medical record, and then the treatment item category to which the core treatment item belongs is determined, such as the category of drugs used, the category of surgical operations performed, etc. Finally, the target historical medical records with the same diagnosis results and the same category of core treatment items are classified into one disease phenotype.

[0093] S104. Extracting keywords of the medical records to be evaluated from the medical records to be evaluated.

[0094] It should be noted that the keywords in the medical record are key information of the diagnosis and treatment logic in the medical record, so the quality of the medical record can be evaluated by evaluating the keywords. Therefore, in the embodiment of the present application, each keyword in the medical record to be evaluated is extracted. For example, the medical record to be evaluated is: the patient had extensive chest and back pain without obvious causes 1 week ago, without radiation, and the degree was more serious with a sense of impending death. After resting for 20 minutes and taking quick-acting heart-saving pills, it improved, accompanied by chest tightness, sweating, dizziness, and headache. For further diagnosis and treatment, he was admitted to the hospital for "unstable angina pectoris" in the outpatient department of our hospital. In the past 2 weeks, the patient was in good spirits, had a good appetite, poor sleep, and normal bowel movements. The keywords extracted from the medical record to be evaluated may include: chest tightness, back pain, impending death, chest tightness, sweating, dizziness, headache, unstable angina pectoris, good spirits, good appetite, poor sleep, normal bowel movements, etc.

[0095] Optionally, the text content in the medical record to be evaluated may be extracted first, and then the text content may be processed by word segmentation, stop word removal, stem extraction, and the like, so as to obtain various keywords in the medical record to be evaluated.

[0096] S105. Based on the weights of the keywords corresponding to the disease phenotype to which the medical record to be evaluated belongs, calculate the similarity between the keywords of the medical record to be evaluated and the keywords corresponding to the disease phenotype to which the medical record to be evaluated belongs, and obtain an evaluation score of the medical record to be evaluated.

[0097] Among them, each keyword corresponding to a disease phenotype is composed of keywords extracted in advance from each historical medical record belonging to the disease phenotype in the medical record database, so the keyword corresponding to a disease phenotype is the keyword that the medical record of the disease phenotype analyzed from a large number of medical records under the disease phenotype should contain, so the quality of the medical record to be evaluated can be determined by comparing the keywords corresponding to the disease phenotype to which the medical record to be evaluated belongs with the keywords contained in the medical record to be evaluated. Specifically, the overlap of the keywords of the two can be compared, and the higher the overlap of the keywords corresponding to the disease phenotype to which the medical record to be evaluated belongs and the keywords contained in the medical record to be evaluated, the more keywords it contains that should be included, and the higher its quality. Therefore, in the embodiment of the present application, the similarity between the keywords corresponding to the disease phenotype to which the medical record to be evaluated belongs and the keywords contained in the medical record to be evaluated is calculated, and the similarity is used as the evaluation score of the medical record to be evaluated.

[0098] Furthermore, since different keywords have different importance for a disease of a certain disease phenotype, the weight of each keyword is also predetermined in the embodiment of the present application. In the embodiment of the present application, the weight of a keyword corresponding to a disease phenotype is the probability of the keyword appearing in a medical record when the medical record belongs to the disease phenotype. Because the greater the probability of a keyword appearing in a medical record of a disease phenotype, the more important it is to the disease phenotype and the less likely it is to be missing, so its weight should be greater. Therefore, in the embodiment of the present application, the probability of its appearance is directly used as its weight.

[0099] Optionally, another embodiment of the present application provides a method for determining the weights of keywords corresponding to disease phenotypes, such as Figure 3 As shown, including:

[0100] S301. Extract keywords from each historical medical record in the medical record database.

[0101] It should be noted that the method of extracting keywords from the medical records to be evaluated should be consistent with the method of extracting keywords from historical medical records to ensure that the same keywords can be extracted, so that the keywords of the medical records to be evaluated can be compared and evaluated using the keywords of the historical medical records.

[0102] S302 , respectively collect and remove duplicate keywords of historical medical records belonging to the same disease phenotype to obtain keywords corresponding to each disease phenotype.

[0103] S303. For each keyword corresponding to each disease phenotype, calculate the conditional probability of the keyword appearing in a historical medical record of the disease phenotype selected randomly from the medical record database, and obtain the weight of the keyword corresponding to the disease phenotype.

[0104] It should be noted that, since a keyword usually appears in the medical records of diseases with multiple disease phenotypes, the probability of its appearance in the medical records cannot accurately reflect its importance to the medical records of a certain disease phenotype. Therefore, in the embodiment of the present application, the conditional probability is used as the weight of the keyword, that is, under the condition of randomly selecting a historical medical record of a disease phenotype, the conditional probability of the keyword appearing in the selected historical medical record is used as the weight. Because the condition of selecting a medical record of a disease phenotype has been set, under this condition, the probability of the keyword appearing can reflect its importance to the medical record of the disease phenotype. Therefore, the weight of each keyword corresponding to a disease phenotype can be calculated by the Bayesian formula, which can be specifically expressed as:

[0105]

[0106] Among them, X represents a certain keyword, and Y represents a certain disease phenotype. Therefore, P(X) represents the probability of event X occurring, that is, the probability of keyword X appearing in a medical record in the medical record database. P(Y) is the probability of event Y occurring, that is, the probability of a medical record randomly selected from the medical record database belonging to disease phenotype Y. Correspondingly, P(X|Y) is the probability of event X occurring under the condition that event Y occurs, that is, the probability of keyword X appearing in a medical record with disease phenotype Y. P(Y|X) is the probability of event Y occurring under the condition that event X occurs, that is, the probability of a medical record with keyword X belonging to disease phenotype Y.

[0107] Therefore, we can calculate the corresponding probability by counting the number of times a keyword appears in the medical record database, the number of medical records belonging to a certain disease phenotype in the medical record database, and the number of times the keyword appears in the medical records belonging to the disease phenotype, and then calculate the weight of the keyword corresponding to the disease phenotype through the Bayesian formula. For example, in 100 electronic medical records, the keyword "cough" appears 45 times; the number of a disease phenotype in the 100 medical records is 25, and the number of times "cough" appears in the medical records belonging to the disease phenotype is 20. Then the weight of the keyword "cough" corresponding to the disease phenotype is:

[0108]

[0109] Optionally, the evaluation score of the medical record to be evaluated may be calculated by comparing the similarity between each keyword of the medical record to be evaluated and each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, and according to the weights of the similar keywords between the two.

[0110] Optionally, in another embodiment of the present application, a specific implementation of step S105 is as follows: Figure 4 As shown, including:

[0111] S401. Vectorize the set of keywords of the medical record to be evaluated and the set of keywords corresponding to the target disease phenotype respectively to obtain the keyword vector of the medical record to be evaluated and the keyword vector of the target disease phenotype.

[0112] The target disease phenotype refers to the disease phenotype to which the medical record to be evaluated belongs.

[0113] Since keywords cannot be used directly for calculations, in an embodiment of the present application, the set of keywords of the medical records to be evaluated and the set of keywords corresponding to the target disease phenotype are first encoded and converted into vectors corresponding to the two sets.

[0114] Optionally, the total set of all keywords contained in the diagnosis result of the medical record to be evaluated can be determined first. Then, the set of each keyword of the medical record to be evaluated and the set of each keyword corresponding to the target disease phenotype are respectively used as the set to be processed, and keyword matching is performed with the total set. If a keyword in the total set is included in the set to be processed, the value corresponding to the keyword is recorded as 1. If a keyword in the total set is not included in the set to be processed, the value corresponding to the keyword is recorded as 0, thereby obtaining a vector consistent with the dimension of the total set, and the vector is the vector after the keyword conversion of the set to be processed. For example, all keywords of the first course record of a medical record of a diagnosed disease include (a, b, c, d, e), that is, the total set of keywords includes (a, b, c, d, e). If the medical record to be evaluated belongs to disease phenotype α. The keywords corresponding to disease phenotype α are (a, b, c, d), and the keywords contained in the medical record to be evaluated are (a, b, c, e). Then, the set of keywords corresponding to the disease phenotype α and the set of keywords contained in the medical record to be evaluated are matched with the total set as the set to be processed. If the set to be processed contains a keyword in the total set, the value corresponding to the keyword is recorded as 1. If the set to be processed does not contain a keyword in the total set, the value corresponding to the keyword is recorded as 0. Thus, the keyword vector of the disease phenotype α is (1, 1, 1, 1, 0), and the keyword vector of the medical record to be evaluated is (1, 1, 1, 0, 1).

[0115] S402. Using the weights of the keywords corresponding to the target disease phenotype, weighted calculations are performed on the keyword vectors of the medical records to be evaluated and the numerical values ​​corresponding to the keywords in the keyword vectors of the target disease phenotype, to obtain the weighted keyword vectors of the medical records to be evaluated and the weighted keyword vectors of the target disease phenotype.

[0116] It should be noted that each keyword has its corresponding weight, and the keyword vector includes the numerical value corresponding to each keyword, so the weight corresponding to each keyword can be used to weight the numerical value corresponding to the keyword in the keyword vector, that is, each numerical value in the keyword vector is multiplied by the weight of the corresponding keyword.

[0117] S403 : Calculate the cosine similarity between the weighted keyword vector of the medical record to be evaluated and the weighted keyword vector of the target disease phenotype to obtain an evaluation score of the medical record to be evaluated.

[0118] It should be noted that the keyword vector of the medical record to be evaluated and the keyword vector of the target disease phenotype are used to represent the keyword sets of the two, so the similarity of the keyword vectors of the two can be used to obtain the similarity of the two in terms of keywords. In the embodiment of the present application, the cosine similarity formula is used to calculate the similarity of the two.

[0119] The present application provides a medical record data evaluation method, which pre-divides the disease phenotypes to which each historical medical record in the medical record database belongs, and extracts keywords from each historical medical record to form keywords corresponding to each disease phenotype, and calculates the weight of each keyword. Among them, the weight of a keyword corresponding to a disease phenotype is the probability of the keyword appearing in the medical record when a medical record belongs to the disease phenotype, so as to reflect the importance of each keyword to the medical record of the disease phenotype. When the medical record needs to be evaluated, the medical record to be evaluated is obtained. Then, based on the characteristic information in the medical record to be evaluated and each target historical medical record, the similarity between the medical record to be evaluated and each target historical medical record is calculated, and based on the similarity between the medical record to be evaluated and each target historical medical record belonging to each disease phenotype, the disease phenotype to which the medical record to be evaluated belongs is determined. Then, each keyword of the medical record to be evaluated is extracted from the medical record to be evaluated, and based on the weight of each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, the similarity between each keyword of the medical record to be evaluated and each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs is calculated to obtain the evaluation score of the medical record to be evaluated. Since the keywords corresponding to the disease phenotype are the key contents that should be included in the medical records of a disease phenotype, the similarity between them and the keywords of the medical records to be evaluated can be calculated to determine the quality of the content of the medical records to be evaluated, so that the evaluation is no longer just based on the surface form, so the quality of the medical records can be effectively guaranteed to be accurately evaluated. And because the evaluation is achieved by calculating the score through the similarity calculation of keywords, differences are allowed, that is, individual differences are allowed, to ensure accurate control of the quality of medical records.

[0120] Another embodiment of the present application provides a medical record data evaluation device, such as Figure 5 As shown, including:

[0121] The first acquisition unit 501 is used to acquire the medical record to be evaluated.

[0122] The similarity calculation unit 502 is used to calculate the similarity between the medical record to be evaluated and each target historical medical record based on the feature information in the medical record to be evaluated and each target historical medical record. The target historical medical record refers to the historical medical record in the medical record database with the same diagnosis result as the medical record to be evaluated.

[0123] The phenotype determination unit 503 is used to determine the disease phenotype to which the medical record to be evaluated belongs based on the similarity between the medical record to be evaluated and each target historical medical record belonging to each disease phenotype.

[0124] The first extraction unit 504 is used to extract various keywords of the medical records to be evaluated from the medical records to be evaluated.

[0125] The score calculation unit 505 is used to calculate the similarity between each keyword of the medical record to be evaluated and each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, based on the weight of each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, to obtain the evaluation score of the medical record to be evaluated. Among them, each keyword corresponding to a disease phenotype is composed of keywords pre-extracted from each historical medical record belonging to the disease phenotype in the medical record database. The weight of a keyword corresponding to a disease phenotype is the probability of the keyword appearing in the medical record when a medical record belongs to the disease phenotype.

[0126] Optionally, in another embodiment of the present application, the medical record data evaluation device further includes:

[0127] The second acquisition unit is used to acquire multiple historical medical records to form a medical record database.

[0128] The preprocessing unit is used to preprocess each historical medical record.

[0129] The phenotype division unit is used to divide each historical medical record into multiple disease phenotypes based on the diagnosis results and core processing items of each historical medical record, and obtain the disease phenotype to which each historical medical record belongs.

[0130] Optionally, in another embodiment of the present application, the medical record data evaluation device further includes:

[0131] The second extraction unit is used to extract each keyword of each historical medical record in the medical record database.

[0132] The merging unit is used to collect and remove duplicate keywords of various historical medical records belonging to the same disease phenotype, so as to obtain keywords corresponding to each disease phenotype.

[0133] The weight calculation unit is used to calculate the conditional probability of the keyword appearing in the selected historical medical record under the condition that a historical medical record of a disease phenotype is randomly selected from the medical record database for each keyword corresponding to each disease phenotype, and obtain the weight of the keyword corresponding to the disease phenotype.

[0134] Optionally, in the medical record data evaluation device provided in another embodiment of the present application, the similarity calculation unit includes:

[0135] The third extraction unit is used to extract each feature information of the medical record to be evaluated and each feature information of each target historical medical record from the medical record to be evaluated and each target historical medical record respectively.

[0136] The first calculation unit is used to divide the number of identical feature information between the medical record to be evaluated and the target historical medical record by the total number of all feature information between the medical record to be evaluated and the target historical medical record for each target historical medical record, so as to obtain the similarity between the medical record to be evaluated and the target historical medical record.

[0137] Optionally, in the medical record data evaluation device provided in another embodiment of the present application, the phenotype determination unit includes:

[0138] The mean calculation unit is used to calculate the mean of the similarity between each target historical medical record belonging to each disease phenotype and the medical record to be evaluated.

[0139] The phenotype determination subunit is used to determine the disease phenotype corresponding to the maximum value of the mean value of the similarity of the medical record to be evaluated as the disease phenotype to which the medical record to be evaluated belongs.

[0140] Optionally, in the above-mentioned medical record data evaluation device, the scoring calculation unit includes:

[0141] The vector processing unit is used to vectorize the set of each keyword of the medical record to be evaluated and the set of each keyword corresponding to the target disease phenotype, respectively, to obtain the keyword vector of the medical record to be evaluated and the keyword vector of the target disease phenotype. The target disease phenotype refers to the disease phenotype to which the medical record to be evaluated belongs.

[0142] The weighting unit is used to use the weights of each keyword corresponding to the target disease phenotype to perform weighted calculations on the keyword vector of the medical record to be evaluated and the numerical values ​​corresponding to each keyword in the keyword vector of the target disease phenotype, so as to obtain the weighted keyword vector of the medical record to be evaluated and the weighted keyword vector of the target disease phenotype.

[0143] The second calculation unit is used to calculate the cosine similarity between the weighted keyword vector of the medical record to be evaluated and the weighted keyword vector of the target disease phenotype to obtain the evaluation score of the medical record to be evaluated.

[0144] It should be noted that the specific working process of each unit provided in the above embodiments of the present application can refer to the implementation process of the corresponding steps in the above method embodiments, and will not be repeated here.

[0145] Another embodiment of the present application provides an electronic device, such as Figure 6 As shown, including:

[0146] Memory 601 and processor 602 .

[0147] The memory 601 is used to store programs.

[0148] The processor 602 is used to execute the program stored in the memory 601. When the program is executed, it is specifically used to implement the medical record data evaluation method provided in any one of the above embodiments.

[0149] Another embodiment of the present application provides a computer storage medium for storing a computer program. When the computer program is executed, it is used to implement the medical record data evaluation method provided in any of the above embodiments.

[0150] Computer storage media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0151] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0152] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A medical record data evaluation method, characterized in that: include: Obtain medical records for evaluation; Based on the characteristic information in the medical record to be evaluated and each target historical medical record, the similarity between the medical record to be evaluated and each target historical medical record is calculated; wherein the target historical medical record refers to a historical medical record in a medical record database having the same diagnosis result as the medical record to be evaluated; Determining the disease phenotype to which the medical record to be evaluated belongs based on the similarity between the medical record to be evaluated and each of the target historical medical records belonging to each disease phenotype; Extracting each keyword of the medical record to be evaluated from the medical record to be evaluated; Based on the weight of each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, the similarity between each keyword of the medical record to be evaluated and each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs is calculated to obtain an evaluation score of the medical record to be evaluated; wherein, one type of each keyword corresponding to the disease phenotype is composed of keywords pre-extracted from various historical medical records belonging to the disease phenotype in the medical record database; one type of weight of the keyword corresponding to the disease phenotype is the probability of the keyword appearing in a medical record when a medical record belongs to the disease phenotype.

2. The method according to claim 1, characterized in that Before obtaining the medical records to be evaluated, the following steps are also included: Acquire multiple historical medical records to form the medical record database; Preprocessing each of the historical medical records; Based on the diagnosis results and core processing items of each of the historical medical records, each of the historical medical records is divided into a plurality of the disease phenotypes, and the disease phenotype to which each of the historical medical records belongs is obtained.

3. The method according to claim 1, characterized in that Before obtaining the medical records to be evaluated, the following steps are also included: Respectively extracting each keyword of each of the historical medical records in the medical record database; Collecting and removing duplicates of the keywords of the historical medical records belonging to the same disease phenotype, respectively, to obtain keywords corresponding to each disease phenotype; For each keyword corresponding to each disease phenotype, the conditional probability of the keyword appearing in the selected historical medical record under the condition that a historical medical record of the disease phenotype is randomly selected from the medical record database is calculated to obtain the weight of the keyword corresponding to the disease phenotype.

4. The method according to claim 1, characterized in that: The calculating the similarity between the medical record to be evaluated and each target historical medical record based on the feature information in the medical record to be evaluated and each target historical medical record includes: Extracting each feature information of the medical record to be evaluated and each feature information of each target historical medical record from the medical record to be evaluated and each target historical medical record respectively; For each target historical medical record, the number of identical feature information between the medical record to be evaluated and the target historical medical record is divided by the total number of all feature information between the medical record to be evaluated and the target historical medical record to obtain the similarity between the medical record to be evaluated and the target historical medical record.

5. The method according to claim 1, characterized in that The determining the disease phenotype to which the medical record to be evaluated belongs based on the similarity between the medical record to be evaluated and each of the target historical medical records belonging to each disease phenotype includes: Calculating the average of the similarities between each of the target historical medical records belonging to each of the disease phenotypes and the medical records to be evaluated; The disease phenotype corresponding to the maximum value of the mean value of the similarity with the medical record to be evaluated is determined as the disease phenotype to which the medical record to be evaluated belongs.

6. The method according to claim 1, characterized in that The step of calculating the similarity between each keyword of the medical record to be evaluated and each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs based on the weight of each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, and obtaining the evaluation score of the medical record to be evaluated includes: Vectorizing the set of keywords of the medical record to be evaluated and the set of keywords corresponding to the target disease phenotype respectively to obtain the keyword vector of the medical record to be evaluated and the keyword vector of the target disease phenotype; wherein the target disease phenotype refers to the disease phenotype to which the medical record to be evaluated belongs; Using the weights of the keywords corresponding to the target disease phenotype, weighted calculations are performed on the keyword vector of the medical record to be evaluated and the numerical values ​​corresponding to the keywords in the keyword vector of the target disease phenotype to obtain the weighted keyword vector of the medical record to be evaluated and the weighted keyword vector of the target disease phenotype; The cosine similarity between the weighted keyword vector of the medical record to be evaluated and the weighted keyword vector of the target disease phenotype is calculated to obtain an evaluation score of the medical record to be evaluated.

7. A medical record data evaluation device, characterized in that: include: A first acquisition unit, used for acquiring the medical records to be evaluated; A similarity calculation unit, configured to calculate the similarity between the medical record to be evaluated and each of the target historical medical records based on the feature information in the medical record to be evaluated and each of the target historical medical records; wherein the target historical medical record refers to a historical medical record in a medical record database having the same diagnosis result as the medical record to be evaluated; A phenotype determination unit, configured to determine the disease phenotype to which the medical record to be evaluated belongs based on the similarity between the medical record to be evaluated and each of the target historical medical records belonging to each disease phenotype; A first extraction unit, configured to extract each keyword of the medical record to be evaluated from the medical record to be evaluated; A scoring calculation unit is used to calculate the similarity between each keyword of the medical record to be evaluated and each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, based on the weight of each keyword corresponding to the disease phenotype to which the medical record to be evaluated belongs, so as to obtain an evaluation score of the medical record to be evaluated; wherein, one type of each keyword corresponding to the disease phenotype is composed of keywords pre-extracted from each historical medical record belonging to the disease phenotype in the medical record database; one type of weight of the keyword corresponding to the disease phenotype is the probability of the keyword appearing in a medical record when a medical record belongs to the disease phenotype.

8. The device according to claim 7, characterized in that Also includes: A second acquisition unit, used for acquiring a plurality of historical medical records to form the medical record database; A preprocessing unit, used for preprocessing each of the historical medical records; The phenotype classification unit is used to classify each of the historical medical records into a plurality of disease phenotypes based on the diagnosis results and core processing items of each of the historical medical records, so as to obtain the disease phenotype to which each of the historical medical records belongs.

9. An electronic device, characterized in that: include: Memory and processor; Wherein, the memory is used to store programs; The processor is used to execute the program, and when the program is executed, it is specifically used to implement the medical record data evaluation method as described in any one of claims 1 to 6.

10. A computer storage medium, characterized in that: Used to store a computer program, which, when executed, is used to implement the medical record data evaluation method as described in any one of claims 1 to 6.